Feed aggregator

Tom Clancy's The Division Resurgence is still expected to land on mobile devices at some point, but we're yet to have an exact release date or even an updated release window. Fortunately, we do have news of upcoming betas, so there's still a decent amount of information to sink our teeth into.

With the most recent trailer having been showcased as part of Ubisoft Forward, there’s no better time to start looking into why The Division Resurgence should be one of your most anticipated upcoming games. Hopefully, we'll hear an update in the coming months, given that the beta is running in September.

Here’s everything we know about what could be one of the best Android games of recent years. From the latest news, to info on what to expect from the gameplay, let's dig into what's been said about The Division Resurgence so far.

The Division Resurgence - cut to the chase 

• What is it? A free-to-play, third-person shooter RPG mobile game
• When can I play it? TBC
• What can I play it on? Android and iOS
• Who’s making it? Ubisoft

The Division Resurgence platforms

(Image credit: Ubisoft)

The Division Resurgence will be released exclusively on Android and iOS devices, shaking things up from the previous The Divison games which were released on console and PC, but offering a new story alongside its new platform. In addition, the game is entirely free to play on mobile devices. Should the release be as streamlined as we are anticipating, it could be a hell of a hit on mobile.

There’s currently nothing to suggest that the game will eventually come to PC or home console, but basing our assumption on how the game appears to be built exclusively for mobile devices right now, we believe that there are no plans for the game to ever leave its mobile exclusivity.

The Division Resurgence story and setting

The Division Resurgence promises a new storyline based on the previous two Division games while maintaining a familiar setting stripped from the first game. The Division Resurgence puts you in the position of becoming a member of the first wave of Strategic Homeland Division agents, amidst a virus outbreak in New York City.

New York will set the scene for the game, as you will need to defend the metropolitan area against waves of enemies and crime syndicates, alongside completing missions to help you explore the area.

There will be a variety of PVE missions on offer, both as part of the central campaign and as unique side missions to help you explore the world outside of your placement in New York. We are yet to learn about what these missions will specifically involve, but we do know that the post-apocalyptic setting is home to numerous hostile enemies you should expect to face during your time with the title.

The Division Resurgence gameplay

In terms of gameplay, we know a pretty limited amount about what to expect from The Division Resurgence as a whole. We’ve received a short, three-minute look at the in-game content, but aside from using a minimap to track a mission, a slight look at how combat plays, and a look at the new compact UI designed specifically for mobile play, there’s a lot to learn still.

Even though The Division Resurgence is set to arrive on a new platform for the franchise, you can expect the game to play in the same style as the previous Division games. Between tense combat and completing missions, you’ll be able to work your way through a skill tree, alongside exploring what the new classes offer and how you can utilize them during your time in Resurgence.

Each class or specialization has a unique skill tree, so whether you’re more of a vanguard or if you’d rather rely on explosives for a quick and easy takedown, you’ll need to focus on improving your own skills to help out teammates and progress through side missions, and the campaign.

In addition, the time it takes to complete central and side missions has been reduced, so you will be able to play in short bursts while on the go, ideal for any mobile gamer. Despite these missions being significantly shorter in completion time in comparison to The Division and The Division 2, that doesn’t mean they lack detail.

You should expect to have your skills put to the test in each mission, whether that’s through a grab-and-go mission or something more hands-on with the enemies. Looting and crafting have also been highlighted to be integral to the game, being the core mechanic when upgrading your weapons and your gear.

The Division Resurgence trailers

The most recent look we’ve had at The Division Resurgence is through a Tactical Action CGI Trailer which was shown as part of Ubisoft Forward Live in June 2023. Within this trailer, we see an animated rendition of the setting, characters, abilities, and combat we can expect to experience throughout the game. We also meet a few enemies and their coinciding attacks, before seeing what looks like a look at actual in-game footage followed by the release window of Fall 2023. This was later followed by a delay to Early 2024.

We’ve scattered a few additional trailers for The Division Resurgence throughout this article, but if you want to find out a little more for yourself, it’s worth looking at Ubisoft's official YouTube channel.

The Division Resurgence beta start date

(Image credit: Ubisoft)

There's an upcoming beta for The Division Resurgence, announced during Gamescom 2025. The game will go back into closed testing sometime in September 2025. You can hear this confirmed in the video below:

The Division Resurgence news

(Image credit: Ubisoft)

The Division Resurgence delayed
Ubisoft released its first quarter 2024-2025 earnings report, revealing that The Division Resurgence has been delayed:

"Rainbow Six® Mobile and The Division® Resurgence are no loger expected in FY25 as the teams are taking the necessary time to ensure that these experiences deliver on expectations with optimized KPIs in the context of a demanding yet very large market."

The Division Resurgence gets a world introduction trailer
Six months after the release of its official reveal, The Division Resurgence world introduction trailer offers a more in-depth look at the world we will explore within the game. Similarly to the first The Division game, we are taken back to New York to embark on some high-risk high-reward missions in an attempt to save what remains of civilization. Between a cinematic introduction and in-game content,

The Division Resurgence receives official reveal trailer
The official reveal trailer for The Division Resurgence launched in 2022, offering the first look at what to expect from the mobile-only title. Transporting us back to New York, the original setting for the first Division game, we get to see the collapse of civilization and agents of the division stepping up to save what remains. A few different areas of what we expect to be the map are shown before the trailer comes to a close with no release date.

You Might Also Like...

• Everything we know about the Nintendo Switch 2
• Our picks for the best FPS games
• Here are the best co-op games to play in 2025

The H2S from Bambu Lab is the XL 3D printer that people have been asking for

• ChatGPT Projects now have their own separate memory
• It stops your projects interfering with the rest of ChatGPT
• Available now in the web interface and coming to iOS and Android in a few weeks

ChatGPT Projects are ways to keep all your work on a particular subject in one place. If you create a new project then you can group together chats, files you’ve uploaded, and custom instructions all in one place.

Say, for example, that you’re planning a trip to go kayaking in the Florida Keys. Instead of having all your conversations about the trip, discussions of routes, and places to say spread all over your ChatGPT chats list, you can group them altogether in a Project, so you can easily refer to them later.

Saved memories

OpenAI has just added a new feature that affects ChatGPT’s memory in relation to projects. So, if we use our Florida Keys tips as an example, it could be that one of your party is a vegetarian, so you can ask ChatGPT to remember within the project that it should always recommend restaurants that have good vegetarian options.

When you get back to your normal ChatGPT conversations, this memory wouldn’t be helpful, so now there’s a way to confine that memory to the project.

Starting a new project is easy – you just click on New Project in the left hand menu on the web interface.

When you open a new project click on ‘More options’ then change Memory from Default to Project-only and you’re good to go. ChatGPT will now draw context only from conversations already in the project and your previous saved memories from outside the project will have no effect on what’s inside the project.

You can activate Project-only memories when you open a new project. (Image credit: OpenAI)Islands of thought

There is one thing to check first though, for this to work, ChatGPT’s memory must be enabled in Settings. Go to Settings, then Personalization then Memory and make sure Reference Saved Memories is turned on.

Currently Project-only memory is only enabled for ChatGPT in the website and in the Windows app, but it’s coming to iOS and Android versions of the app in the coming weeks.

With Project-only memories, your projects can now become the tightly-focused and self-contained islands of thought they were always designed to be.

You might also like

• AI music is fine until it starts pretending to be real people
• I finally got ChatGPT to stop asking 'Want me to…' at the end of every response – here’s how to do it
• Your Grok chats are now appearing in Google search – here’s how to stop them

The H2S comes hot on the heels of the H2D full laser combo and standard H2D 3D printer (see my review here), but with a few major differences that make it a superb option for any enterprise business requiring a machine that offers scale, reliability and speed.

It might not have the headline-catching dual nozzle that so interested the 3D printing market, but with much the same form factor, just with a single nozzle rather than dual nozzle tool head, it offers something a little different.

Essentially, Bambu Lab is touting this as a bigger X1C, which has become a staple for many design studios due to its print quality, flexibility, multiple filament printing, and of course, speed. Now, with the H2S taking yet another leap in scale and performance, and what's more, it's also coming in at a relatively affordable price point.

In use, the H2S is impressive, and the single-nozzle hotend is highly tuned for precision and quality. The speed, while quoted as faster than the X1C, in real-world tests using standard settings is on par; however, tweak the options and the full speed of the machine is unleashed, seeing a standard 20-30% boost in print times across the board. The H2S also has a far larger print bed, enabling more potential with the style and type of prints you're able to produce. There's absolutely no doubt that, when it came to side-by-side print quality, the X1C is one of the market leaders, but the H2S just took the edge during my test.

One feature that appeals to me about the H2S is that, compared with the H2D, it is far more affordable and there is a laser option as well, in fact, the new H2S line exactly reflects the H2D, just at a more affordable price.

What this essentially means is that, for any product design studio or anyone looking for the best 3D printer for quality, the Bambu Lab H2S is a balanced option, especially if you mainly print with a single filament with the occasional use of multifilament.

Bambu Lab H2S: Price and availability

There are three versions of the H2S available: the Base, the AMS Combo, and the Laser Full Combo. The two more advanced versions, the AMS Combo and the Laser Full Combo, are available now for $1,499 and $2,099, respectively. Bambu Lab will also be releasing the Base unit, which is just a single-nozzle machine, for $1,249.

Each model offers an upgrade route, so if you are working on a budget and just want an exceptionally good single FFM printer, you can start with the Base, then upgrade with the AMS Pro 2 to the AMS Combo. If you want to go all the way, you can upgrade to the Laser Full Combo, but only pick out the additions that you need.

Bambu Lab H2S: Design

(Image credit: Bambu Lab // Future)

Like the H2D, the H2S has almost an identical form factor, so once again it's a large-scale machine that will take up a sizable amount of your desktop, and now that I have three of these machines side by side, that doesn't leave a lot of space for much else.

The fact is, however, that with the H2S, it cuts away a lot of the frills of the H2D and the excitement those machines create. Instead, the H2S goes back to the core of 3D printing, offering a single nozzle toolhead and a few additional features that help elevate what's possible. It positions itself as a successor to the phenomenal X1C, or if not a successor, then a larger-scale version that takes on the new form factor of the H2 series.

At the base level, the first thing you notice alongside that nozzle is the larger print volume, which measures in at 340 × 320 × 340 mm, making it one of the largest on the market. It also once again utilises the CoreXY system, offering high-speed printing, both single and multifilament, when coupled with the AMS Pro 2.

Once again, the 3D printer is fully enclosed, so if you want to print with more advanced materials, that is perfectly possible, with a heated bed that reaches up to 120°C and a 65°C actively heated chamber. This is great for materials like ABS, nylon, and ASA.

One of the features that I've really liked across all of the H2 series is that large touchscreen. It's just really intuitive, beautifully designed, and very quick to respond, which is not always the case.

Inside, they've also boosted the number of sensors, with five flame sensors covering monitoring, emergency stop, and airflow optimisation, once again almost mirroring the H2D series, but building on the feedback and experience of those first H2 machines.

When it comes to the major differences between the H2S and the H2D, aside from that single nozzle toolhead, most of the other differences come in the technology and steps forward in development. For instance, there's second-generation extrusion motion technology with a PMSM servo motor that exerts a 67% increase in extrusion force. There's also a 20 kg feedback load for precise filament control.

Once again, we see the Vision Encoder, which is optional, but will boost the motion accuracy for prints under 50 microns. This is essentially industrial-grade precision, so it will really only be of interest if you're using this for high-precision product design. However, if you're into gaming and want to make desktop game figures, then this is going to be a great feature. I have noted that there are several 3D artists now producing FFF-ready miniatures. With the Vision Encoder having industrial-grade precision, it could mean that printing miniatures on an FFM machine should now, in some cases, be possible, meaning that you don't have to deal with messy resin.

Bambu Lab has also taken a look at monitoring and intelligence, and the H2S includes 23 sensors inside the machine that are monitoring the print process and conditions. This is coupled with three onboard cameras, one for live view, one for the nozzle, and one for monitoring. This compares to just one on the H2D. There's also smarter fault detection if a print does go wrong, and I have to say, during the three-week test, I didn't have any print failures, even when I tried with an impossible print, the machine just detected it and paused.

Now, while the form factor of the machine is identical to the H2D, when it comes to the build volume, because it only has a single nozzle rather than two, the full print bed can be used, so you have a larger print volume of 340 × 320 × 340 mm compared with 325 × 320 × 325 mm for single-filament printing on the H2D.

Bambu Lab also says that the H2S is a high-speed machine, capable of printing up to 1000 mm/s and with acceleration over 20,000 mm/s². However, during my tests, I'd say they are both pretty much on a par until you delve manually into the settings.

Once again, there's a heated chamber, and in the H2S, this is capable of reaching around 65°C, exactly the same as the H2D. Likewise, the hardened steel nozzle seems to be the same, as does the 5-inch touchscreen.

One of the big questions I had when the machine arrived was: because it is a single-nozzle machine, does it still have the capability of using a laser and cutter? There are, of course, three versions of the H2S, including the Laser Full Combo version that offers laser engraving, cutting, digital cutting, and precision drawing capabilities.

Bambu Lab H2S: Features

(Image credit: Bambu Lab // Future)

Take a look through the machine's features, and aside from the single nozzle, which is obvious, the main feature aside from technology enhancement is the larger print area, which means that this is essentially a large successor to the X1C. Through the test, the speed of the H2S impressed. Put alongside the X1C on standard prints, print speeds were essentially similar until you delve into the settings. However, there are more options, and if you really want to push the speed of the printer, it is possible to get up to a 30% faster printout of the H2S than the X1C, although there is a small drop in quality.

The maximum speed of the printhead goes up to 1000 mm/s with 20,000 mm/s² acceleration. So, while the acceleration matches the X1C, the print speed essentially doubles, with a real-world speed increase of around 30%.

Like the H2D, the single nozzle of the H2S is capable of reaching 350°C, which makes it a perfect option for printing engineering-grade materials such as ASA and nylon. The H2S is a fully enclosed environment, which helps ensure optimal printing conditions for those materials, and with the filtration also helps to remove any toxins from the materials as they print.

Once again, Bambu has included a whole series of sensors that will detect if any print goes wrong and stop proceedings before issues are caused. Through the test, aside from running out of filament, I only had one occasion where the error-checking needed to kick in, due to some filament falling onto the print bed from a previous print. The new machine now includes 23 sensors and three onboard cameras. During the test, I didn't have any tricky situations with printing. The one issue that could've caused a problem was instantly recognised, and the printing was paused until I removed the filament, then the print completed successfully.

What I also liked was that when it came to filament changes due to running out, the process of restarting the print was absolutely spot-on. There wasn't the usual worry about layer misalignment that can often happen.

One of the big new features of the extruder is that it now includes a PMSM servo motor that can apply 67% more force. While I couldn't visually see the difference, the extrusion of the filament did seem exceptionally smooth for all materials tried, including Carbon fill and TPU. When I compared prints between the H2D and the H2S, there was very little difference; it was only when testing a specific model that I could see a very slight uplift in the print quality coming from the H2S.

Like the ability to add a laser cutter and digital cutter to the machine, there's also the option to add the Vision Encoder, which you'll need if you're going to be using this for professional product design or any dimensionally critical work. The Vision Encoder can be used for any print that needs 50-micron precision.

Through the software, there's also hole and contour compensation to assist with more intricate models. Again, the intelligent monitoring will stop if there is an issue. During the test, I tried to push the printer with a few excessive overhangs just to see what would happen, and the machine essentially paused as the print failure was detected.

As with the H2D machines, there's AMS Pro 2, and again, the new model is fully compatible with the original AMS and the AMS HT. You can also link several AMS modules together so you can extend printing to more than just four filaments, up to 16, and some people have even claimed it can be pushed further.

Bambu Lab H2S: Performance

(Image credit: Bambu Lab // Future)

Getting started with the H2S is as simple as any other Bambu Lab 3D printer, as it essentially comes pretty much ready to go. In this review, I'm looking at the Bambu Lab AMS Combo, so you get the 3D printer and the AMS Pro 2 to enable multifilament printing.

Getting it all set up is extremely straightforward. First, the packaging is stripped away, the AMS is popped on top, one tube connects the AMS to the machine, and then there's a data cable that also connects between the machine and the AMS. Once that's done, the activation key is popped into the back along with the power cable, and the machine is switched on.

After running through the initialisation and calibration process, which is all handled by the machine, you can then install the four filaments and get started with connecting to Bambu Studio and producing your first print.

As ever with this machine, there are a few prints just to get you started, and it shot through these in no time at all, with the 3D Benchy model looking exceptionally good, fine detail and a high-quality finish to all of the outer walls and using the default settings this took 33 minutes with all checks.

As I went through the test, there were some larger models to print to create a Shadow Force suit of armour, and this is where the machine really started to come into its own. The ability to print at a scale that is normally only possible with large-scale printers, combined with the high nozzle quality, meant that the end result had a surface cleanliness and consistency that you don't usually get with large-scale consumer 3D printers.

What I really liked was that, for quite a few of these large-scale prints, I did run out of filament partway through. However, the filament detection worked exceptionally well. As with many machines, it detected the issue, but what impressed me was that feeding in the new filament aligned perfectly, continuing the print from exactly where it left off. Adhesion to the print bed was sufficient that there was no movement or issue, even when the print had been left for a few hours in between filament changes.

For single filament printing, the print speed was exceptionally good, although at present, there isn't a huge array of presets for materials. So, if you are new to 3D printing and want high-speed prints, it is a little tricky to adjust the settings to get exactly what you want. However, if you are experienced with Bambu Studio, then delving into the quality, strength, speed, and support settings will enable you to fine-tune for those higher print speeds.

At those high speeds, the quality actually remains exceptionally good. While you can see a slight reduction in outer surface quality, for prototyping or quickly printing ideas, the quality is beyond what I've seen with most 3D printers of this type, size and price.

Going on to multifilament printing, again, the speed of the print was good. But, as with all other Bambu Lab AMS system printers, the volume of waste that gets kicked out of the back is quite extraordinary. So, while multifilament 3D printing is obviously a major pull for this machine, the fact that it still produces so much waste is a bit of an issue.

The quality of the 3D prints, however, is once again exceptionally good, with a high-quality finish and one of the best surface finishes for a multifilament print that I've seen. Multifilament print times are obviously slower on the H2S than compared with the H2D. That second nozzle makes a huge difference to both the amount of waste produced and overall duration. Still, because of the high-speed printing that is possible with the H2S, the speed was still good.

As an example, the Robo Tortoise four-colour print took just over 24 hours. On the H2D with its two nozzles, this print took only 20 hours. So, while there is a reduction in speed between the H2S and the H2D for multifilament prints, it isn't huge. The biggest difference really is the amount of waste produced.

As I ran through the test prints, I became more taken with this machine. While the single nozzle isn't as headline-grabbing as the dual nozzle of the other H2D machines, it's still exceptionally high quality and essentially makes this machine a top option for any designer looking to prototype or even run small production batches.

The quality of the casing and the size give you plenty of flexibility in how you use it, and the fact that it has that heated chamber enables you to use engineering materials such as nylon, ABS, and ASA. What really impressed me was that, across multiple prints, there were no real failures.

Bambu Lab H2S: Print quality

(Image credit: Bambu Lab // Future)

Target 25 = X: 24.88mm / 0.12mm Error | Y: 24.96mm / 0.06mm Error

Target 20 = X: 19.87mm / 0.13mm Error | Y: 19.90mm / 0.10mm Error

Target 15 = X: 14.87mm / 0.13mm Error | Y: 14.80mm / 0.20mm Error

Target 10 = X: 9.90mm / 0.10mm Error | Y: 9.80mm / 0.20mm Error

Target 5 = X: 4.90mm / 0.10mm Error | Y: 4.80mm / 0.20mm Error

X Error Average = 0.114

Y Error Average = 0.152

X & Y Error Average = 0.133

Fine Flow Control – Score of X

Fine Negative Features – Score of X

Overhangs – Score of 5

Bridging – Score of 5

XY Resonance – Score of 5

Z-Axis Alignment – Score of 5

Total Score – 29 / 30

Bambu Lab H2S: Final verdict

(Image credit: Bambu Lab // Future)

At the end of the test, there's no doubt that I'm impressed once again with this Bambu Lab machine. The H2S takes the scale of single-nozzle 3D printing up a notch, and unlike many other large-scale 3D printers, it offers a brilliant enclosed environment. This enables you to print with engineering materials, not just mainstream PLA that can be used with open frame machines.

Due to the CoreXY mechanics, you also have ultra-fast print speed, reliability, and the dimensional accuracy that comes with it. The test prints that came out of this machine were some of the cleanest I've seen, and at this scale and consistency, it's really difficult to match with any other model.

While the machine compares very closely to the other H2D models, just with a single nozzle, when it comes to print quality and speed, it takes a small leap forward for single-filament printing. Where it really comes into its own is in being able to print larger than any other Bambu Lab printer, and the scale of those prints is genuinely exceptional.

Although the prints Bambu sent me for testing were clearly optimised to show the printer at its best, I also tried it with other complex single-filament prints. It coped with contours, overhangs, and overall print quality exceptionally well. As ever, the AMS Pro 2 is a standout feature, and since I was reviewing the combo version, I was able to give that a proper test with the machine and once again it works exceptionally well, with the only downside being the amount of waste that is produced.

The quality of the multifilament print was again exceptional, and the fact that it's going through a single nozzle just gives you a little more real estate to print large. The full-scale helmet from the Cosplay armour I printed came out with exceptional quality and a support structure that was easy to remove.

One of the big points I would highlight while using this printer was not only the speed and scale, but also the quality of the surface finish.

Surfaces were exceptionally smooth, and when I printed the Autodesk Kickstarter test model I've been using for many years, it was almost a full 30 out of 30. The print surface was exceptional, and the only thing that let it down slightly was dimensional accuracy, although my thought is that, if I used the Vision Encoder, I'd likely hit that full 30 out of 30 for the first time.

Regarding the fine flow test, while I've had many 3D printers that can produce a very clean result, there are often slight signs of angel hair filament, with only the AnkerMake M5C previously producing a model with this level of cleanliness.

The other point I really like about the H2S is its simplicity. The base unit isn't expensive considering. If you're looking at this as an enthusiast or business user, the fact that you can buy the base unit for just over $1,000 and then upgrade all the way to the Laser Combo means it's relatively affordable and offers a solid upgrade path for the future.

At the end of the test, once again, Bambu Lab has demonstrated why it leads the market with its 3D printers, and the H2S single nozzle combo version reviewed here once again shows the level that all other 3D printer manufacturers need to aim for. However, with the likes of the Original Prusa XL and SnapMaker U1, while it's great to see a larger print area single nozzle machine it feels like there is a move now into multitool head options which could leave Bambu Lab behind in the technology stakes.

Image 1 of 2

(Image credit: Bambu Lab // Future)Image 2 of 2

(Image credit: Bambu Lab // Future)Should I buy the Bambu Lab H2S?Buy it if...

You need a large build volume and speed.

Once again, this printer comes in the superb H2 form factor that offers a fully enclosed environment. This time, the single nozzle enables an even larger print area that should suit most home and small business projects.

You need precision for engineering projects

While a fully enclosed environment is nice for some users, if you're working in education or in the professional sector and need to print in engineering materials, this enclosed environment with its heated chamber is absolutely essential.

Don't buy it if...

You need native dual-nozzle multi-material

While single-filament printing is exceptionally good, if you need to use support materials as standard, then upgrading to an H2D will probably be a better idea.

You're a casual or first-time hobbyist

Starting with one of the best 3D printers on the market might seem like a great idea, but with so many superb printers at a fraction of the Price, it might be worth starting with a cheaper printer before investing in the H2S.

For more tech for your crafting studio , I've tested the best laser engravers.

Hasselblad X2D II 100C: two-minute review

If you've read any of my articles about Hasselblad's 'medium-format' mirrorless cameras, you'll know I have a soft spot for the premium Swedish camera maker's X-series.

They're stunning minimalist cameras with quality craftsmanship that are simple to operate and a joy to use. Naturally, it also helps that the recent models can shoot the best-looking photos this side of $10,000.

The latest addition to the series is the X2D II 100C, which is an update of my dream camera; it's refined in all the right ways, and it costs much less than its three-year-old predecessor, pushing it closer to reality for photographers on the edge of splashing out.

If I were to pick one camera for landscape photography, this would be it. You can expect highly-detailed 100MP resolution, gorgeous natural colors that need little to no editing, and 16-bit files with rich HDR, along with probably the best image stabilization performance for this full-frame-dwarfing format.

Usually cameras in this 44 x 33mm sensor format are largely confined to the studio for still life and static subjects, but the X2D II is way more versatile than that, especially since it gets an all-new autofocus system, equipped with subject-tracking continuous autofocus and LiDAR.

For the first time with a Hasselblad I felt confident using its autofocus for shooting portraits, given its largely reliable people-detection skills. There's no obvious gap in autofocus performance between Hasselblad and a Fujifilm GFX model now.

I also felt comfortable shooting without a tripod, getting sharp shots handheld with shutter speeds in excess of one second.

Don't get me wrong – I still needed to be more involved in the process to get the results I wanted than when shooting with my full-frame Nikon, but the X2D II has strong automatic chops too.

Otherwise, for the most part the shooting experience felt largely the same as with the original X2D 100C, which is fine by me.

The X-E5 can be purchased in a bundle with this nifty, space-saving 23mm pancake lens. (Image credit: Tim Coleman)

The X2D II 100C has what for me is probably the best camera design bar none: a gloriously minimalist body with exotic curves, a high-quality build with a slight reduction in weight compared to its predecessor, a larger-than-average tilting touchscreen with gorgeous color and clarity, plus various refinements that include a new joystick for adjusting the autofocus point and more.

A built-in SSD provides 1TB of storage for those mammoth 200MB raw files, and a CFExpress Type B card slot offers another storage option.

Some downsides remain: it has middling battery life, there's still no video recording capabilities (which for some could be a plus, but for me would be a nice to have), while the quality of the 5.76m-dot viewfinder display feels only okay.

And while this is the best-performing Hasselblad yet, Fujifilm GFX rivals can shoot faster, while full-frame alternatives are faster still. The X2D II is a steady shooter, just not one for action.

But again, that's fine – a camera that can deliver such rich detail can't be expected to sustain high-speed burst sequences.

Improved autofocus performance was probably my top wish for a X2D 100C successor, and that wish has come true, alongside other design refinements which improve the overall user experience.

A pleasant additional surprise is the X2D II 100C's list price: $7,399 / £6,400 (AU$ pricing TBC). That's a good deal less than the X2D 100C, even if it's still a pricey premium snapper.

Hasselblad X2D II 100C: price and availability

• Available now for $7,399 / £6,400 (Australia pricing TBC)
• Costs less than the original X2D 100C
• Launched alongside a new XCD 35-100mm f/2.8-4 zoom lens

Hasselblad products are pricey, and with good reason – the build quality is top-drawer, as is the medium-format sensor tech, plus there's a 1TB SSD built-in. Yet despite several refinements in this latest model, it actually costs much less than the original X2D 100C, which was launched in September 2022.

The price difference depends on the region, but in the UK the X2D II 100C costs £1,000 less than its predecessor. It's still pricey, as is the new 35-100mm zoom lens (28-80mm effective), which costs $4,599 / £4,100 (Australia pricing TBC), but this Hasselblad is edging closer to Fujifilm in terms of pricing.

• Price score 4/5

Hasselblad X2D II 100C: specs

Sensor:

100MP 'medium-format' sized 43.8 × 32.9mm

Video:

None

Cont. shooting:

3.3fps (approx)

Viewfinder:

5.76m-dot OLED

LCD:

3.6-inch, 2.36m-dot tilting touchscreen

Battery (CIPA rating):

Up to 327 shots, extended to 466 shots with HDR off

Weight:

29.6oz / 840g (with battery)

Dimensions:

5.85 x 4.17 x 2.95 inches / 148.5 x 106 x 75mm

Hasselblad X2D II 100C: design and handling

• Very similar design to its stunning, minimalist predecessor
• Slightly lighter than before, new joystick added
• Large tilting touchscreen with stunning display
• 1TB built-in SSD and CFExpress Type B card slot

Having tested the Hasselblad X2D 100C at length, the X2D II feels very familiar. At 840g with battery, the latest model is slightly lighter, but it retains the same form factor, with a wonderfully minimalist layout and an exaggerated grip. Given the sensor format, the X2D II is impressively compact.

The finish on the high-quality body is different, and I prefer the new version both in appearance and feel.

We have the same touchscreen as before, and it's delightful. At 3.6 inches it's larger than most, plus it's wonderfully bright, rich in color and contrast, and its touch response is spot on, which is something that can't be said for all cameras – touch response feels more smartphone-like than camera-like.

Initially, pulling out the tilt screen stills feels a little fiddler than it could be, and the tilt motion is stiff; but it feels durable, and the up and downward tilt range is handy, especially for shooting at low angles.

Image 1 of 10

(Image credit: Tim Coleman)Image 2 of 10

(Image credit: Tim Coleman)Image 3 of 10

(Image credit: Tim Coleman)Image 4 of 10

(Image credit: Tim Coleman)Image 5 of 10

(Image credit: Tim Coleman)Image 6 of 10

(Image credit: Tim Coleman)Image 7 of 10

(Image credit: Tim Coleman)Image 8 of 10

(Image credit: Tim Coleman)Image 9 of 10

(Image credit: Tim Coleman)Image 10 of 10

(Image credit: Tim Coleman)

That tilt range also means the screen clears the viewfinder, which sticks out from the body, for unimpeded viewing from low angles. The large 5.76m-dot viewfinder display is lovely too, although it's far from the absolute best viewfinder I've used.

Once again we get a 1TB built-in SSD for storing images – and you'll need it, with each raw file coming in at a whopping 200MB. There's also the option to use removable media in the form of CFXepress Type B, which is one of the fastest card types today – I made use of both options during my review period.

Image 1 of 8

(Image credit: Tim Coleman)Image 2 of 8

(Image credit: Tim Coleman)Image 3 of 8

(Image credit: Tim Coleman)Image 4 of 8

(Image credit: Tim Coleman)Image 5 of 8

(Image credit: Tim Coleman)Image 6 of 8

(Image credit: Tim Coleman)Image 7 of 8

(Image credit: Tim Coleman)Image 8 of 8

(Image credit: Tim Coleman)

One of the headline new features is an autofocus system that's totally new to Hasselblad, which I'll get onto in the next section, and in terms of the design this has prompted the addition of a joystick. This can be used to manually select the autofocus point – I find it a more intuitive method than the touchscreen. Despite the addition of the joystick, which is very welcome, the design still feels wonderfully minimal.

The same goes for the UI – menus are still super stripped-back and easily navigable, even with a new range of autofocus modes. The omission of video recording helps here, while there are no color profile options at all, and I really appreciate how quick and easy it is to make adjustments to settings.

I have only one niggle with design – the protective door for the USB-C port (the port which can be used to charge the camera), pops open all too easily, especially when using a camera strap.

• Design and handling score 5/5

Hasselblad X2D II 100C: features and performance

• 5-axis sensor-shift IBIS with up to 10EV stabilization
• New subject-detection autofocus recognizing humans, animals and vehicles
• Continuous shooting up to 3.3fps

The original X2D's autofocus is poor by today's standards – I used it to shoot a wedding, and I quickly reverted to my old DSLR for moments when I needed autofocus with a quick response.

Hasselblad has sensibly focused its attention on various performance improvements for the X2D II, and chief among them is a completely new autofocus system. The X2D II's focusing is a different proposition altogether, almost matching Fujifilm GFX standards.

This is the first Hasselblad camera with continuous autofocus, plus the first with subject-detection autofocus, being able to recognize and track humans, animals and vehicles. It also utilizes LiDAR focusing, an improvement that comes courtesy of drone maker DJI (which bought Hasselblad some years ago, and which introduced the built-in SSD in previous models).

I tested the camera on all of the subject types it's able to recognize, and it performed well, but don't expect the same degree of autofocus sophistication you'll get from, say, recent Sony / Canon / Nikon models.

Image 1 of 2

I shot this self portrait remotely using Hasselblad's Phocus software, triggering the camera with my Apple iPad (Image credit: Tim Coleman)Image 2 of 2

Animal subject detection picked up this bird for sharp focusing, no problem (Image credit: Tim Coleman)

Still, given it was my first time shooting portraits on location with a Hasselblad, I felt confident that the focus was going to be pin-sharp on my subject's eyes, and that's a big deal. The AF was also able to track my children's pet guinea pigs and recognize birds and horses.

Hasselblad says image stabilization performance has also been improved, and it's now rated up to 10EV, which would be a world-first for any camera in any format.

(Image credit: Tim Coleman)

I was using the X2D II with the new 35-100mm lens, and put stabilization to the test. Based on 10EV of stabilization, at the wide-angle end of the lens zoom range I would expect the camera to be able to shoot sharp handheld shots with shutter speeds of 15 seconds.

Not a chance – that simply isn't the case. I was getting sharp shots like I was before with the X2D, down to around three seconds, which is more like 7.5EV. Nonetheless, it's still an incredible performance for such a high-resolution camera.

Despite its improved autofocus performance, the X2D II, like its predecessor, is not a camera for action photography. Its burst shooting performance is limited to around 3fps, which by today's standards is very slow. No, this is a camera for single shots really, like landscapes and portraits.

• Features and performance score: 4.5/5

Hasselblad X2D II 100C: image quality

• Stabilized 100MP 'medium-format' 44 x 33mm sensor
• 16-bit raw files and 'end to end' HDR capture
• No video recording whatsoever

Image quality for the most part is much the same as from the X2D, but Hasselblad didn't need to make any improvements here – the original model was already capable of taking the best-looking pictures I've ever made with a camera.

Not convinced? Check out this Hasselblad X2D vs Nikon Z6 II feature I wrote.

The X2D II utilizes the same 44 x 33mm sensor with 5-axis stabilization as before, shooting 100MP stills with 14-bit or 16-bit color-depth options. What is new, however, is what Hasselblad calls 'end to end' high dynamic range (HDR) capture – you simply check the box in image quality section of the menu to activate it.

HDR, as the name suggests, increases the range of tones in an image. Without it, the X2D II already has a wide dynamic range given the large sensor format and pixel density, but with it we get over 15EV of dynamic range. Check out some HDR landscape photos in the gallery, below.

Image 1 of 7

(Image credit: Tim Coleman)Image 2 of 7

(Image credit: Tim Coleman)Image 3 of 7

(Image credit: Tim Coleman)Image 4 of 7

(Image credit: Tim Coleman)Image 5 of 7

(Image credit: Tim Coleman)Image 6 of 7

(Image credit: Tim Coleman)Image 7 of 7

(Image credit: Tim Coleman)

I took photos with and without HDR active (in raw and JPEG simultaneously), and really I could take or leave HDR – the raw files are stunning either way. Yes, you get more visible detail in shadow and highlight areas, but you also lose some of the punch of the raw files as a result.

And there's plenty of latitude to brighten shadows and darken highlights of raw files for HDR-like results using editing software. Perhaps the main benefit to in-camera HDR is that you can skip the editing process for the times when you want more detail in the tonal extremes. It could also mean that shooting in JPEG-only format is a more realistic option. I still think HDR capture is an excellent feature.

Image 1 of 8

(Image credit: Tim Coleman)Image 2 of 8

(Image credit: Tim Coleman)Image 3 of 8

(Image credit: Tim Coleman)Image 4 of 8

(Image credit: Tim Coleman)Image 5 of 8

(Image credit: Tim Coleman)Image 6 of 8

(Image credit: Tim Coleman)Image 7 of 8

(Image credit: Tim Coleman)Image 8 of 8

(Image credit: Tim Coleman)

If you're a camera owner who's used to a choice of in-camera color profiles, like you get with virtually all other cameras, such as Fujifilm's film simulations, then you'll quickly notice the absence of those here. There's not even a black-and-white mode.

It's another example of Hasselblad's minimalist approach which I appreciate, especially when the natural colors look this good – they're somehow vivid and natural at the same time – so unless you like to dial the saturation slider up to unnatural levels, stylize your images, or go monochrome, you really don't need to spend time tweaking colors in post. All the images in the gallery above are completely unedited.

The impressive image stabilization on board and improved autofocus performance really help too, meaning you can get sharp handheld shots more easily.

Hasselblad has left out video recording once again – the X2D II is purely for photographers. That will no doubt delight photography purists, but I can't see any harm in adding video recording, besides perhaps clogging up the menus.

One thing to note about all of the images included in my galleries – they are original unedited files. With 100MP to play with, there's plenty of cropping potential to focus in on certain details, like the distant person sat on a hillside, framed by the setting sun.

There's also various aspect ratios available in camera, such as Hasselblad's legendary X-Pan panoramic format.

• Image and video quality score: 5/5

(Image credit: Tim Coleman)Hasselblad X2D II 100C: Test scorecardHasselblad X2D II 100C

Attributes

Notes

Rating

Price

A premium camera for a premium price, but it's surprisingly and significantly cheaper than it's predecessor

4 / 5

Design

Simply one of the best camera designs of all time

5 / 5

Feature and performance

Much improved autofocus, built-in SSD storage, but slow burst shooting speeds

4.5 / 5

Image quality

The best looking photos you can get, for the price. No video recording though.

5 / 5

Should I buy the Hasselblad X2D II 100C?

(Image credit: Tim Coleman)Buy it if...

You want natural-looking images from the off
Hasselblad's 100MP, 16-bit Raw files are packed with detail and the stunning colors look right from the off

You appreciate camera design
I dig Hasselblad's minimalist design philosophy – the X2D II not only looks the part, but it feels great in the hand and comes with a refreshingly simply UI.

You're a pro portrait and landscape photographer
If you want the absolute best image quality for portrait and landscape photography and don't often shoot action, then the X2D II sets the bar.

Don't buy it if...

You shoot video
An obvious one, but the X2D II has no video capabilities – it's purely for photographers.

You shoot any kind of action photography
Yes, the X2D II has subject tracking autofocus which makes it more reliable, but it's no match for industry-leading systems, while continuous burst shooting is a slow 3fps approx.

You want supreme battery life
Up to 330 shots is a pretty poor battery life for such a camera. That said, it's a photography-only camera with slow continuous shooting performance, so you're unlikely to rinse the battery quickly.

Hasselblad X2D II 100C: Also consider

(Image credit: Tim Coleman)

Like the sound of the Hasselblad X2D II 100C but keen to know what else is out there? Here are a couple of solid alternatives…

Fujifilm GFX100 II
Fujifilm's flagship camera with the same 44x33mm sensor format is the GFX100 II. It also has a 100MP sensor, but it has superb video recording capabilities too. It's a little pricier than the X2D II.

Read our in-depth Fujifilm GFX100 II review

Sony A7R V
The best full-frame alternative to the X2D II is the high-resolution Sony A7R V. It has the same pixel density – 61MP on a stabilized full-frame sensor, together with a wide dynamic range and impressive autofocus skills. It costs around 40% less than the X2D II too.

Read our in-depth Sony A7R V review

How I tested the Hasselblad X2D II 100C

• Hasselblad loaned me the X2D II 100C for three weeks ahead of its launch
• I had the new XCD 35-100mm f/2.8-4 lens to test it with
• I've taken camera and lens pairing on several outdoor photography shoots

I've had plenty of chances to use the X2D II with the new XCD 35-100mm f/2.8-4 lens ahead of their launches, taking the pairing on a few landscape photography trips, plus for candid portraits of humans and animals.

I've shot during daylight and nighttime, outdoors and in, and paired the camera with the latest (beta test) version of Hasselblad's Phocus app for remote control of the camera and image uploading.

My testing took place over the course of about one month.

First reviewed August 2025

• OneNote for Windows 10 loses support at the same time as the OS
• New OneNote on Windows app replaces it with new features
• Users can update by using the in-app migration ribbon

Microsoft has warned users its OneNote note-taking app is set to lose support for Windows 10 machines from October 14, 2025, and with this landmark comes an important software update.

From that deadline, Microsoft OneNote will be a read-only app, meaning notes can be viewed but not edited or synced.

No further updates, bug fixes or security patches will be provided in line with Windows 10 also losing support on October 14.

OneNote won’t be supported in Windows 10 from October 2025

“As we continue to invest in a more secure, modern, and capable note-taking experience, we’re consolidating our efforts into a single, more powerful OneNote on Windows app,” OneNote Product Manager Bola Soneye wrote in a blog post announcing the news.

By streamlining OneNote, Microsoft is promising faster delivery of new features, long-term support and a “foundation for future innovation in OneNote.”

To move from the classic Windows 10 version of OneNote to the new OneNote on Windows app, users must sync their notes and then use the in-app migration ribbon to ‘Move to the newest version of OneNote’.

The new app can be downloaded from the Microsoft Store for free, but users must sign in with work, school or personal accounts to unlock it.

The blog post also details specific processes for Enterprise and Education customers.

Looking ahead, Microsoft’s roadmap has plenty of entries for OneNote, including a whole host of Copilot-based improvements like summarization, Q&A, voice tools, image and video support and audio overviews.

The OneNote change is just part of a broader shift happening as Windows 10 makes room for Windows 11. Earlier in June 2025, Windows 11 adoption finally caught up with Windows 10 (via Statcounter), before surpassing it one month later.

Soneye concluded by urging OneNote users to migrate now rather than leaving it until the last minute to avoid disruptions.

You might also like

• These are the best productivity tools around today
• Cut costs by using the best free office software on offer
• I can't believe it's taken Microsoft OneNote so long to add this basic feature - and I might actually use it now

On a planet that can feel increasingly challenged, we asked activist Edgard Gouveia Jr. about his latest efforts to improve life on Earth, what "artivism" is — and what he dreams of.

(Image credit: Marlena Waldhausen for NPR)

Annoyingly, The Gilded Age season 4 isn't going to be on our screens for an age (pun intended) yet. In fact, we don't actually have a release window at all, which isn't too surprising given that season 3 only wrapped a few weeks ago. However, the last few episodes told us a lot about what we can expect going forward, and all eyes are on George (Morgan Spector) and Bertha (Carrie Coon).

George’s life hung in the balance in the closing scene of season 3 episode 7. He’d previously won out against Clay (Patrick Page) and Mr. Sage (Peter McRobbie), getting the business back on track in spite of their rivalry. However, this came at a price. A courier soon visited George’s office, pulling a gun on his secretary before aiming it straight at George himself. The gun went off, the screen goes black and we had no idea whether he’s alive or died.

In episode 8, we find out that George was indeed shot, but managed to make a miraculous recovery. But once George is well enough, he promptly tells wife Bertha that he’s not even sure if he loves her or trusts her, leaving for New York without so much as an amicable thought between them. We've currently got no idea if George is even going to return for season 4, but we do know Bertha has to rebuild her life on her own, likely vicariously living through Gladys (Taissa Farmiga).

However, I don't think it's George or Bertha who deserve all the attention in The Gilded Age season 4, or should be the storyline that's going to change 19th century New York from the way that we know it. For me, a new era lies in Gladys herself, and it's her recent marriage and impending pregnancy that will have a lot to answer for.

Gladys' marriage and pregnancy will change The Gilded Age season 4 for good, mark my words

Taissa Farmiga as Gladys in The Gilded Age. (Image credit: HBO)

Gladys announces she is pregnant right in the episode's final moments, with George seemingly out of the picture. At its most basic level, the final moments of season 3 mean Bertha got everything she'd been working towards, but now has been left without a stable family unit of her own. Watching Gladys wed the Duke of Buckingham, and essentially become part of British aristocracy by proxy, was what she always wanted, but it's the consequences we need to be paying attention to.

Firstly, Glady's story is actually based on real life. She's loosely based on Consuelo Vanderbilt, an American socialite and member of the Vanderbilt family who married the 9th Duke of Marlborough during the real Gilded Age. However, theirs became known as a loveless marriage that was 'forced' on her by her mother, with their union thought of as being advantageous to everyone expect for Consuelo herself. Sound familiar?

On top of this, Gladys is who will bring us closest to creator Julian Fellowes' biggest global hit, Downton Abbey. We're still a way off time wise, with The Gilded Age still being the latter stage of the 1800s while Downton Abbey originally began in 1912, but if Gladys stayed with her Duke, had his baby and moved to England, the timelines cross over enough for her to have be middle aged when season 1 of Downton Abbey rolls around. If that doesn't smell like a spinoff, I don't know what does.

Regardless of which direction Gladys' story actually takes, it'll no doubt change the trajectory of The Gilded Age's future – and that's before we even considering how all of this affects Bertha. But who knows whether Gladys will realise she's living Bertha's dream and quit while she's ahead?

"I hope I don't have to play pregnant too long," Taissa Farmiga told Marie Claire. I don't want to have to wear a corset with a pregnancy belt. Let's just throw this out there – maybe Julian reads this. But also I'll do what I have to do!"

All three seasons of The Gilded Age are streaming on HBO Max in the US, NOW TV in the UK and Paramount+ in Australia.

You might also like

• IT: Welcome to Derry – everything you need to know about the new Max horror series
• Harry Potter TV show: everything we know so far about the upcoming HBO adaptation
• The Last of Us season 3: everything we know so far about the hit HBO Max show's next chapter

• Intel has a new 'Gamer Days 2025' promotion running until September 7th
• You get Battlefield 6 for free with certain CPUs or prebuilt PCs
• There are also some chunky discounts to be had with some CPUs

Intel's latest angle to try and shift more of its desktop processors is a promotion tied in with Battlefield 6.

Wccftech reports that Intel has kicked off its 'Gamer Days 2025' campaign – running through until September 7th – which involves some beefy discounts on some of its CPUs (as well as prebuilt PCs containing those processors). However, the big draw for many will be the free copy of Battlefield 6 that's bundled with this offer.

To pick out some examples from the current generation of Arrow Lake CPUs, the flagship Intel Core Ultra 9 285K has been reduced by 12% on Amazon (so in the US it's $530 instead of $600 now), plus that Battlefield 6 freebie. The more mainstream Core Ultra 7 265K processor is the real attraction here, though, with a discount of 36% at Amazon currently (meaning a reduction from an MSRP of $399 down to $259).

This isn't just about Arrow Lake, though, as some previous-generation chips are also reduced. These include the Intel Core i5-14600K which is currently out of stock at Amazon in the US (but is down to $150 at Newegg after a discount code is applied, a seriously tempting proposition at that price level).

In total, there are almost 20 processors involved in this promotion on Amazon, and PC builders such as CyberPowerPC and Origin are in the mix when it comes to the prebuilt rigs included in the offer in the US (or there's the likes of Scan and Overclockers in the UK).

Analysis: heavy hitter of a freebie is needed

(Image credit: EA)

Battlefield 6 represents $70 of value in the US, so if you were intending to buy the game anyway, grabbing one of these CPUs with that outlay knocked off leaves them pretty cheap indeed in some cases.

The price cuts in themselves are nice, but it's the game offer that's doing a lot of the heavy lifting here, as we've already seen these kinds of discounts for Intel chips – even the current-gen models.

Or I should say especially the current-gen Arrow Lake CPUs, as these are rather lackluster in terms of their gaming performance, so Intel needs help getting some sales momentum behind them. That goes for past-gen chips, too, which have the shadow of previous stability issues still hanging over them – consumers aren’t going to forget that episode in a hurry.

You might also like...

• Is Intel set for a great CPU comeback? Nova Lake in 2026 could copy AMD's playbook for a key gaming boost
• What is a processor: Your CPU explained in plain terms
• No, Windows 11 PCs aren't 'up to 2.3x faster' than Windows 10 devices, as Microsoft suggests – here's why that's an outlandish claim

• Last week, the Fitbit app got a redesign with Dark Mode
• The much-requested feature has been a long time coming, as many competitor apps have been using this design for years
• Here's how to switch Dark Mode on and off using Settings

The Fitbit app is undergoing some big changes. To coincide with the launch of the Google Pixel Watch 4 (you can read our early impressions in our Google Pixel Watch 4 hands-on review) it's getting a personal AI health coach in the US and, as far as we're aware, UI changes as well.

However, before those changes come into effect, Google has given the Fitbit app a significant facelift already, with the launch of Dark Mode.

The Fitbit app, since its inception, has always been set against a bright off-white backdrop regardless whether the rest of your phone is in Dark Mode or not.

It's resisted change even though competitors for the crown of best fitness app, such as Apple Health and Garmin Connect, have been on dark backgrounds to make parsing through complex graphs and planning workouts easier on the eyes.

(Image credit: Garmin/Shutterstock)

In my opinion, it's crazy that despite the popularity of the best Fitbits, it's taken so long for the companion app to get a Dark Mode. It's a simple inversion that makes the experience of using the app so much better for most people.

However, if I was being completely cynical, it might have taken so long because Google simply didn't know what to do with Fitbit.

I've written plenty about Google's neglect of the brand while folding the best hardware features into its Pixel Watch series, but it seems as though after last year's app redesign, Dark Mode and this year's heavy investment into the AI health coach, Google's finally seeing a way for Fitbit to exist within Google's complex ecosystem going forward.

When I opened my Fitbit app this morning, Dark Mode was already enabled. However, in case yours hasn't switched over automatically or you're looking for manual adjustment, here's how to toggle Dark Mode on and off.

(Image credit: Future)

• Ensure your Fitbit app is updated to version 4.50. If not, navigate to Software Updates in your phone's settings
• In the Fitbit app, tap your profile image and go to Fitbit Settings
• In Settings, tap the new Theme option
• You can choose between System Default, Light or Dark options
• System Default will match Fitbit to your phone's theme, so if you use Dark Mode on your phone's operating system, Fitbit will switch automatically

You might also like...

• Google didn't show its AI health coach in action – here are 5 features I hope we'll see when it drops
• Google’s Pixel Watch 4 just beat Apple to a lifesaving satellite trick – here’s how it works
• As a smartwatch expert, I'm hoping the Pixel Watch 4's best feature is the start of a bold new era in wearable tech

• Researchers find 9.3/10 flaw in Docker Desktop for Windows and macOS
• The bug allows threat actors to compromise underlying hosts and tamper with data
• A fix was quickly released, so users should patch now

Docker has patched a critical severity vulnerability in its Desktop app for Windows and macOS which could have allowed threat actors to fully take over vulnerable hosts, exfiltrate sensitive data, and more.

The vulnerability is described as a server-side request forgery (SSRF) and, according to the NVD, it “allows local running Linux containers to access the Docker Engine API via the configured Docker subnet.”

“A malicious container running on Docker Desktop could access the Docker Engine and launch additional containers without requiring the Docker socket to be mounted,” Docker said in a follow-up security advisory. “This could allow unauthorized access to user files on the host system. Enhanced Container Isolation (ECI) does not mitigate this vulnerability.”

Not all systems are affected in the same way

The bug was discovered and reported by security researcher Felix Boulet. It is now tracked as CVE-2025-9074 and was given a severity rating of 9.3/10 (critical).

However, a separate researcher, Philippe Dugre, stressed that the risk is not the same on all platforms, noting it’s actually somewhat greater on Windows, compared to macOS.

This is due to the safeguards baked into the macOS operating system. Dugre managed to create a file in the user’s home directory on Windows, but not on macOS:

"On Windows, since the Docker Engine runs via WSL2, the attacker can mount as an administrator the entire filesystem, read any sensitive file, and ultimately overwrite a system DLL to escalate the attacker to administrator of the host system," Dugre explained.

"On MacOS, however, the Docker Desktop application still has a layer of isolation and trying to mount a user directory prompts the user for permission. By default, the docker application does not have access to the rest of the filesystem and does not run with administrative privileges, so the host is a lot safer than in the Windows case," he added.

Docker fixed it in Desktop version 4.44.3, so users are advised to upgrade as soon as possible.

Via BleepingComputer

You might also like

• Docker could still be hosting a whole load of potentially malicious images - putting users at risk
• Take a look at our guide to the best authenticator app
• We've rounded up the best password managers

Kilmar Abrego Garcia has been taken into custody and faces deportation again, this time to Uganda. And, Trump seeks to fire a member of the Board of Governors of the Federal Reserve.

(Image credit: Anna Moneymaker)

In a new proposed class action lawsuit against Amazon [via The Hollywood Reporter], the company has been accused of “bait and switch" (a type of fraudulent activity) on Prime Video. This means they are allegedly misleading consumers into thinking they’ve purchased content when they’re only getting a license to watch, which can be revoked at any time if Amazon loses the rights to a title.

The potential lawsuit claims: "violations of California unfair competition, false advertising and consumer legal remedies laws. It seeks unspecified damages, including disgorgement of profits and punitive damages for allegedly intentionally malicious conduct." Filed in Washington federal court on August 22 2025, the proposal claims that Amazon is "misrepresenting the nature of movie and TV transactions during the purchase process".

In case that's not clear, let's take the unwatched digital version of Conclave I 'bought' on Prime Video so my parents didn't miss out on this year's Oscar hype as an example. I have a digital copy, but if Prime Video's licensing agreements were to change, so too could the version of Conclave I have access to. If Amazon were to no longer have the rights to the title, my parents would lose the movie.

As the complaint points out, “you receive a license to the video and you agree to our terms," meaning that what you actually get for parting with your money is written in the small print. But should Prime Video be allowed to tell subscribers that they've "bought" a movie, and what does this mean for us users in the long run?

Prime Video’s new class action lawsuit proves we need to invest in more physical media

Prime Video has a huge back catalog, but are we really buying it? (Image credit: Amazon)

Before we go any further, let's not forget that this isn't the first lawsuit of this kind Prime Video has had against it. In 2020, a separate lawsuit alleged "unfair competition and false advertising over the practice". While Amazon has not yet publicly commented on the new class action, it claimed in 2020 that using the word "buy" isn't deceptive to subscribers because consumers already understand that their purchases is subjective to license agreements. Five years later, and I'd say that likely isn't the case.

Back in 2023, a Californian legislature brought the problem to the forefront again. Gamers found that their access to The Crew would be stopped after Ubisoft shut down the game's servers, inspiring the 'Stop Killing Games' movement that took aim at publishers destroying previously-bought consumer titles.

However, it's changes to Californian legislature this year work to our new lawsuit's advantage. Essentially, a state law has barred the use of the word 'purchase' in a transaction unless "it offers unrestricted ownership of the product." Obviously, our Prime Video small print doesn't fit into this, and Amazon can hardly afford to lose such a huge profit share as California (if it was its own separate country, California would be the fourth largest economy in the world).

We don't yet know what any of this means for streamers with a Prime Video subscription on a wider level, but to me, it's an incredibly stark reminder that we need to keep investing in physical media as much as possible. Yes, it's more expensive than paying a flat fee every month for all the content you can possibly want. But it's like dating: if you become more intentional in what you invest in, the results are lifelong.

If you have physical copies of movies and TV shows that you love, you can never be parted from them, and it's the only way we can now guarantee the security of what we buy. Maybe it's time for the best streaming services to revert to the good old days of sending us discs in the post to watch and return when we're done with them, just like Netflix did in the late 2000s.

You might also like

• I can’t wait for Fallout season 2 to dive into New Vegas – here are 6 things I’m desperate to see when the Prime Video show returns
• Conrad’s bachelorette speech is exactly like the books in The Summer I Turned Pretty season 3 episode 7, but it still leaves a sour taste
• 007: Road to a Million season 2 is better than any new James Bond idea Amazon could ever come up with

Walk into most organizations today and ask what they're spending on SaaS. Odds are, no one can give you a confident answer. Not because they don't want to, it's because no one actually knows.

Ask a different question: who owns SaaS spend in your company? You'll likely hear three things: "Finance handles it," "That's IT's job," or "Honestly, it depends.”

And therein lies the real problem. While companies are dropping anywhere from $9,000 to $17,000 per employee annually on software, most organizations have zero clue what they're actually buying.

The explosion of software tools across every function, only exacerbated by AI, has quietly created a gap between what companies think they're managing and what they're actually managing. And that gap is getting more expensive by the month.

SaaS sprawl is worse than you think

Here's how it happens: your marketing team signs up for Canva Pro, your sales team gets Calendly, design jumps on Figma, and engineering grabs another GitHub license. Meanwhile, IT is already paying for Adobe Creative Suite, Microsoft has calendar functionality, you've got design tools in your existing stack, and there's a company-wide GitHub Enterprise account sitting unused.

This isn't just wasteful spending. It's what we call SaaS sprawl, and it's quietly bleeding companies dry. Recent data shows organizations use an average of 112 SaaS applications, with large enterprises using up to 447 different tools. And I think this is actually underrepresented. When every department acts like its own startup, you end up with a technology Frankenstein that nobody can control or understand.

When you factor in that companies waste 30-50% of their SaaS budgets on unused licenses, and missed renewal dates can cost upwards of $200,000 per instance, it’s hard to understand why so many are not addressing this problem head on. When there's no centralized intake or contract visibility, things slip through. You renew tools no one's using. You pay above market rates because you don't benchmark. You get hit with surprise auto-renewals.

The AI acceleration problem

And, just when some companies thought they had SaaS sprawl under control, AI came along and hit the gas pedal. We're seeing the late 2010s SaaS explosion all over again, but this time it's powered by artificial intelligence.

We’re in the middle of a perfect storm. Leadership wants teams to be AI-enabled, to experiment, to learn. They're actively encouraging employees to test new tools and find ways to work more efficiently. Meanwhile, IT teams are desperately trying to control the sprawl that's already spiraling out of control.

Guess who wins? The credit card.

Employees are swiping corporate cards to try the latest AI writing tool, testing out OpenAI subscriptions, or spinning up Zapier automations without any security review or budget coordination. Each purchase seems small and reasonable. A $20 monthly subscription here, a $50 annual plan there. But multiply that across every department, every team, every curious employee, and you've got a massive problem.

The conflicting stories are everywhere. Leaders preach innovation and experimentation while finance teams watch budgets explode. IT departments create approval processes while employees find workarounds. Everyone wants to be AI-first, but nobody wants to be the one who says no to the next breakthrough tool.

Shadow IT: The innovation myth

Here's where things get interesting. Some people claim Shadow IT and now Shadow AI drives innovation. They're wrong. Anyone claiming Shadow IT drives innovation isn't actually fostering an innovative environment.

When 40% of IT spending happens outside formal oversight, that's not innovation. That's broken processes. Your procurement workflows are failing to meet company needs quickly enough, so people are going rogue.

Sure, it looks like innovation on the surface. Employees find new tools, solve problems quickly, and move fast. But here's what's really happening: you're diverting time, money, and focus from actual innovation and R&D investments that could drive the company forward.

Real innovation happens when teams can explore new ideas without bypassing controls. If the only way to get work done is to go around IT or procurement, that's not agility, it's dysfunction. And it's expensive.

The security nightmare we’re all ignoring

It’s not just pure budget that is the problem, Shadow IT and AI and SaaS Sprawl are all creating security holes that many are simply not addressing. Every unauthorized app is a potential entry point for bad actors. IBM found that one in three data breaches involved Shadow IT, with the average breach costing around $4.9 million.

When someone in engineering or marketing signs up for a random productivity tool using their work email, they're potentially exposing company data. No security review, no IT approval, no encryption standards. Just click, sign up, and hope for the best.

The compliance risks are equally terrifying. Use a non-GDPR-compliant tool for EU customer data? That's a potential fine. Healthcare company using a random file-sharing app? Hello, HIPAA violations. These types of risks are happening right now at companies that think they have things under control.

Where sprawl lives

Interestingly, SaaS sprawl doesn't always come from obscure tools. It often comes from the biggest names in tech. At Tropic, we’ve found that some of the most common drivers of tool overlap and Shadow IT include:

• Zoom, Microsoft, Slack, Google – Multiple collaboration tools per organization
• Figma, Canva, Adobe – Design tool overlap with no license governance
• Salesforce, Calendly, DocuSign – Sales tools stacked on top of each other
• GitHub, JetBrains, Atlassian – Dev tools used inconsistently across teams
• Dropbox, Apple, Amazon, OpenAI – Personal subscriptions tied to work email

No one sets out to buy the same tool twice. But without visibility, it happens all the time. Every new vendor means more contracts to track, more renewals to manage, more security reviews to conduct, and more relationships to maintain. The administrative overhead alone can eat up significant resources.

When spreadsheets become expensive

A lot of finance and IT teams are still trying to manage all this complexity with spreadsheets. That's like trying to navigate a modern city with a paper map from 1995. Even a 1% error rate on $50 million of spend can waste $500,000 annually.

Dig deeper and this isn’t just a tooling issue, it's an ownership issue. Procurement or finance thinks IT is managing it. IT assumes finance has the numbers. Finance is tracking spend, but not usage. Legal might only get involved post-signature. So, things fall through the cracks.

Let's talk ROI

Here's something most people don't talk about enough: every dollar saved on procurement and purchasing has an immediate impact on the bottom line. Unlike new sales revenue, a dollar saved can be pure profit.

Reducing SaaS spend by just 6% delivers the same profit lift as a 20% increase in top-line revenue. And that's before you factor in the benefits of reduced risk, stronger compliance, and faster purchasing cycles.

We've seen companies recover hundreds of thousands—sometimes millions—just by tackling renewals earlier, consolidating tools, and validating usage.

What smart companies are doing instead

The fix isn't shutting down software purchases. Not only is that impossible, but you’d have a disgruntled workforce on your hands. It is, however, about enabling them with structure. The companies that are winning aren't locking down every software request. They're treating software spend like the strategic lever it is.

Here's what best-in-class companies are doing:

• Centralizing intake. Giving teams one place to request or renew software.
• Building a software inventory. Not just contracts, but owners, usage, and cost.
• Reviewing renewals 90–180 days out. Not two weeks before expiration. Get ahead of things to determine if you need other tools and create savings.
• Using benchmarking data. So, you don't overpay for tools that should cost less.
• Measuring utilization. If you bought 500 seats and only used 320, ask why.

None of this slows people down. In fact, it makes it easier for teams to get what they need, faster because the path is clear, the data is ready, and approvals don't sit in a black hole.

The time to act

Every month you wait is money walking out the door. Those auto-renewals are happening whether you're paying attention or not. The unused licenses are accumulating. The security risks are multiplying.

But don’t fear. You don't need to solve everything at once. Start with visibility. Figure out what you're actually buying. Identify the obvious waste. Cancel the subscriptions nobody is using.

Software isn't slowing down. And with AI in the mix, things are only getting more complex. This is your moment to get control, not by over-regulating, but by creating the visibility and structure your teams need to move fast, spend wisely, and innovate securely.

Your choice is simple: act now, or pay later. The meter is running either way. You don't need 200 tools to move fast. You need the right 20 and a way to manage them well.

We've featured the best business plan software.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro

Arbuscular mycorrhizal fungi build efficient transportation networks underground to connect to plants that they trade with.'/>

Fungi and plants have something to teach humans about global trade and cooperation

(Image credit: Loreto Oyarte Gálvez)

Levoit CirculAir Oscillating Fan: two-minute reviewProduct info

For this review, I tested the UK version of the Levoit CirculAir. A version with the same product code is available in the US, but there are minor differences compared to the UK model, including fewer fan speeds and slightly different onboard controls.

The Levoit CirculAir Oscillating Fan, also known as the Levoit Pedestal Air Circulator Fan in the UK, might be the best fan I've tested. I’ve been seriously impressed with the strong performance, convenience, and quiet operation.

This fan can be configured in two ways: either at full height or shorter, by simply unscrewing a section of the stand. This lower height is a little bit tall for use on a tabletop, but perfect for circulating air from heaters or floor-standing air conditioners.

(Image credit: Future)

It offers automatic horizontal and vertical oscillation, which makes for great air circulation, and also makes it really easy to fine-tune the position of the airflow just using the remote. The only thing it’s missing here is the option to select different oscillation ranges; instead, it's limited to 90 degrees horizontally and 120 degrees vertically.

On the front of the stand there’s a well-lit LED screen that displays the selected mode, and either the fan speed or the ambient temperature. Below this are the onboard touch controls (which vary slightly depending on region), and the CirculAir can also be controlled with the easy-to-use circular remote, which is magnetic for convenient storage on the back of the air circulator.

(Image credit: Future)

The airflow proved impressively strong during testing, and I was very grateful for it during a recent heatwave. Even at its lowest speed, I could feel a good breeze from 8ft / 2.4m, and I ran out of space to back up any further at 18ft / 5.5m away when the fan was at its top speed of 12, so it’s clear it has super-strong airflow when running full blast.

Despite its strength, the CirculAir runs satisfyingly quietly. I measured a volume of just 29dB from around 2ft away on fan speed one, 37dB at speed five, and 51dB at speed 12. This means it can operate with whisper-quiet volume at its lowest speed, and only gets as loud as moderate rainfall.

All this for a not-too-painful price tag: this fan is available to purchase at Amazon at a list price of $89.99 / £119.99, but I've already spotted a few discounts, so you might be able to avoid full whack. That's the short version; read on for my full Levoit CirculAir Oscillating Fan review.

Levoit CirculAir Oscillating Fan review: price & availability

• List price: $89.99 / £119.99
• Availability: US and UK

The Levoit CirculAir Oscillating Fan, AKA the Levoit Pedestal Air Circulator Fan in the UK, has a list price of $89.99 / £119.99, and is available at VeSync in the US, and at Amazon in both regions.

This is a good price considering the impressive performance the Levoit CirculAir delivered during testing, but I recommend keeping an eye out for offers, too, as there appear to be fairly regular deals available at Amazon. For example, at the time of writing, there’s a 10% limited-time deal knocking the price down to $80.99 in the US, and the price was regularly dropping to below £95 in the UK a couple of months ago.

I’ve tested the UK model, and while the model number is the same in the US, there are some slight differences, which I make note of in the following sections.

• Value for money score: 5 out of 5

Levoit CirculAir Oscillating Fan review: specs

Type:

Pedestal air circulator

Speeds:

9 (US) / 12 (UK)

Oscillation:

Automatic: 90 degrees horizontal, 120 degrees vertical

Dimensions:

43.7 x 12.5 x 14.5 in / ‎36.5 x 110.8 x 32 cm

Weight:

12.7lb / 5.8kg

Control:

Onboard buttons and remote control

Timer:

Yes

Additional modes:

Eco, Turbo, Sleep

Levoit CirculAir Oscillating Fan review: design & features

• Automatic vertical and horizontal oscillation
• 2 height options and 4 fan modes to choose between
• Onboard touch controls and remote control

The Levoit CirculAir is a pedestal air circulator with two height options, so it can be placed on the floor or a tabletop. I personally found its shorter height to be a little taller than I’d prefer for a desktop fan, but this is a good option to have if you happen to be short on floor space. This shortened height could also prove useful if you wanted to, for example, position the fan in front of a heater or floor-standing AC unit, as it would be more efficient at circulating the heated or cooled air from that position.

I found it satisfyingly easy to get the Levoit CirculAir setup in either configuration – it was a simple case of screwing the poles together and into the base. The base is just the right weight: heavy enough that the fan felt sturdy on my medium-pile living room carpet, but not too heavy to move it around when required.

(Image credit: Future)

There are two ways to control the Levoit CirculAir: by using the circular remote or the large touch panel on the front of the stand. The remote was just the right size; small enough to be easy for me to hold and reach all the buttons, but large enough that the buttons were well-spaced and the icons were easy to read.

The large LED panel has a generous screen that displays either the fan speed or ambient temperature, as well as which mode the CirculAir is currently set to. The controls vary slightly on the UK model from the one in the US, but all controls are accessible on the remote in both regions.

(Image credit: Future)

The number of fan speeds also varies slightly between regions, curiously, with 12 fan speeds on offer on the UK model I tested, but only nine in the US. There are a few different modes that can be cycled through: Normal mode (where I could choose the fan speed), Turbo mode (which amps up the fan), Sleep mode (which silences the responsive beeps and dims the display), and Eco mode (which makes use of the CirculAir's internal temperature sensor and adjusts the fan speed to suit).

Unlike most desktop fans, the Levoit CirculAir boasts automatic vertical oscillation, along with the standard horizontal oscillation. There isn’t the option to choose the degree of oscillation, sadly, so it’s either 90-degree horizontal oscillation and 120-degree vertical oscillation, or nothing. I did have the choice to use just one or the other, or both, though. Not only does this range of oscillation equal better air circulation, but it also means that it could be set to just the right angle from the comfort of my couch using the remote, which I loved.

• Design score: 4.5 out of 5

Levoit CirculAir Oscillating Fan review: performance

• Easy to use, responsive controls
• Super strong airflow
• Satisfyingly quiet operation

Controlling the CirculAir was straightforward, as the touch buttons on the well-lit display were simple to understand and responsive, as were the controls on the remote. I really liked that the remote control could be magnetically clipped onto the back of the air circulator, as this made it really easy to store and retrieve.

The Levoit CirculAir has delightfully strong airflow, and I found it to be a real help cooling me down during a recent heatwave here in the UK. In fact, it has the strongest airflow out of the fans I’ve tested to date, to the point where I wasn't able to confirm how far the air travelled when it was running at its top speed, as I ran out of room at 18ft / 5.5 meters.

The airflow strength was great at lower speeds, too, as I could still feel a pleasantly cooling breeze from an impressive 8ft / 2.4m away at speed 1, and 15ft / 4.5m at speed 6, which means it reaches further at half power than the Dyson Purifier Cool Formaldehyde TP09 does at full blast.

(Image credit: Future)

I love it when a fan has an automatic mode, and 'Eco mode' (as it's called in the CirculAir's case) came in particularly handy during the recent heatwave. The CirculAir adjusted the fan speed in response to the rising ambient temperature throughout the day, though I found it to be much more conservative in terms of the fan speeds it applied compared to my current favorite air circulator, the MeacoFan Sefte 8in. This is good news for energy saving, though, and doesn’t make it any less efficient, really, considering the concentration of air kicked out even at the lower fan speeds.

The volume levels were pleasingly quiet, especially when taking into account the CirculAir’s punchy airflow. I got a reading of just 29dB from about 2ft away on fan speed 1, 37dB at speed 5, and 51dB at speed 12, so it was whisper quiet at its lowest, and only as loud as moderate rainfall when running at top speed.

(Image credit: Future)

Its low volume meant I had no problem at all sleeping with it toward the foot of my bed. I really appreciated the automatic vertical and horizontal oscillation at nighttime, too, as being able to control both with the remote meant I could get the angle of the airflow exactly where I wanted it without needing to get up and manually adjust it.

After testing, I have no complaints about the Levoit CirculAir. It was easy to use and responsive, and it operated at low volumes while delivering impressively strong airflow.

• Performance score: 5 out of 5

Should I buy the Levoit CirculAir Oscillating Fan?

Section

Notes

Score

Value for money

This air circulator’s fantastic performance and versatility make it great value for money, but it’s always worth keeping an eye out for discounts.

5/5

Design

With 2 height options, automatic horizontal and vertical testing, plus responsive controls with clever remote storage, there’s plenty of versatility.

4.5/5

Performance

The airflow is seriously strong, and yet the CirculAir runs pleasingly quietly. The automatic oscillation proved particularly convenient.

5/5

Buy it if...

You want strong airflow
This air circulator certainly packs a punch! With impressively strong airflow even at its lower speeds, verging on ridiculous when running at full blast.

You want automatic cooling
The CirculAir’s Eco mode is a handy feature to have when the ambient temperature fluctuates, as it’ll adjust the fan speed to suit.

You want a fan with adjustable height
Having two different height options is great, whether you want to move it from the floor to a table, or you want to get the most efficient air circulation, depending on the position of your heater or air conditioner.

Don't buy it if...

You want something smart
If you love your smart tech and prefer to operate things from your phone, then this won’t be the air circulator for you.

You’re looking for a desk fan
Even at its shortened height, the CirculAir is a bit too tall, and the airflow is likely too strong to be a personal fan.

You want to fine-tune the oscillation
Unlike fans like the Dyson Cool CF1, which offers angles of 15 to 70 degrees, the Circulair doesn’t have the option to fine-tune the range of the oscillation.

Levoit CirculAir Oscillating Fan review: Also consider

Levoit CirculAir Oscillating Fan

Shark FlexBreeze

Meaco MeacoFan 1056P

Price

$89.99 / £119.99

$199.99 / £199.99

$189.99 / £149.99

Speed settings

9 US / 12 UK

5

12

Oscillation

Automatic: 90 degrees horizontal, 120 degrees vertical

Automatic:180 degrees horizontal

Manual: 55 degrees vertical

Automatic: 60 degrees horizontal, 80 degrees vertical

Timer

Yes

Yes

Yes

Controls

Touch buttons, remote control

Buttons, remote control

Buttons, remote control

Power

Corded

Corded or cordless

Corded

Additional modes

Eco, Turbo, Sleep

Boost, Mist

Sleep, Natural, Eco

Height adjustable

Yes

Yes

Yes

Dimensions

43.7 x 12.5 x 14.5 in / ‎36.5 x 110.8 x 32 cm

37 x 13.8 x 13.8in / 94 x 36 x 36cm

32.3-37.6 x 12.6 x 13.4 in / 82-95.5 x 32 x 33.9cm

Weight

12.7lb / 5.8kg

12.2 lb / 5.7kg

12lb / 5.4kg

Shark FlexBreeze
This is a great option if you like the idea of taking your fan outside and making use of its misting attachment on sunny days. It can be used as a pedestal or table fan, plus it can run cordless, and our reviewer was pleased with the battery life. It only has five fan speeds, but it still delivers airflow that can be felt up to 70 feet away. If you’d like to learn more, check out our full Shark FlexBreeze review.

Meaco MeacoFan 1056P
Another impressively quiet fan, this height-adjustable air circulator proved a hit with our reviewer. It delivered a great performance when tested during hot weather, and proved pleasingly quiet, even at higher speeds. This is another fan with clever fan storage, too, as the circular remote magnetically clips to the front. To find out more, have a read of our full Meaco MeacoFan 1056P review.

How I tested the Levoit CirculAir Oscillating Fan

• I tested the Levoit CirculAir Oscillating Fan for a week
• I tried out the different modes and features
• I tested its volume and how strong the airflow was

I used the Levoit CirculAir Oscillating Fan in our photo studio, my home office, and my bedroom over the course of a week. I evaluated the assembly process and build quality closely before using it passively and in specific tests that allowed me to assess the available features.

I tried the CirculAir at both heights, oscillating, and static. I tested out the different modes, particularly Eco mode, to see how the CirculAir would respond to fluctuating temperatures.

I assessed the airflow by checking the distance I could still feel a cooling breeze on the top, middle, and lowest settings. I used a decibel meter on my iPhone to check the noise levels, taking the readings from around 2ft / 600mm away, ensuring the fan wasn’t blowing directly into the microphone.

• First reviewed: August 2025
• Read TechRadar's Reviews Guarantee

You might also like

• 'You get the best of both worlds' – an expert reveals why an air circulator is better than a regular fan this summer
• Pedestal fan vs tower fan: what's the difference?
• Air circulator vs fan: what's the difference, and which one should you buy?

With data becoming a more valuable business asset than ever, informing decision-making, enhancing operational efficiency, and enabling businesses to gain a competitive edge, hyper-personalization is transforming how businesses interact with customers. Taking data analysis to a new frontier through the deployment of AI at scale to identify hidden consumer patterns and preferences, hyper-personalization is fast becoming the new standard for businesses seeking to attract and retain customers.

However, with increased use of data comes more challenges and responsibilities. As more people become more digital savvy, they also become more conscious of their data and how it is used. This presents a necessary challenge for businesses: how to offer high-quality but ethical personalization offerings.

How hyper-personalization is delivering breakthrough value amid rapidly shifting market demands

Tailored offerings boost engagement, and foster customer loyalty, and as a result, hyper-personalization strategies are growing exponentially because of their high-value returns.

According to IBM, effective personalization programs can reduce customer acquisition costs by up to 50% as machine learning and advanced analytics can predict customer preferences and automate decision-making, with data and analytics turning raw information into meaningful insights that drive measurable performance.

Advanced AI systems further enhance these capabilities, supporting sophisticated personalization at scale and enabling continuous adaptation as new data becomes available.

These enhancements increase agility and business dynamism – a necessity for businesses operating in today's economic and geopolitical landscape. Hyper-personalization enables companies to make faster, smarter business decisions that align with rapidly changing market conditions.

By leveraging real-time data and advanced analytics, organizations can quickly identify and respond to emerging trends, customer preferences, and competitive threats, allowing them to adapt their strategies and operations in near real-time.

How businesses are approaching hyper-personalization

Having undergone a wide adoption, personalization strategies are offering businesses the ability to reshape everything from customer experience to product development, proving personalization to be one of the most strategic use cases of artificial intelligence.

Across industries, leading organizations are seizing the power of AI and advanced analytics and automation to deliver tailored experiences to remain competitive. Examples include product recommendations on online stores and streaming services based on previous search history. AI is also being utilized to humanize customer interactions in insurance, and tailoring treatment plans unique to patient biology in healthcare.

Tech native platforms, which have long leveraged the advantages of personalization, are accustomed to these strategies, but businesses across traditional industries – healthcare, manufacturing, retail, automotive – are increasingly investing in personalization to keep pace.

The legacy systems that defined these industries for decades are being succeeded by AI powered, data-integrated solutions – a testament to the growing recognition that these technologies provide actionable insights.

How businesses can navigate the complexities and opportunities of AI adoption at scale

Given the increasing circulation and use of data, as well as the commercial imperative to leverage detailed customer preference data, the scale up of hyper-personalization strategies is a complex process, calling for robust AI regulation and data privacy frameworks.

Strong data governance is inherent to a sustainable hyper-personalization scale-up, especially when attempting to elevate the AI profile within a business. This includes establishing clear policies on data collection, usage, and retention, as well as ensuring compliance with evolving privacy regulations such as GDPR, and implanting robust cybersecurity systems which mitigate data breaches.

Workforce transformation is also a critical consideration. There is a need to upskill and reframe workforce training to foster a culture of innovation that works in tandem with AI additions. Traditional sales and marketing roles are evolving and fundamentally changing at a rapid pace compared to operations that were commonplace only a decade ago. Now, there is greater emphasis placed on data analysis, ethical AI model development, and AI literacy.

Managing risk in AI adoption also emerges as intrinsic to a hyper-personalization scale-up. Adopting AI systems is only a start, but managing risk becomes the next focus, to mitigate algorithmic bias and false results. This involves regular auditing of AI models, monitoring for unintended consequences, and embedding ethical considerations into the AI lifecycle.

Businesses are operating on new terrain. Hyper-personalization has made reaching customers easier than ever – the real challenge now is how best to employ these tools to anticipate the needs of customers before they know themselves. There is a fine balance to strike here – companies need to invest money and manpower into the ethical growth of their AI and data strategies, or they risk eroding consumer trust.

We've featured the best AI chatbot for business.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro

The rise of AI is reshaping business technology at an unprecedented pace. From IT to HR, finance to customer service, few departments remain untouched by the wave of automation and intelligence sweeping companies and industries today. However, with this surge in interest comes a growing challenge: distinguishing between truly transformative AI tools and those merely dressed up in buzzwords.

For CIOs and business leaders, the mandate has shifted from exploration to execution. Deploying the wrong AI solution doesn’t just stall progress; it burns time, budget, and internal credibility. The challenge now is clear: cut through the noise, ensure enterprise-grade security, and back only the AI that drives measurable impact.

Perception problems around AI

At a surface level, many AI solutions look the same: slick interfaces, automated responses, bold claims. But there is a distinct difference between basic AI bots and true agentic AI. Some AI products automate tasks only within rigid boundaries, while agentic AI is designed to think, act, and adapt with no intervention required.

The confusion often stems from how AI is marketed. Some platforms tout predictive insights but rely on limited or shallow data, resulting in misleading outputs. Others claim “full autonomy,” yet still depend heavily on human input. Most are wrappers for outdated automation, only a few are truly built to drive action, solve real problems, and evolve with your environment.

Similarly, many products only scratch the surface by simply passing user prompts to large language models (LLMs) through an API - what you might call a very thin layer of AI. They look impressive at first, but lack any meaningful depth.

This creates a perception problem. AI is either seen as a cure-all or dismissed as hype. In reality, the value lies between. Real improvements in productivity and efficiency come from using the right tools, not just any tools.

The shift from automation to autonomy

Although hype still surrounds AI, we’re also seeing real progress as it evolves from basic automation to true autonomy. In IT specifically, autonomous AI is starting to take on entire workflows from start to finish, including resolving low-level support tickets without any intervention from IT personnel, even though end users may still interact with the AI.

The depth of these solutions is critical. When AI systems layer orchestration, coordinate multiple processes, or use specialized agents for different tasks, they become much more than a simple interface to a language model. And when they can take informed action on real business systems, drawing on an organization’s unique data and historical context rather than merely offering recommendations, that’s when you see what can truly be considered a deep AI product.

The effect on an organization is threefold. For end users, it delivers a zero-time SLA experience: instant support, self-service resolution, and frictionless access to help anytime. This shift dramatically improves the digital employee experience (DEX), which is now a key driver of productivity and satisfaction in mature IT environments. For IT teams, it frees up hours each week, reduces backlog, and improves response times. For the organization, it cuts costs without compromising quality and enables scalable IT support without additional hiring. However, with this power comes responsibility. IT leaders must ensure these systems operate within clear guardrails, especially when interacting with sensitive data, employee devices, or live environments.

A central concept here is closed-loop AI. These systems are designed to ensure that inputs remain within the organization’s control. Unlike open models that may use your data to enhance results elsewhere, closed-loop systems are built with enterprise-grade governance in mind. This approach gives IT leaders greater confidence to adopt AI without compromising security or compliance.

Three warning signs of hype

To effectively evaluate AI tools, it’s important to look past the branding and focus on the core mechanics. Here are three common red flags:

Lack of specificity: If a product claims to “revolutionize business” but cannot point to a specific workflow or use case it improves, that is a concern.

No explainability: If you can’t trace how a decision was made, or what data was used to make it, that’s a sign of a black-box system. Trustworthy AI should be auditable and understandable, especially in high-stakes enterprise settings.

No real learning or depth: If the AI lacks any meaningful learning mechanism or only relies on a small, shallow set of data points, it’s unlikely to improve over time. True AI products get smarter by processing large, relevant datasets, whether through training robust models or continuously absorbing business context. Without this depth, you’re often looking at a thin layer that may impress in a demo but quickly fall short in the real world.

As more tools claim to offer autonomy, it’s more important than ever to understand what to look for in a reliable AI solution and what to avoid.

What to look for instead

Instead of getting distracted by flashy demos or inflated claims, decision-makers should evaluate AI tools based on three key pillars:

Relevance and integration: Is it trained on data that reflects your business context, and can it be customized to fit your company’s workflows, policies, and operational guidelines? Just as important, will it integrate with your existing tech stack or require major reengineering? AI works best when it adapts to how your organization already operates, not the other way around.

Transparency: Can you understand and control how it works?

Impact: Does it save time, reduce errors, or improve outcomes in measurable ways? Does it actually do the work? Are there any stats or data points that can show proven impact?

Ultimately, the strongest AI solutions build layers of capability, from orchestration to specialized agents to learning engines that can take real action, creating something far more valuable than tools that simply pass prompts to a language model. They don’t just mimic intelligence; they deliver tangible value by empowering teams to focus on strategic work, improving efficiency, and generating a clearly demonstrable return on investment.

The future Is functional, not flashy

The future of AI in enterprise technology will not be defined by the tools with the boldest announcements or the most dramatic demos. Instead, it will be shaped by smart, adaptable systems that take ownership of tasks from start to finish and operate independently within clearly defined parameters. These tools quietly improve everyday operations and deliver consistent results with minimal oversight.

AI on its own is no longer enough. To truly deliver value, it needs to be connected to real-time systems, historical data, and the operational context where it’s deployed. That’s what unlocks its full potential. When AI is paired with an on-the-ground agent and backed by rich historical insights, it can go beyond recommendations and solve problems autonomously. It’s the combination of real-time visibility, institutional memory, and intelligent execution that makes for a truly transformative solution.

For IT leaders, the goal is not to chase hype, but to make informed decisions by asking tough questions, demanding clarity from vendors, and staying focused on business outcomes.

We've featured the best AI website builder.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro

Cambridge Evo 150 SE: Two-minute review

A little over four years after it launched its well-regarded Evo 150 music streaming amplifier, Cambridge has given it the once-over and rechristened it Evo 150 SE. The differences are not numerous, but they are noteworthy nonetheless – and that’s even before you take into account the fact that this new model is actually less expensive than the model it replaces.

A neat form factor, with swappable side panels a particularly nice touch, a big, bright display and several very decent control options are all carried over. So is the expansive selection of digital and analogue, wired and wireless input options. Power, at 150 watts per channel into 8 ohms, is unchanged too – but now the Hypex NCOREx Class D amplification is ‘tuned by Cambridge’. This is by far the most significant differential between the 150 SE and the outgoing 150.

It makes its presence felt, too, in the weighty, punchy (and ultimately fractionally overstated) low frequency reproduction. But it hasn’t affected the machine’s ability to create a big soundstage, to extract and deliver lots of detail, to manage a rhythm confidently, or to generate a large and persuasive soundstage.

Add in great ergonomics, a fine standard of build and finish and the ability to drive a couple of pairs of speakers without alarms, and it’s apparent the Cambridge Evo 150 SE is a very worthwhile proposition indeed. Even if it doesn’t look all that different to the product it replaces.

Oh, it's red and no mistake (Image credit: Future)Cambridge Evo 150 SE review: Price and release date

• Release date: July 2025
• Price: $3,299 / £1,999 / AU$5,999

The Cambridge Evo 150 SE launched in July 2025, and in the United Kingdom it sells for £1,999 – the same price as the outgoing Evo 150 had hit after launching at £2,249 back in 2021. In the United States you should expect to pay $3,299, while in Australia it’s currently listed at AU$5,999.

Of course, the market for streaming amplifiers is hotting up nicely – brands as venerable as NAD and as up-and-coming as Eversolo have products contesting this region of the market. So simply having a decent reputation for this sort of thing is only going to carry Cambridge (and the Evo 150 SE) so far…

Cambridge Evo 150 SE review: Features

• 150 watts of Hypex NCOREx ‘Tuned by Cambridge’ Class D amplification
• Numerous wired and wireless, analogue and digital inputs
• Four-layer PCB

There was very little wrong with the feature-set of the original EVO 150, so Cambridge has (very sensibly) left well enough alone. Or, at least, it has in all but one very significant area.

So what’s carried over from the outgoing model? Well, there’s the remarkably generous selection of physical and wireless inputs and outputs, for starters. Where analogue stuff is concerned, the Evo 150 SE has a line-level unbalanced input on stereo RCAs, a balanced equivalent via XLRs, and a moving magnet phono stage for use with a turntable accessed by another pair of RCAs. Digital inputs run to a digital coaxial input, a digital optical socket, an HDMI ARC and a USB-B input.

Wireless stuff is handled by dual-band wifi and Bluetooth with aptX HD codec compatibility. Wi-fi, of course, means the Cambridge can deal with AirPlay 2, Chromecast, Spotify Connect and TIDAL Connect - and it also gives access to internet radio. The 150 SE is Roon Ready, too.

As far as outputs go, there are binding posts for two pairs of speakers and a 3.5mm headphone socket. In addition, there’s a pre-out for a subwoofer, corresponding pre-outs for use with a power amplifier. And the Cambridge is a Bluetooth transmitter as well as a receiver, so it can drive your wireless headphones.

Incoming digital signals are handled by an ESS Sabre ES9018K2M digital-to-analogue converter that can deal with resolutions of up to 32bit/384kHz and DSD256. The Cambridge is compatible with all popular file types, and quite a few unpopular ones too.

So far, so very-similar-indeed-to-the-old-model. But what makes this machine an ‘SE’ is the fact that Cambridge has worked in such close collaboration with Hypex on its NCOREx Class D amplification that here it’s branded as ‘Tuned by Cambridge’. And it is, at least according to the companies involved, ready and able to produce “warm, dynamic and controlled sound”.

Features score: 5 / 5

Note the 'wings' (Image credit: Future)Cambridge Evo 150 SE review: Design

• 39 x 317 x 352mm (HxWxD)
• Large, bright, full-colour display
• Swappable side panels

Just as with ‘features’, not a lot has changed where the design of the Evo 150 SE is concerned - it demonstrably wasn’t broken, and so Cambridge has chosen not to fix it. Unlike the ‘features’ section, though, where there’s a difference here between the 150 and this 150 SE it is absolutely trifling.

So your £1999 buys you a nicely proportioned, beautifully made and flawlessly finished box with a bright, crisp full-colour display on the fascia that will display album artwork or virtual VU meters as well as providing menu and set-up options. The Evo 150 is supplied with two pairs of side-panels that attach magnetically and can easily be swapped - one is a real wood veneer, the other a dark grey slatted alternative.

So yes, the difference between the Evo 150 and this Evo 150 SE is - hold on to your hat - the knurling of the two-part volume dial/input selector. It’s less knurled than it was before.

Design score: 5 / 5

Everything in its right place… (Image credit: Future)Cambridge Evo 150 SE review: Sound quality

• Detail and dynamism in equal measure
• Big, organised soundstage
• Slightly overconfident low-frequency reproduction

There are differences between the way the Evo 150 SE sounds when delivering a DSD64 file of Tears For Fears’ Head Over Heels via some network-attached storage compared to the same song streaming from the free tier of Spotify, of course – but one of the most admirable things about this Cambridge machine is how consistent it sounds. No matter which of its inputs you’re using, its fundamental character comes through loud and clear.

And broadly speaking, its character is a lovely balance between ‘detail’ and ‘energy’, and between ‘scale’ and ‘dynamism’. The Evo 150 SE is capable of revealing the finest, most minor details in a recording, and giving them the appropriate amount of emphasis relative to the overall performance. It invests music with the sort of vigour and momentum that can bring it to life, without ever threatening to lose control of proceedings. It creates a large and well-defined soundstage, and ensures that every element of recording (no matter how numerous the elements might be) gets the required elbow-room to express itself – though it delivers music as a unified and singular piece at the same time. It has the sort of dynamic headroom that allows it to track the shifts in intensity or volume during the course of a recording without sounding pressured or stressed.

At the top of the frequency range, the Evo 150 SE attacks with chunky determination. Treble sounds have brilliance and a fair amount of bite, but they’re substantial enough to prevent any hardness or edginess creeping in. It’s a similar story through the midrange: the Cambridge is forward and direct, but never pushy, and it reveals a big amount of information, both broad and fine, about what’s going on there. Voices, in particular, benefit no end from this eloquence and positivity.

The tonal balance up to this point, and down into the low frequencies too, is almost (but not quite) neutral – there’s just a hint of warmth that suits the overall character of the Evo 150 SE just fine. The frequency response, though, is slightly skewed. The Cambridge puts slightly more emphasis on the bottom end than is absolutely ideal, with the result that the overall presentation is slightly tiled towards bass. It’s not that the low end here lacks detail or is in any way ponderous – the Evo 150 SE controls the low frequencies carefully at their attack and decay, and expresses rhythms confidently as a result. But while there are doubtless plenty of listeners who will interpret this slight bottom-heaviness as ‘exciting’, it serves to make the Evo 150 SE sound less than neutral. And to demand you take some care with partnering equipment, especially loudspeakers.

Having praised the Cambridge for the consistency of its sound through its numerous inputs, it’s nevertheless worth pointing out just what a gem the Evo 150 SE’s phono stage is. A vinyl copy of The Cinematic Orchestra’s Every Day sounds full, dynamic to almost comical degree, lavishly detailed and is delivered with complete positivity. Yes, that tilt towards the bottom end is still in evidence – but the way the Cambridge handles the various rhythms and tempos goes an awfully long way towards making up for it.

Sound quality score: 4.5 / 5

The bulk of my testing was done with the Sennheiser IE900 or the Austrian Audio 'The Composer' (Image credit: Future)Cambridge Evo 150 SE review: Usability & setup

• Remote handset or app control
• Simple to get attach to your local network
• Big-print on-screen menus are nicely realised

Control is available via the on-screen menus and the input dial, or via the perfectly adequate yet entirely unremarkable remote control handset. Or you can do what all sensible people will do, and use the fourth (and current) version of the StreamMagic control app. This is a comprehensive, logical, stable and usable app, which offers a lot of options – I’m particularly keen on the ability to include only those inputs you’re using on the home screen. It’s good for integrating your favourite music streaming services, saving half a dozen internet radio stations as presets, and plenty more besides.

As far as ‘setup’ goes, it’s simply a question of making the physical connections you want to make, and then getting the 150 SE on to your local network. It behaves as an AirPlay speaker when fresh out of the box, so finding it and hooking it to your network couldn’t be easier (unless you’re using an Ethernet connection to your router, which makes it easier still). Then load up your streaming services and internet radio stations, nominate the inputs you’re using to the front of the app, and away you go.

Usability & setup score: 5 / 5

This white light (for aptX Lossless) was oddly hard to come by when using sources able to handle it… (Image credit: Future)Cambridge Evo 150 SE review: Value

When the original Evo 150 launched back in 2021, it cost £2249 – and everyone seemed to think that was fair enough. Somehow Cambridge has managed to retain everything that was great about the original, dial in some upgraded amplification and lower the asking price to an extremely eye-catching £1999. So yes, there’s splendid value for money here.

Value score: 5 / 5

Note the new 'rails' on the casework to help with cooling (Image credit: Future)Should you buy Cambridge Audio Evo 150 SE?Buy it if...

You admire well-realised industrial design
The Evo 150 SE is an understatedly swish looker, and those magnetically attached side panel options are never less than gratifying.

The description of sound as ‘thumping’ strikes you as a positive
You’ll never find yourself craving greater low-end heft or substance.

You own (or will own) some capable loudspeakers
The Cambridge is an adept performer in so many ways - and it deserves the speakers that can do that fact some justice.

Don't buy it if...

Your speakers or source equipment aren’t of a similar standard
See above, really - if you want the Evo 150 SE to sound like the money’s-worth (which it most definitely can) your sources and speakers need to step up.

You crave perfect sonic balance
There’s a definite flavour to the way the Cambridge sounds - and as with all flavours, it’s very much a matter of taste.

(Image credit: Cambridge)Cambridge Audio Evo 150 SE review: Also consider

In 2025 the Naim Uniti Atom is an old stager – but that just means it’s matured, rather than being in any way past it. The audio balance is more neutral than the Evo 150 SE, and what the Naim lacks in inputs (there’s no phono stage, for instance) it more than makes up for in both tangibles and intangibles.

iFi makes strong design choices and you love to see it (Image credit: Future)How I tested the Cambridge Audio Evo 150 SE

• Two-week testing period
• Bowers & Wilkins speakers
• Rega P1 turntable; network storage; Tidal Connect, internet radio

With a pair of Bowers & Wilkins 702 S3 Signature at the business end, with a Rega Planar 1 turntable as a physical source and a mixture of Tidal Connect, network-attached storage, internet radio and various devices attached via Bluetooth at the other end, the Cambridge Evo 150 SE always held its own.

No genre of music seemed to be off-limits, and it wasn’t as ruthless with lower quality content as it might have been - try as I might, I couldn’t make it insist on hi-res stuff. Which means it was a pleasure to listen to in pretty much every circumstance.

• Read more about how we test

First reviewed August 2025

• Russia has already banned WhatsApp and Telegram
• Users flocked to Google Meet, which could also get banned
• WeChat-based Max platform will come pre-installed on an phones soon

Russia is considering banning Google Meet as part of a broader crackdown on foreign tech, according to The Moscow Times reporting.

Roskomnadzor, Russia’s federal agency responsible for the country’s media, has already banned voice and video calls via WhatsApp and Telegram, and the country is set to expand restrictions even further.

With users seeking alternatives, Google Meet quickly became one of the most popular alternatives – one that Roskomnadzor is now also looking to ban.

Russia bans Google Meet

Google Meet recently faced widespread disruptions, with over 2,000 individual reports complaining about frozen calls, missing video/audio and unexpected shutdowns, however Roskomnadzor has publicly denied any involvement.

“People started experimenting with other platforms after blocking calls in WhatsApp and Telegram, apparently, overloads led to certain failures,” Deputy Chairman of the State Duma Committee on Information Policy, Information Technologies and Communications Andrei Svintsov said in a conversation with Abzats.

“Roskomnadzor has not taken any measures to restrict the operation of the Google Meet video conferencing service in Russia, the agency’s press told Interfax,” news agency Interfax wrote in a (translated) Telegram post.

The post also described Meta as “extremist,” linking WhatsApp with fraud, sabotage and terrorism. WhatsApp had over 97 million users in Russia in July 2025 according to Reuters.

“However, applications that can monitor our citizens, transfer information to Western special services, may well be blocked,” Svintsov added.

Industry analysts are expecting an imminent ban on Google Meet to tie in with Russia’s promotion of Max, a state-backed app set to be pre-installed on all new smartphones from September, based on China’s WeChat.

Max, which is still in a testing phase, has amassed around 18 million users already.

You might also like

• We’ve listed the best online collaboration tools
• Enhance worker output with the best productivity tools
• Russia's WhatsApp rival to be pre-installed on new smartphones and tablets from September – here's what we know