TechRadar News

• Nvidia briefly became the first company ever to reach a $4 trillion valuation
• Data centre revenue and AI form huge part of Nvidia's success
• Despite brief dip, it remains the world's most valuable company

After months battling it out to become the world's most valuable company, and briefly holding that title on several occasions, Nvidia has become the first company ever to hit a staggering $4 trillion market cap.

Although the company's value has since dipped slightly, to just $3.972 trillion at the time of writing, it remains the world's most valuable company ahead of Microsoft and Apple – the only two other companies to have passed the coveted $3 trillion mark.

Although fluctuations continue to see the three companies change positions, Nvidia continues to see strong growth due to continued demand for AI chips.

Nvidia becomes the world's first $4 trillion company

Much of Nvidia's success can be attributed to the launch of ChatGPT in November 2022, as its stock has risen more than 15x in the past five years as a result of the demand for AI.

More recently, Nvidia shares are up 15% month-over-month and 22% year-to-date.

The news comes in the year after Nvidia hit $2 trillion (February 2024) and $3 trillion (June 2024) valuations.

Internally, Nvidia's earnings also continue to grow, with first-quarter revenue up 12% quarter-over-quarter and a health 69% year-over-year. The company also revealed that data centre revenue also up 73% compared with last year, accounting for over 88% of the company's total revenue.

"Global demand for Nvidia's AI infrastructure is incredibly strong. AI inference token generation has surged tenfold in just one year, and as AI agents become mainstream, the demand for AI computing will accelerate," CEO Jensen Huang explained.

This marks a departure from Nvidia's past, when it was known for its gaming GPUs. With Nvidia stock worth just 1% of what it's worth today eight years ago, it's unclear how much more growth the company could be set to experience.

You might also like

• We've listed the best cloud hosting providers
• Check out our roundup of the best AI tools and best AI writers
• Nvidia officially becomes the world's most valuable company — now can it stay there?

• A new update for The Last of Us Part 2 lets you play the game's story in chronological order
• Naughty Dog says Chronogical mode will allow players to "will gain even deeper insight into Part 2’s narrative"
• New trophies and Uncharted 4-themed skins tied to the mode have also been added

Naughty Dog has released a surprise update for The Last of Us Part 2 that lets you play the game's story in chronological order.

The new Chronological mode is now available for free and allows players to experience the game's non-linear story beats in the order that they take place.

Spoiler alert! If you've played The Last of Us Part 2, this means that those Ellie and Joel flashbacks that are scattered throughout the story will now be featured chronologically near the beginning of the game.

This also means that Joel's gut-punch death won't kick-start the emotional journey and will come much later, and we'll learn about Abby's motivations a lot sooner.

"Through the new Chronological mode, we believe players will gain even deeper insight into Part 2’s narrative," Naughty Dog explained in a PlayStation Blog post.

"Those who have already played will know its story is told non-linearly, as Ellie and Abby’s motivations, realizations, and emotional stakes unfold across myriad flashbacks and present-day storylines. While this structure is very intentional and core to how our studio wanted Part II’s themes and narrative beats to impact players, we always wondered what it would be like to experience this story chronologically. And now finally, we can answer that question."

One of the game's strong suits is its narrative structure, and although the Chronogical mode is a neat addition that I'll gladly experiment with, I don't think it will offer the same emotional payoff the original format does.

"It was no small feat to bring The Last of Us Part II’s story chronologically together, given that Part II’s story is so meticulously put together," the studio said. "We’re grateful to the developers both at Naughty Dog and our partners at Nixxes to make the Chronological mode as smooth as possible. And while we of course recommend players still new to the game to play through Part 2's story as was originally developed, the team’s hard work has paid off with a fascinating new way to enjoy this chapter."

In addition to this new mode, new trophies tied to the experience have been added, along with two new Uncharted 4: A Thief's End costumes for Joel and Tommy, inspired by Nathan Drake and Sam Drake.

These skins can be unlocked by completing the narrative in Chronological mode and will be usable in No Return, The Last of Us Part 2 Remastered's survival mode.

You might also like...

• Best PS5 games 2025 - the top 26 PlayStation 5 titles to play right now
• I’ve spent 40 hours exploring Death Stranding 2: On the Beach, and it’s an incredible sequel that builds upon its unique predecessor to become a masterpiece
• Marvel Rivals season 3 is making big changes to how you earn your rewards, but you can still claim extra goodies by simply watching your favorite Twitch streamer play the game

• The Google Pixel 10 Pro Fold has appeared on Geekbench
• Its scores are far lower than those of the Samsung Galaxy Z Fold 7
• However, it's still an improvement on the Pixel 9 Pro Fold's scores

Samsung just unveiled its line of foldable phones for 2025, headlined by the Samsung Galaxy Z Fold 7, and next up it’s Google, with the Google Pixel 10 Pro Fold expected to land in August – but if you're looking for a foldable and power is your priority, a new benchmark suggests you’re better off buying the Z Fold 7 than waiting for Google's new foldable.

A Geekbench listing for the Pixel 10 Pro Fold (spotted by NotebookCheck) includes a single-core score of 2,276 and a multi-core result of 6,173. In isolation those numbers might not mean much, but they’re well below the results we’ve seen for the Samsung Galaxy Z Fold 7.

While Geekbench hasn’t yet listed an average score for the Z Fold 7, recent results include the likes of 2,826 and 2,552 for single-core, and 9,053 and 8,639 for multi-core. It’s likely that when an average is reached it will be similar to the Samsung Galaxy S25 Ultra’s averages, as that phone has the same chipset and amount RAM, and that’s sitting at 2,878 for single-core and 9,510 for multi-core.

(Image credit: Geekbench)Expectedly underpowered

All of those numbers are much higher than the Google Pixel 10 Pro Fold’s results. Now, we’d take the Pixel’s results with a pinch of salt, since these would have been recorded on a pre-release handset, so further optimization could take place. And this is just one result, so it could end up being an outlier.

But there’s nothing overly surprising here, as the Google Tensor G5 chipset the phone will probably use is unlikely to match the Galaxy Z Fold 7’s Snapdragon 8 Elite chipset. That’s the case every year, with Google’s chipsets proving less powerful than rival options.

Still, this is an improvement on the Google Pixel 9 Pro Fold, which tends to record single-core scores of just under 2,000, and multi-core scores of around 4,500.

So going by this one Pixel 10 Pro Fold result, it should be a significant upgrade on the Pixel 9 Pro Fold at least, and while it probably won’t be the speediest phone around, it’s unlikely to feel lacking in power in everyday use.

We should find out for sure in August, with leaks pointing to August 20 as the announcement date for the whole Google Pixel 10 line.

You might also like

• Leaked Google Pixel 10 prototype photos give our best look yet at the phone
• Google Pixel 10 series colors and storage capacities leaked, with some weird limitations on pairings
• The Google Pixel 10 Pro series might barely be an upgrade

It’s finally official: The Big Bang Theory spinoff Stuart Fails to Save the Universe has been green lit by HBO Max, following earlier reports from 2023 that it had been in development. We don’t know when it will begin to film or air, but fans can start to get excited.

Or can we? This isn’t the first spinoff we’ve seen for the franchise, with CBS and Paramount+ partnering up to bring us seven seasons of Young Sheldon and new kid on the block Georgie & Mandy’s First Marriage.

Even though there was overlap between the hit comedy series and its initial prequel, The Big Bang Theory (TBBT) hasn’t been on our screens since 2019. But making a comeback with a spinoff like this has me worried, and not in a quirky Sheldon Cooper way.

Stuart Fails to Save the Universe isn’t the part of The Big Bang Theory universe we want to see

Stuart (Kevin Sussman), Raj (Kunal Nayyar) and Sheldon (Jim Parsons) in The Big Bang Theory. (Image credit: CBS)

If you could choose to follow-up with any element of The Big Bang Theory, it definitely wouldn’t be with comic book store owner Stuart (Kevin Sussman). When we last saw him, he’d finally moved in with the love of his life Denise (Lauren Lapkus), who he’d originally hired to help in the store after it became popular.

Here’s what HBO Max tells us to expect from Stuart Fails to Save the Universe: “Tasked with restoring reality after he breaks a device built by Sheldon and Leonard, accidentally bringing about a multiverse Armageddon. Stuart is aided in this quest by his girlfriend Denise, geologist friend Bert ( Brian Posehn), and quantum physicist/all-around pain in the ass Barry Kripke (John Ross Bowie). Along the way, they meet alternate-universe versions of characters we’ve come to know and love from The Big Bang Theory. As the title implies, things don’t go well.”

It’s a clever get-out-of-jail-free card for HBO Max. As the synopsis suggests, fan favourites from TBBT might return, but it isn’t guaranteed. The most obvious follow-up to the main series would have been to catch up with Sheldon (Jim Parsons), Penny (Kayley Cuoco), Leonard (Johnny Galecki) or Howard (Simon Helberg). Choosing to follow a minor character fans might not remember or care about is the very definition of the meme “go girl, give us nothing”.

Can we then assume that a left-field approach like Stuart Fails to Save the Universe could actually damage the legacy of TBBT? We talk a lot about existing IPs being left alone in favour of original storytelling, and this could be a great example. It’s currently unclear what value the new spinoff will add, other than choosing to appease fans with surprise cameos as and when it wants to.

There's proof in the spinoff pudding, too. While Young Sheldon became a surprise smash hit that ruled the Nielsen ratings, Georgie & Mandy’s First Marriage has been much more mixed. Granted, it’s been renewed for a second season, but after the single-cam format of Young Sheldon, it returned to TBBT’s signature multi-cam approach. Combined with a laugh track, the audience reactions suggest the time for classic sitcom formats has been and gone.

If Stuart Fails to Save the Universe can lean into what came before in TBBT without making it a superficial cameo soup with its sights set on high ratings, then there’s potential for a job well done. If not, well… Stuart might fail to save his own show.

You might also like

• Peacemaker season 2: everything we know so far
• The Last of Us creators confirm exactly what I expected for season 3 of the hit HBO Max show
• Harry Potter TV show: everything we know so far about the upcoming HBO adaptation

• Samsung says the tri-fold foldable will be out in 2025
• The Tab S11 tablets are coming too – probably in September
• We already have a host of new devices to test out

We saw plenty of new product announcements at the Samsung Galaxy Unpacked 2025 event yesterday, but we didn't quite get all of the news we were expecting – and Samsung has gone on the record to clarify that more launches are on the way.

First up, as reported by The Korea Times (via Android Authority), Samsung mobile chief TM Roh has said the long-awaited Galaxy tri-fold phone should be coming before the end of 2025, so adjust your savings plans accordingly.

"We are working hard on a tri-fold smartphone with the goal of launching it at the end of this year," Roh told journalists. "We are now focusing on perfecting the product and its usability, but we have not decided its name."

One name in the running is apparently the Samsung Galaxy G Fold, but watch this space. The device will seemingly come with two hinges and three sections to its main display, and is also rumored to cost quite a bit (as you would expect).

Tablets are coming too

The Samsung Galaxy Tab S10 Ultra (Image credit: Philip Berne / Future)

Next up, there's the Samsung Galaxy Tab S11 tablet series, following on from the models we saw last year, including the Samsung Galaxy Tab S10 Plus. There's an update here as well, again courtesy of Android Authority.

According to a Samsung executive, the Galaxy Tab S11 devices will be "coming through shortly" in line with the "traditional cadence" – which most likely means a September launch this year, to match the 2024 schedule.

Leaks around the Samsung Galaxy Tab S11 tablets have been pretty thin on the ground so far, but we are expecting both a standard model and an Ultra edition to match the Samsung Galaxy Tab S10 Ultra – as Samsung takes on the best iPads once more.

In the meantime, we've got hands-on impressions of all the devices Samsung did announce yesterday: the Galaxy Z Fold 7, Galaxy Z Flip 7, Galaxy Z Flip 7 FE, Galaxy Watch 8, Galaxy Watch 8 Classic.

You might also like

• More Samsung Galaxy tri-fold specs have leaked
• These are the best Samsung phones you can buy
• One UI 8 features coming to your Galaxy phone

• OpenAI plans to launch an AI-powered web browser, report claims
• It could arrive “in the coming weeks” and compete with Google Chrome
• This would give OpenAI access to extensive user data, Reuters says

The browser wars could be about to heat up in a big way, with ChatGPT-maker OpenAI apparently set to launch its own web browser in the coming weeks. According to Reuters, that could put significant pressure on Google Chrome and potentially “fundamentally change how consumers browse the web.”

Citing “three people familiar with the matter,” Reuters reports that the browser will have an artificial intelligence (AI) chat interface that would keep many user interactions within the chat window rather than linking out to external websites.

As well as that, the browser might integrate OpenAI’s AI agent – dubbed Operator – that would allow the app to “carry out tasks on behalf of the user.” This could include “booking reservations or filling out forms” on the websites you use.

OpenAI’s browser is apparently built using Google’s open-source Chromium tech, which powers Chrome, Edge, and many of the other best web browsers. OpenAI’s product is due to launch “in the coming weeks,” Reuters’ sources believe.

Analysis: The battle for your data goldmine

(Image credit: OpenAI)

OpenAI faces stiff competition in the browser world. Google Chrome currently enjoys a stranglehold position, with roughly two-thirds of the available market share.

AI firm Perplexity and web firms Brave and The Browser Company have also launched their own AI browsers. Like OpenAI’s rumored effort, Perplexity’s browser can perform tasks on your behalf.

Reuters suggests a clear motive for OpenAI: user data. Running its own browser would allow the company to harvest as much information as possible from users, which could then go back into training its AI models and providing other monetization opportunities.

After all, Chrome “provides user information to help Alphabet target ads more effectively and profitably, and also gives Google a way to route search traffic to its own engine by default,” Reuters says. An OpenAI browser would give the AI firm a similarly powerful access route to lucrative data.

If you’re concerned about your privacy, then, OpenAI’s browser is likely to be ringing alarm bells in your head. OpenAI has faced criticism for its data collection practices, as has Google Chrome. As someone who's used Firefox for over two decades, that has me worried.

When scooping up users’ private data is an incentive for launching a browser – as Reuters implies it might be – serious caution is advised. We’ll have a clearer idea of all this when OpenAI’s web browser launches later this year, so keep your eyes peeled.

You might also like

• ChatGPT explained – everything you need to know about the AI chatbot
• Best web browser of 2025
• These are the worst web browsers for sucking up all your data, so you may want to stop using them

• The iPhone 17 Air could be sold in black, silver, light gold, and light blue shades
• It may also have an A19 Pro chipset, rather than the base A19
• However, previous leaks disagree with that chipset claim

If you’ve got your sights set on the iPhone 17 Air, you can also now start thinking about what color you'd like your slimline iPhone to be, as a new leak has revealed the four shades that Apple will apparently sell the device in.

According to leaker Majin Bu – citing a previous leak from Fixed Focus Digital which they claim to have now “confirmed” – the iPhone 17 Air will be available in black, silver, light gold, and light blue shades.

The black shade is apparently a “dark and understated tone”, the silver model is said to be “bright and clear”, the light gold is described as being “warm and soft”, and the light blue as “very light” and “evoking a clear sky”.

(Image credit: Majin Bu)

You can see how these colors might look in the image Majin Bu provided above, though it sounds like the shades shown are probably guesses based on how their sources described the colors.

In any case, we’d take this information with a pinch of salt – although the use of mostly quite pale, delicate shades seems fitting for phone that’s likely to be billed as slim and elegant.

Some surprising chipset news

In other iPhone 17 Air news, leaker Fixed Focus Digital (via Phone Arena) has claimed that the handset will have an A19 Pro chipset. That’s at odds with previous leaks, which suggested only the iPhone 17 Pro and Pro Max would have this chipset, while the iPhone 17 and iPhone 17 Air would have an A19.

However, they added that the iPhone 17 Air will get a version of the A19 Pro with a five-core GPU while the iPhone 17 Pro models will get a six-core version, in which case the Pro models would still have the most powerful chipset.

They also mention RAM, saying that the standard iPhone 17 will have 8GB, while the other three models will have 12GB. So if this is all correct then the upcoming iPhones could have three different power tiers rather than the usual two.

We’re a bit skeptical of this since it would be a change of form for Apple, and since other leaks point to the iPhone 17 Air having the base A19 chipset, but if Apple does boost the power that could help to make up for the Air only having one rear camera lens, which is something that’s widely rumored to be the case.

We should find out exactly how powerful the iPhone 17 Air is – and what colors it’s available in – sometime in September, as that’s when it will likely launch, alongside the rest of the iPhone 17 series.

You might also like

• The iPhone 17 Pro Max could have a much bigger battery than any previous model
• iPhone 17 Air leaked specs suggest it's likely to beat the S25 Edge in one key way – and lose out in another
• The foldable iPhone’s cameras might leave it trailing behind the Galaxy Z Fold 7

• Small businesses are less confidence with AI than larger ones
• Half of businesses say AI has become critical
• Training and policies should be more comprehensive

New research has claimed barely one in 10 (12%) SMEs have invested in AI-related training for their staff.

The report from The Institute of Coding revealed nearly one in three (29%) SMEs now see a lack of training as their biggest obstacle to AI tools adoption, with a further one in two (52%) citing a lack of internal skills and knowledge as the main battier.

Moreover, the research depicts a troubling picture for smaller companies when compared with their larger counterparts – 82% of medium businesses expressed confidence working with AI compared with 37% of smaller businesses.

Small businesses are struggling to adopt AI

Around one in two (51%) of the SMEs surveyed agreed that AI could now be perceived as critical, but only around half of these (27%) believe they can safely and effectively implement AI tools.

As such, The Institute of Coding is warning of a growing AI readiness divide between different types of companies, launching its own free short AI courses aimed at all career levels.

Looking ahead, companies that are on the verge of being left out by the AI revolution are now calling for greater support from the government.

Three in five (59%) call for national AI skills strategies to support businesses of all sizes, with three-quarters asking for clearer guidance on the AI skills they're likely to need in the next three to five years.

"This isn’t just about individual business success – it’s about ensuring the entire UK economy can participate in the AI transformation," Professor Rachid Hourizi MBE, Director of the Institute of Coding, explained.

The report goes on to explain that, if micro businesses and sole traders are not explicitly included in national plans, AI will become concentrated, not democratized.

You might also like

• These are the best AI writers around today
• AI still isn't living up to its full potential for many business
• Prepare for the future with the best online learning platforms

• Former Microsoft executive Peter Moore has said that if Xbox hadn't spent so much time and money repairing Xbox 360 consoles, "I'm not sure the Xbox brand would be around today"
• Moore called it a "defining moment" for the company
• He also believes that if Xbox did fall, the games industry would have suffered

Former Microsoft lead Peter Moore believes the Xbox brand might not be around today if the company hadn't focused on repairing countless Xbox 360 consoles.

In an interview with The Game Business, Moore suggested that the company was in hot water trying to figure out how to fix the Xbox 360's "red ring of death", but determined that spending the $1.15 billion on repairs is what ultimately saved its future.

"It took us a while to figure out what was going on," Moore said (thanks, GamesRadar). "Were the fans in the right place? [...] Trying to figure out whether, you know, [...] wrapping the towel around it would create more heat, which would rejoin some of the issues of the cracks in some of the units. All of this was going on, and it was stressful. But the one thing I will always say is, you know, this was, for us, a defining moment.

"If we hadn't done what we did, I'm not sure the Xbox brand would be around today."

Moore continued, saying, "We felt that that was money well spent to hang on to a brand that we built, that we felt had huge viability going forward – and of course it does," he said. "And doing good by the gamers."

However, Moore believes that if Xbox had fallen, the game industry as a whole would have suffered.

"Rising tide lifts all ships. Microsoft's entry into the market created a massive rising tide," he said. "They put billions of dollars into marketing, advertising, R&D, and I don’t think the industry would be anywhere close to what it is today if Microsoft wasn't involved."

In other news, Microsoft recently made significant cuts to its gaming division, laying off an estimated 9,000 staff members, and canceling multiple projects such as the Perfect Dark reboot, Everwild, and an unannounced MMO from The Elder Scrolls Online developer.

You might also like...

• Best Xbox Series X games 2025: smash hits for Microsoft's top console
• I’ve spent 40 hours exploring Death Stranding 2: On the Beach, and it’s an incredible sequel that builds upon its unique predecessor to become a masterpiece
• Marvel Rivals season 3 is making big changes to how you earn your rewards, but you can still claim extra goodies by simply watching your favorite Twitch streamer play the game

This week, Samsung Galaxy Unpacked unveiled its latest smartphones, the Galaxy Z Fold 7, Flip 7, and Flip 7 FE, all integrated with Google Gemini AI features. To sweeten the deal for potential customers, Google announced a special offer: six free months of Google AI Pro for those who purchase the new phones. This premium subscription service includes access to the Gemini 2.5 Pro model, the Google Veo 3 AI video generator, two terabytes of cloud storage, and early access to upcoming AI features.

Of course, once those six months are up, you'll have to pay the standard $20 a month to keep your subscription. But, Google likely believes more than a few people will be happy to pay after they get accustomed to its AI toolkit. The psychology behind this is as simple as free samples at the grocery store. Google isn’t trying to sell you a subscription right now because it thinks you won't want to give it up just because it isn't free anymore.

It's a pretty impressive set of features. Veo 3 is one of the most powerful consumer-facing video generators available. And Gemini 2.5 Pro is far more coherent in conversation than its predecessors.

Gemini try and buy

It's easy to imagine how Google hopes the six months will go. You might spend a month fiddling with Veo and creating movies about your dog going on adventures. Or start turning to Gemini to summarize very long emails, and eventually every email. Or you might get a great recipe from a random prompt to Gemini and soon use it to plan your every meal. By the end of six months, Google's AI might just be what you turn to a dozen times a day as a reflex. By the time Google asks for $20 a month, you might even consider it a bargain.

That's Google's dream scenario, but it comes with a risk. Google is betting that people will find these tools indispensable. But, if people take it for granted for six months, they might resent having to pay for it, no matter how much you enjoy playing with Veo 3 and talking to Gemini. Nobody likes the feeling of having something useful pulled away unless you pony up. That's probably even more likely when it comes bundled with a device that already costs over a thousand dollars. There’s a version of this where the user relationship becomes less “wow, this is useful,” and more “wait, I have to pay extra for that now?”

But, I wouldn't be surprised if a scenario somewhere between the extremes still makes Google happy. Tying its most advanced AI tools to Samsung’s brand-new hardware is smart. The Galaxy Z Fold 7 and Flip 7 are devices people buy because they want the bells and whistles. They're practically built for people who like to show off. In other words, the people most likely to find ways to use AI enough to justify $20.

But it cuts both ways. Because if the experience feels too essential, people will feel punished when it disappears. And if it doesn’t feel essential enough, they won’t bother subscribing. The six-month trial is walking a very fine line between generosity and locking someone in to an AI future.

You might also like

• Google just announced 5 new Gemini features coming to Android, and it’s good news for fans of foldable smartphones
• I tried Google’s new Gemini-powered clothing app – here’s how you can use AI to find the perfect outfit
• Adding Google Gemini to Samsung's Ballie AI robot sounds impressive, but I'm not sure it matters

At a time when many believe that oversight of the Artificial Intelligence industry is desperately needed, the US government appears to have different ideas. The "One Big Beautiful Bill Act" (OBBBA)—recently given the nod by the House of Representatives—includes a 10-year moratorium on state and local governments enacting or enforcing regulations on AI models, systems, or automated decision-making tools.

Supporters claim the goal is to streamline AI regulation by establishing federal oversight, thereby preventing a patchwork of state laws that could stifle innovation and create compliance chaos. Critics warn that the moratorium could leave consumers vulnerable to emerging AI-related issues, such as algorithmic bias, privacy violations, and the spread of deepfakes.

Basically, if the AI sector is the Wild West, no one will be allowed to clean up Dodge.

Why should we care?

History may not literally repeat itself, but there are historical patterns and trends that we can view and hopefully be informed by, and our history books are packed with examples of technology reshaping the lives of the workforce.

And be it in the form of James Watt’s steam engine or Henry Ford’s moving assembly line, the cost of the progress brought by fresh technology is regularly paid by the large numbers of people sent home without a pay packet.

And AI will cost jobs too.

Experts such as those at McKinsey, the Lancet, or the World Economic Forum (WEF) may not agree on exact numbers or percentages of lost jobs, but the consistent message is that it will be bad:

• 30% of US work hours across all sectors will be automated by 2030 says McKinsey
• 25% of medical administrative tasks could vanish by 2035 according to a Lancet study
• 39% of existing skill sets will become outdated between now and 2030 warns WEF

Of course, as with all new technologies, new jobs will be created. But we can’t all be prompt engineers.

The Great Brain Robbery

Essentially, those hit hardest by the bulk of new technologies from the Spinning Jenny onwards were the ones engaged to carry out physical work. But AI wants to muscle in on the intellectual and creative domains previously considered uniquely human. For example, nonpartisan American think tank the Pew Research Center reckons 30% of media jobs could be automated by 2035.

And those creative jobs are under threat because creatives are being ripped off.

Many AI models are trained on massive datasets scraped from the internet, and these often include articles, books, images, music and even code that are protected by copyright laws, but AI companies lean heavily towards take-first-ask-later. Obviously, artists, writers, and other content creators see this practice as unauthorized use of their intellectual property and they argue that ultimately, it’s not even in the best interests of the AI sector.

If AI takes work away from human creatives—devastating creative industries already operating on thin margins—there will be less and less innovative content to feed to AI systems which will result in AI feeding off homogenized AI content – a derivative digital snake eating its own tail.

A smarter way forward would be to find a framework where creatives are compensated for use of their work to ensure the sustainability of human produced product. The music industry already has a model where artists receive payments via performing rights organizations such as PRS, GEMA and BMI. The AI sector needs to find something similar.

To make this happen, regulators may need to be involved.

Competitive opportunity versus minimizing societal harm

Without regulation, we risk undermining the economic foundations of creative and knowledge-based industries. Journalism, photography, literature, music, and visual arts depend on compensation mechanisms that AI training currently bypasses.

The United Kingdom and the European Union are taking notably different paths when it comes to regulating AI. The EU is pursuing a strict, binding regulatory framework, an approach designed to protect fundamental rights, promote safety, and ensure ethical use of AI across member states. In contrast, the UK is currently opting for a more flexible approach, emphasizing innovation and light-touch oversight aiming to encourage rapid AI development and attracting investment.

But this light-touch strategy could be a massive misstep – one that in the long term could leave everyone wishing we’d thought things through.

While AI enthusiasts may initially be pleased with minimal interference from regulators, eventually AI businesses will come up against consumer trust, something they absolutely need.

While AI businesses operating in Europe will be looking at higher compliance costs, there is also a clearer regulatory landscape and therefore more likely to be greater consumer trust – a huge commercial advantage.

Meanwhile, AI businesses operating in light-touch markets (such as the UK) need to consider how their AI data practices align with their (and their competitors’) brand values and customer expectations. As public awareness grows, companies seen as exploiting creators may face reputational damage. And a lack of consumer confidence could lead to a shift in mindset from previously arm’s-length regulators.

Regardless of the initial regulatory environment, early adopters of ethical AI practices may gain competitive advantages as regulatory requirements catch up to ethical standards. Perhaps the wisest way forward is to voluntarily make Dodge City a better place, even if there’s no sheriff in town – for now.

I tried 70+ best AI tools.

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

As organizations continue to adopt AI tools, security teams are often caught unprepared for the emerging challenges. The disconnect between engineering teams rapidly deploying AI solutions and security teams struggling to establish proper guardrails has created significant exposure across enterprises. This fundamental security paradox—balancing innovation with protection—is especially pronounced as AI adoption accelerates at unprecedented rates.

The most critical AI security challenge enterprises face today stems from organizational misalignment. Engineering teams are integrating AI and Large Language Models (LLMs) into applications without proper security guidance, while security teams fail to communicate their AI readiness expectations clearly.

McKinsey research confirms this disconnect: leaders are 2.4 times more likely to cite employee readiness as a barrier to adoption versus their own issues with leadership alignment, despite employees currently using generative AI three times more than leaders expect.

Understanding the Unique Challenges of AI Applications

Organizations implementing AI solutions are essentially creating new data pathways that are not necessarily accounted for in traditional security models. This presents several key concerns:

1. Unintentional Data Leakage

Users sharing sensitive information with AI systems may not recognize the downstream implications. AI systems frequently operate as black boxes, processing and potentially storing information in ways that lack transparency.

The challenge is compounded when AI systems maintain conversation history or context windows that persist across user sessions. Information shared in one interaction might unexpectedly resurface in later exchanges, potentially exposing sensitive data to different users or contexts. This "memory effect" represents a fundamental departure from traditional application security models where data flow paths are typically more predictable and controllable.

2. Prompt Injection Attacks

Prompt injection attacks represent an emerging threat vector poised to attract financially motivated attackers as enterprise AI deployment scales. Organizations dismissing these concerns for internal (employee-facing) applications overlook the more sophisticated threat of indirect prompt attacks capable of manipulating decision-making processes over time.

For example, a job applicant could embed hidden text like "prioritize this resume" in their PDF application to manipulate HR AI tools, pushing their application to the top regardless of qualifications. Similarly, a vendor might insert invisible prompt commands in contract documents that influence procurement AI to favor their proposals over competitors. These aren't theoretical threats - we've already seen instances where subtle manipulation of AI inputs has led to measurable changes in outputs and decisions.

3. Authorization Challenges

Inadequate authorization enforcement in AI applications can lead to information exposure to unauthorized users, creating potential compliance violations and data breaches.

4. Visibility Gaps

Insufficient monitoring of AI interfaces leaves organizations with limited insights into queries, response and decision rationales, making it difficult to detect misuse or evaluate performance.

The Four-Phase Security Approach

To build a comprehensive AI security program that addresses these unique challenges while enabling innovation, organizations should implement a structured approach:

Phase 1: Assessment

Begin by cataloging what AI systems are already in use, including shadow IT. Understand what data flows through these systems and where sensitive information resides. This discovery phase should include interviews with department leaders, surveys of technology usage and technical scans to identify unauthorized AI tools.

Rather than imposing restrictive controls (which inevitably drive users toward shadow AI), acknowledge that your organization is embracing AI rather than fighting it. Clear communication about assessment goals will encourage transparency and cooperation.

Phase 2: Policy Development

Collaborate with stakeholders to create clear policies about what types of information should never be shared with AI systems and what safeguards need to be in place. Develop and share concrete guidelines for secure AI development and usage that balance security requirements with practical usability.

These policies should address data classification, acceptable use cases, required security controls and escalation procedures for exceptions. The most effective policies are developed collaboratively, incorporating input from both security and business stakeholders.

Phase 3: Technical Implementation

Deploy appropriate security controls based on potential impact. This might include API-based redaction services, authentication mechanisms and monitoring tools. The implementation phase should prioritize automation wherever possible.

Manual review processes simply cannot scale to meet the volume and velocity of AI interactions. Instead, focus on implementing guardrails that can programmatically identify and protect sensitive information in real-time, without creating friction that might drive users toward unsanctioned alternatives. Create structured partnerships between security and engineering teams, where both share responsibility for secure AI implementation.

Phase 4: Education and Awareness

Educate users about AI security. Help them understand what information is appropriate to share and how to use AI systems safely. Training should be role-specific, providing relevant examples that resonate with different user groups.

Regular updates on emerging threats and best practices will keep security awareness current as the AI landscape evolves. Recognize departments that successfully balance innovation with security to create positive incentives for compliance.

Looking Ahead

As AI becomes increasingly embedded throughout enterprise processes, security approaches must evolve to address emerging challenges. Organizations viewing AI security as an enabler rather than an impediment will gain competitive advantages in their transformation journeys.

Through improved governance frameworks, effective controls and cross-functional collaboration, enterprises can leverage AI's transformative potential while mitigating its unique challenges.

We've listed the best online cybersecurity courses.

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 first-generation system is slower than tape but aims to scale up rapidly by 2030
• Cerabyte’s roadmap involves physics so advanced it sounds like sci-fi with helium ion beams
• Long-term capacity hinges on speculative tech that doesn’t yet exist outside lab settings

Munich-based startup Cerabyte is developing what it claims could become a disruptive alternative to magnetic tape in archival data storage.

Using femtosecond lasers to etch data onto ceramic layers within glass tablets, the company envisions racks holding more than 100 petabytes (100,000TB) of data by the end of the decade.

Yet despite these bold goals, practical constraints mean it may take decades before such capacity sees real-world usage.

The journey to 100PB racks starts with slower, first-generation systems

CMO and co-founder Martin Kunze outlined the vision at the recent A3 Tech Live event, noting the system draws on “femtosecond laser etching of a ceramic recording layer on a glass tablet substrate.”

These tablets are housed in cartridges and shuttled by robotic arms inside tape library-style cabinets, a familiar setup with an unconventional twist.

The pilot system, expected by 2026, aims to deliver 1 petabyte per rack with a 90-second time to the first byte and just 100MBps in sustained bandwidth.

Over several refresh cycles, Cerabyte claims that performance will increase, and by 2029 or 2030, it anticipates “a 100-plus PB archival storage rack with 2GBps bandwidth and sub-10-second time to first byte.”

The company’s long-term projections are even more ambitious, and it believes that femtosecond laser technology could evolve into “a particle beam matrix tech” capable of reducing bit size from 300nm to 3nm.

With helium ion beam writing by 2045, Cerabyte imagines a system holding up to 100,000PB in a single rack.

However, such claims are steeped in speculative physics and should, as the report says, be “marveled at but discounted as realizable technology for the time being.”

Cerabyte’s stated advantages over competitors such as Microsoft’s Project Silica, Holomem, and DNA storage include greater media longevity, faster access times, and lower cost per terabyte.

“Lasting more than 100 years compared to tape’s 7 to 15 years,” said Kunze, the solution is designed to handle long-term storage with lower environmental impact.

He also stated the technology could ship data “at 1–2GBps versus tape’s 1GBps,” and “cost $1 per TB against tape’s $2 per TB.”

So far, the company has secured around $10 million in seed capital and over $4 million in grants.

It is now seeking A-round VC funding, with backers including Western Digital, Pure Storage, and In-Q-Tel.

Whether Cerabyte becomes a viable alternative to traditional archival storage methods or ends up as another theoretical advance depends not just on density, but on long-term reliability and cost-effectiveness.

Even if it doesn't become a practical alternative to large HDDs by 2045, Cerabyte’s work may still influence the future of long-term data storage, just not on the timeline it projects.

Via Blocksandfiles

You might also like

• These are the best mobile workstations you can buy right now
• We've also listed the best mini PCs for every budget
• Another dual Intel GPU card with 48GB VRAM launched as Arc Pro emerges as cheap Nvidia and AMD alternative for AI crowd

• AOOSTAR’s NEX395 has the power, but the cooling system remains a complete mystery
• Radeon 8060S beats RX 7600 XT in specs, making external GPU pairing confusing
• Without OCuLink, the eGPU dock likely suffers major bottlenecks in real-world tasks

AOOSTAR NEX395 is the latest in a growing field of AI-focused mini PCs which comes in a box-like casing that departs from the more common designs found in the segment.

The company says the NEX395 uses AMD’s flagship Strix Halo processor, a 16-core, 32-thread chip with boost speeds up to 5.1GHz.

It includes 40 RONA 3.5 compute units and appears to support up to 128GB of memory, most likely LPDDR5X given the compact casing.

Memory capacity matches rivals, but key hardware details are missing

This level of memory is in line with other mini PCs targeting AI development workflows, especially those involving large language models.

However, no details have been confirmed regarding storage, cooling, or motherboard layout.

The device looks more like an oversized SSD enclosure or an external GPU dock than a full-fledged desktop system.

Its slim, rectangular, vent-heavy design completely deviates from the usual cube or NUC-style mini PCs.

Holding it in your palm feels more like gripping a chunky power bank or a Mac mini cut in half, definitely not what you’d expect from a 16-core AI workstation.

The layout makes you question where the thermal headroom or upgradable internals even fit.

The AOOSTAR NEX395 includes an integrated Radeon 8060S GPU, part of the Ryzen AI MAX+ 395 APU.

However, it also sells an external eGPU enclosure featuring the Radeon RX 7600 XT.

Given that the integrated GPU already offers a newer architecture and more compute units than the RX 7600 XT, the use case for pairing the two is unclear.

Also, the NEX395 does not appear to support high-speed eGPU connectivity like OCuLink, which would limit bandwidth for external graphics support.

Port selection includes dual Ethernet ports, four USB-A ports, USB-C, HDMI, and DisplayPort outputs, along with a dedicated power input, suggesting reliance on an external power brick.

Without confirmed thermal design or sustained performance metrics, it’s unclear whether this system can function reliably in roles normally filled by the best workstation PC or best business PC options.

Unfortunately, the pricing details for the NEX395 are currently unavailable.

Given the $1500–$2000 range of comparable models such as the HP Z2 Mini G1a and GMKTEC EVO-X2, AOOSTAR’s model is unlikely to be cheap.

Via Videocardz

You might also like

• These are the best mobile workstations you can buy right now
• We've also listed the best mini PCs for every budget
• Another dual Intel GPU card with 48GB VRAM launched as Arc Pro emerges as cheap Nvidia and AMD alternative for AI crowd

• OnexGPU Lite reuses the same chip but adds Thunderbolt 5 to stay relevant
• The RX 7600M XT continues to show up while RDNA4 remains nowhere in sight
• AMD keeps winning eGPU slots while Nvidia remains largely absent from this niche segment

The external GPU market has been quietly evolving in recent years, and AMD appears to be securing a rather strange lead in this niche.

The debut of OnexGPU Lite makes it the 11th known eGPU powered by an AMD Radeon RX 7000 series chip, and it’s now clear vendors are consistently choosing AMD over Nvidia for their modular graphics solutions.

However, the reason(s) behind this momentum remains unclear, especially when broader market trends still favor Nvidia for desktop and mobile gaming.

Thunderbolt 5 takes the spotlight

The OnexGPU Lite is the latest entry in a growing list of eGPUs using the Radeon RX 7600M XT, a mobile RDNA3 GPU with a known 120W power ceiling.

Although not the best GPU in AMD's lineup, it has become a go-to for modular setups.

According to Onexplayer, the Lite version is currently undergoing beta testing and will launch "soon," but there is no confirmed price, release date, or detailed spec sheet.

Unlike the higher-end OnexGPU 2 that features the Radeon RX 7800M, the Lite version isn’t targeting raw power.

Instead, it seems designed to balance portability and futureproofing, with one key upgrade: support for Thunderbolt 5.

This is a notable development, as it marks one of the first eGPUs to adopt the new interface.

Onexplayer claims Thunderbolt 5 will mean "PCIe bandwidth will be doubled," although the actual PCIe tunneling remains at 64Gbps, the same as OCuLink.

What sets Thunderbolt 5 apart is its ability to support both power delivery and display output over a single cable, features that OCuLink lacks.

This emphasis on all-in-one connectivity is likely to appeal to creators using a laptop for video editing or for Photoshop.

For them, fewer cables and more streamlined setups can make a real difference.

Still, the reliance on the RX 7600M XT, with no sign of RDNA4 hardware on the horizon, does raise questions about performance ceilings.

That said, it appears that the selling point of this device will be the inclusion of Thunderbolt 5, but whether this will justify its place in a market still searching for a truly compelling external graphics solution remains to be seen.

Without more powerful mobile chips available, vendors are essentially repackaging the same core GPU in new chassis with slightly upgraded ports.

The AMD-centric trend in the eGPU space might seem surprising, but it could reflect pricing, power efficiency, or driver integration preferences.

Via Videocardz

You might also like

• Check out the best 3D modeling software for 3D printing and more
• We've rounded up the best portable monitors available now
• Lenovo set to introduce white ThinkPad laptops with ambitious design

This week, Samsung is boldly unveiling some of its most remarkable folding phones ever. They're thinner, lighter, smarter, and, yes, more expensive. In the US, at least, that trend may continue in more dramatic fashion if President Trump moves forward with his 25% tariff on goods produced in South Korea.

Perhaps you didn't realize that despite its ubiquity in the US, Samsung is based and operated out of South Korea. Like many global tech companies, it manufactures products at its home base, as well as in Vietnam, India, and Taiwan. In the US, President Donald Trump is trying to drag manufacturing back to the US shores and doing so mostly through the coercion of tariffs, which are basically taxes applied to all goods shipped into the US. It's a cost that some worry will eventually be passed along to the consumer.

While not directly addressing the price of the now more expensive Galaxy Z Fold 7 and other Samsung mobile devices, Samsung Executive Vice President of Mobile Experience Dave Das said, during a Samsung Unpacked breakfast panel this week in response to a question on the impact of tariffs, "I'll say the chips have not fully fallen where they may."

Das joked, "The situation is so fluid, so rapidly changing, that I feel like I have to check my phone right now to make sure whatever I'm saying is still applicable."

Samsung, Das contends, could be in a better position – at least as it refers to mobile products – to weather these fast-changing global trade circumstances. "I think one of Samsung's greatest strengths is how agile and flexible we are," said Das, referring to Samsung's skills in manufacturing and supply chain management.

His team is gaming out various scenarios, but they are also keeping the lines of communication open. "We are working closely with this administration to ensure that no matter what, Samsung is able to deliver the best products, the best experiences, the best services to US consumers at an attractive price and a competitive price."

Das didn't talk specifically about any product or reference the $100 price increase on the latest Z Fold model, though it's fair to assume that this adjustment is less about tariff concerns and more about more expensive components (the new 200MP sensor) and manufacturing (4.2mm thickness).

A dynamic situation

Flexibility in the rapidly evolving tariff picture is key, noted Das, adding that the team wants to manage and "work with the administration, again, to ensure we stay on course, and focused and we're delivering great products."

It's a solid and rational answer in the face of what may be some irrational forces. Keeping track of where the US Administration is applying tariffs and by how much is almost impossible because it has changed if not by the hour, then certainly by the day.

As I write this, the tariffs on South Korea could equal 25%. By the time you read it, it could be lower or higher. What will matter to consumers most, though, is what they'll be paying for the Samsung Galaxy Z Flip 7, 7 Flip FE, Z Fold 7, and all those wonderful Galaxy S25 handsets.

You may also like

• 5 things Samsung could announce at Galaxy Unpacked to ...
• Is the Galaxy Z Fold ready to go Ultra? Samsung just set ...
• How to watch Samsung Galaxy Unpacked on July 9
  

• AOOStar RX 7600XT eGPU runs hot but slower than its full desktop counterpart
• Not all laptops will handle USB4 reverse charging or PCIe 4.0 bandwidth gracefully
• At 61 decibels, the cooling solution trades thermal control for constant ambient noise

As more compact computing solutions crowd the market, mini PC vendors are increasingly turning to external graphics units to offer an upgrade path.

AOOStar is the latest to join this trend with the release of its XG76XT eGPU, built around AMD’s Radeon RX 7600XT and supporting 16GB of GDDR6 memory on a 128-bit interface.

This desktop-grade GPU is based on the RDNA 3 architecture, built using a 6nm process, and features 32 compute units.

Performance limits and thermal design

Marketed as a modular solution for users seeking to enhance visual performance without transitioning to a full desktop, the device’s specifications appear solid on paper.

The graphics processor supports a game clock of 2470 MHz and a power ceiling of 150W in this enclosure, down from the GPU’s full desktop TGP of 190W.

This limitation could affect sustained performance, especially in thermally demanding applications.

However, for those seeking a compromise between mobile convenience and graphical muscle, it may offer a boost, particularly when integrated GPUs fall short for tasks such as editing high-res images or handling multiple 4K displays.

The enclosure includes a custom vapor chamber cooling solution, a full copper heatsink, and a fan housed under a honeycomb-style top grill.

While this setup appears capable of keeping thermals in check, the noise level under load reportedly reaches up to 61 decibels.

That’s not whisper-quiet by any standard, and it could be disruptive in shared or silent workspaces.

AOOStar XG76XT supports both Oculink and USB4, which allow hot swapping and offer up to 100W reverse power delivery, potentially charging your laptop over the same cable.

This might seem convenient for those using a laptop for video editing or for Photoshop, although not all systems will support these features equally.

USB4 relies on PCIe 4.0 lanes, which improve bandwidth over legacy eGPU approaches, but performance bottlenecks compared to internal GPUs are still possible.

On the display side, the XG76XT features one HDMI 2.1 port, two DisplayPort 2.1 outputs, and a Type-C port that supports DisplayPort 1.4 with 15W power delivery.

At ¥3399 (roughly $470), the pricing is not unreasonable for an eGPU with a current-generation GPU.

Yet for anyone looking for the best GPU for demanding creative work or high-end gaming, internal desktop cards in a traditional tower still offer better performance per dollar.

At the time of writing, this device is out of stock and there is no confirmed global release or restock date.

Via Videocardz

You might also like

• These are the best AMD graphics cards on the market
• We've also listed the best mobile workstations around
• Rising prices of DDR4, DDR5, GDDR6 RAM set to cause havoc on PC, smartphone and GPU prices

In case you missed it, Samsung just announced its latest raft of foldables in its latest Galaxy Unpacked presentation, but we actually already had the opportunity to go hands-on with the new phones ahead of the main event.

That's right, we've actually touched and used the impressively thin Samsung Galaxy Z Fold 7, as well as the new Z Flip 7 and Flip 7 FE, Samsung's first budget-first foldable. But wait, there's more... Samsung also announced the new Watch8 and Watch8 classic, packed out with a host of cool new features and a world-first smartwatch antioxidant index.

To hear our initial thoughts, join myself, Hamish Hector and Axel Metz, as well as friend of the show and special guest, Nick Benton from TA Tech Tips.

Make sure to subscribe to our YouTube channel, or if you prefer an audio-only podcast experience you can listen along on Spotify, or Apple Podcasts. There, you'll find all of our previous episode, including our CES and gaming specials.

So, what are you waiting for? Find out who we think the Z Flip 7 is actually for, why we're calling the Z Flip 7 an S25 Ultra competitor why we think this is the most excitable foldable phone launch event from Samsung in a while.

You might also like

• Hands on: I opened the case file on the new Galaxy Z Flip 7 FE, and it was all black and white
• The Samsung Galaxy Z Fold 7 captures the essence of Ultra in a ludicrously thin device, and I still can't believe it
• Hands on: The Samsung Galaxy Watch 8 Classic is a Galaxy Watch Ultra that’s dressed to impress

Samsung's second Galaxy Unpacked event of the year is over – and it was another big one for fans of foldables and smartwatches.

As predicted by the (many) leaks and rumors, Samsung launched a trio of new foldables. Leading the pack was the new Z Fold 7, but we were also treated to more affordable alternatives in the Z Flip 7 and Z Flip 7 FE (which appears to be the Z Flip 6 in disguise).

It wasn't just a big foldables fest either, with Samsung also delivering the Galaxy 8 series. Alongside the new Galaxy Watch 8, we also saw the return of the rotating bezel with the Galaxy Watch 8 Classic, a smartwatch that also does a very good impersonation of a dress watch.

But what else did Samsung reveal (or fail to reveal) at what will likely be its last Unpacked event of 2025? Here are all of the big headlines in one handy place...

1. The Samsung Galaxy Z Flip 7 finally brings a usable cover screen to the party

(Image credit: TechRadar)

In recent years, it’s felt like the Galaxy Z Flip series has been one step behind the competition – but that finally changes with the new Galaxy Z Flip 7.

Samsung has seemingly innovated in all the right places with its latest flip foldable, which boasts a big, beautiful cover display (it’s now 4.1 inches!) and a host of meaningful longevity upgrades like a stronger hinge and a more powerful Exynos 2500 chipset.

The Flip 7’s internal display has been beefed up, too. At 6.9 inches, it’s bigger and wider than the inner screen on the Galaxy Z Flip 6, and as we noted in our hands-on Galaxy Z Flip 7 review, this size increase “makes the phone feel more like a Galaxy S25 Ultra and less like an oddly-shaped foldable in the hand.”

Other upgrades include a larger battery and improved waterproofing, but the Galaxy Z Flip 7 is all about the new display experience – both screens are bigger and better, and as a result, Samsung’s Flip series no longer trails behind the likes of the Motorola Razr Ultra.

2. The Samsung Galaxy Z Fold 7 is an ‘Ultra’ level phone with a price to matchImage 1 of 4

(Image credit: Lance Ulanoff / Future)Image 2 of 4

(Image credit: Lance Ulanoff / Future)Image 3 of 4

(Image credit: Lance Ulanoff / Future)Image 4 of 4

(Image credit: Lance Ulanoff / Future)

The Samsung Galaxy Z Fold 7 was arguably the star of this latest Unpacked event. Far from just being an iterative upgrade, this was an all-round improvement on the already-excellent Samsung Galaxy Z Fold 6, as we’ve described in our hands-on Samsung Galaxy Z Fold 7 review.

There are of course the expected power upgrades, with this phone gaining a Snapdragon 8 Elite chipset and – at the top end – an extra 4GB of RAM. But Samsung has also equipped the Galaxy Z Fold 7 with a new 200MP primary snapper, and increased the sizes of both screens.

Perhaps most importantly, this is by far the slimmest and lightest Z Fold yet, coming in at just 215g (that’s less than the Samsung Galaxy S25 Ultra!), and just 8.9mm thick when folded and 4.2mm thick when unfolded.

But all of this comes at a high price, with the Galaxy Z Fold 7 starting at $1,999.99 / £1,799 AU$2,899 – which in the US is $100 more than last year’s phone and in Australia is AU$150 more. Ouch.

3. The Galaxy Z Flip 6 has been reborn as the Galaxy Z Flip 7 FEImage 1 of 4

(Image credit: Philip Berne / Future)Image 2 of 4

(Image credit: Philip Berne / Future)Image 3 of 4

(Image credit: Philip Berne / Future)Image 4 of 4

(Image credit: Philip Berne / Future)

For the first time, Samsung has launched an FE version of one of its foldable phones, giving potential buyers a cheaper entry point than usual. But if you’ve been keeping an eye on foldable phones then the Samsung Galaxy Z Flip 7 FE might seem strangely familiar to you.

That’s because this is basically just the Samsung Galaxy Z Flip 6 all over again. It has exactly the same dimensions, it weighs the same, and they both have a 6.7-inch 120Hz foldable display and a 3.4-inch 60Hz cover screen.

@techradar

♬ original sound - TechRadar

Their cameras are the same too (though in fairness they also match the standard Z Flip 7’s camera specs), and both phones have a 4,000mAh battery.

The main differences are simply the chipset and RAM, with the Samsung Galaxy Z Flip 7 FE having an Exynos 2400 and 8GB, rather than a Snapdragon 8 Gen 3 and 12GB. So the newer phone is actually less powerful – but it also costs less, starting at $899 / £849 / AU$1,499. Whether that lower price will be enough to tempt buyers remains to be seen.

4. The Samsung Galaxy Watch 8 is slimmer and brighter than before

(Image credit: Future)

It wasn’t just phones at Unpacked – Samsung also unveiled some wearables, including the Samsung Galaxy Watch 8, and this has a new slimmer 8.6mm thick design, and a much brighter screen that can reach 3,000 nits.

It’s not only the hardware that’s been upgraded either, as the Galaxy Watch 8 also has some new software features, such as Gemini on your wrist, and a new Running Coach, which can put you on one of 160 different running plans.

@techradar

♬ original sound - TechRadar

There’s also now an Antioxidant Index, which can give you an idea of whether you’re eating enough fruit and vegetables (though deep down you probably know the answer to that already).

It’s not all good news – our hands-on Samsung Galaxy Watch 8 review noted for example that the battery life is still too short, but the upgrades that have been made sound promising. There's also a new Classic version, for fans of rotating bezels...

5. The Samsung Galaxy Watch 8 Classic is this year’s answer to the Ultra

(Image credit: Future)

Last year, Samsung launched the Samsung Galaxy Watch Ultra, but rather than getting a new Ultra model this year, we’ve got the Samsung Galaxy Watch 8 Classic.

The Classic has a rotating bezel, much like previous Classic models, and our hands-on Samsung Galaxy Watch 8 Classic review was as taken with this as ever. But it also borrows from the Ultra with, for example, a programmable Quick Button. And while this wearable is definitely dressed to impress, it’s not small, coming in at a chunky 10.6mm thick.

Most of the other highlights are on the software side, with the Watch 8 Classic packing on-wrist Gemini, sleep apnea detection, and the same new Running Coach app as the standard Galaxy Watch 8.

At $499 / £449 / AU$899 this is a pricier prospect than the standard Samsung Galaxy Watch 8, but it’s also sure to turn more heads.

6. Google Gemini comes to the Galaxy Watch 8 – and all Wear OS smartwatches

(Image credit: Samsung)

One of the big AI announcements at this Unpacked event was that Google Gemini would be coming to the Galaxy Watch 8 series. And now Google has followed that up by revealing that Gemini is rolling out across all Wear OS watches, including those from Oppo, OnePlus, Xiaomi and, of course, Pixel watches.

This means you’ll be able to talk to Gemini without having your phone nearby, and will let you ask your smartwatch questions using natural language. In theory, you should get speedy, concise answers too, but we’ll be looking to try this out for ourselves before passing judgement.

Google says Gemini will start to roll out to Wear OS watches “over the coming weeks”, as long as your watch is running Wear OS 4 or higher.

7. The One UI 8 rollout is landing soon (for some)

(Image credit: Samsung / Google)

If it doesn’t feel that long since One UI 7 landed, it wasn’t – Samsung’s take on Android 15 arrived as recently as December 2024. But One UI 8 will soon take the baton just six months on, with the new OS pre-installed on the Z Fold 7 and Z Flip 7 – both of which ship on July 25.

So, what’s new with One UI 8? As we covered in our recent rumor roundup, you can expect to see an improved AI audio eraser, plus several other AI improvements. These include a more powerful Gemini Live feature (now also available on the Z Flip 7’s Flex Window) and a better version of Circle to Search, which can tap into Google’s AI Mode for more in-depth answers to your questions.

One of our big unanswered questions, though, is exactly when One UI 8 will rollout to older Samsung devices like the Galaxy S25? Samsung didn't address that at its latest Unpacked, which will leave some worried that it could drag on in a similar vein to One UI 7.

8. The Z Fold 7 ditches the S Pen

(Image credit: Future / Carlos Pedrós)

One more controversial piece of news from Samsung’s second Unpacked this year is that the Z Fold 7 has dropped support for the S Pen. Why? It’s all in aid of that slimmer foldable design.

By ditching the digitizer layer in the display that supports the S Pen, Samsung has been able to get the Z Fold 7 down to just 8.9mm thick when folded and only 4.2mm when unfolded. For some, the lack of S Pen support will be an annoying compromise, but Samsung clearly thinks it’s one that's worth making.

@techradar

♬ original sound - TechRadar

To be fair, the Z Fold series has never come with an S Pen, nor has it had a slot where you could stash one away. It was also only possible to use it on the foldable’s inner display, rather than the cover display. But Samsung may have incurred the wrath of power users who see the Fold series as the natural heirs to its old Note family.

9. Samsung's tri-foldable still isn't ready for prime-time

We had hoped to see Samsung officially reveal its answer to the Huawei Mate XT (above) at Unpacked 2025, but the wait goes on (Image credit: Huawei)

All of the stars were pointing to an official teaser of the Samsung Galaxy tri-foldable, or G Fold as it's been referred to in rumors, at this Unpacked 2025 event. But we didn't even get a five-second trailer.

This was disappointing, as in recent days we've seen the first signs of the tri-foldable appear in One UI 8, along with credible rumors about mass production starting in September. That led us to think that Samsung would at least start building the hype for its next big foldable design.

But there are also likely good reasons why it didn't, including the tri-foldable's expected exclusivity. Not only is it rumored to be prohibitively expensive (early whispers predicted a tag in the region of $3,000-$3,500), there's a chance it may also only go on sale in South Korea and China, initially.

If that's the case, then a reveal at a global event like Unpacked 2025 would make less sense. But it doesn't make us any less excited about seeing it later this year.

You might also like

• Hands on: The Samsung Galaxy Z Fold 7 captures the essence of Ultra in a ludicrously thin device, and I still can't believe it
• I went hands-on with the Samsung Galaxy Z Flip 7, and it matches – or maybe even beats – the Motorola Razr Ultra in one key area
• Hands on: The Samsung Galaxy Watch 8 Classic is a Galaxy Watch Ultra that’s dressed to impress

• Gaming handheld prototype spotted with AMD Ryzen AI Max+ 395 inside
• Strix Halo offers triple the performance of current handheld APU solutions
• GPD could combine console-class graphics with workstation-level AI processing

GPD has made a name for itself by building compact gaming consoles and ultra-portable laptops, often blurring the lines between console and PC.

Its devices, like the Win series and Pocket line, have found favor with enthusiasts looking for performance in tight form factors and, in 2024 GPD expanded its lineup with a dual-screen OLED laptop aimed at creators - we even said the GPD Duo “may be the best mobile workstation ever.”

Now, a newly leaked image spotted by Videocardz suggests GPD could be preparing something far more ambitious - a gaming handheld powered by AMD’s high-end Ryzen AI Max+ 395 APU.

GPD STX Halo handheld prototypeSource：QQ pic.twitter.com/nhSf5xucR3June 28, 2025

Power efficiency concerns

The prototype “GPD STX Halo handheld”, shared on X by HXL (see above), shows a 3DMark TimeSpy score of 10,393, with a graphics score of 10,366 and a CPU score of 10,552. That’s a steep jump from what current Strix Point-based handhelds can achieve.

The chip inside, which is a Ryzen AI Max+ 395 with Radeon 8060S graphics, packs 16 Zen 5 CPU cores and 40 RDNA 3.5 Compute Units.

It’s AMD’s most powerful integrated GPU to date, offering more than twice the GPU hardware found in chips like the Ryzen 7 8840U or 8945HS.

Based on the benchmark figures in HXL's image, GPD's prototype could outperform current-generation handhelds by a wide margin.

Power efficiency remains a concern, however. As Videocardz notes, Strix Halo has a default TDP of 120W, which would normally be too much for a handheld. The architecture can scale down to 20-25W however, and early testing indicates it still retains a performance edge over Strix Point even when throttled.

It’s unclear if this is close to being a final product or just an internal test unit as GPD hasn’t made any official announcements yet.

That said, the prototype points to a possible future where handheld consoles offer desktop-class power, and we’re definitely here for that.

More from TechRadar Pro

• These are the best rugged laptops around today
• And these are the best laptops for graphic design you can buy right now
• The first true dual screen laptop is finally here for productivity fans