Apple’s Irish tax arrangements were scrutinized for more than a decade, and has now, finally, culminated in one of the largest corporate tax recoveries ever ordered by the European Union.

Apple has been fighting a decade-long battle over how much it owed Ireland in corporate taxes. The company was accused by the European Union of exploiting a loophole to avoid reduce its tax liability, despite employing more than 5,000 people in the country.

Now, new filings, seen by The Financial Times, show that Apple paid $17 billion in taxes to Ireland in 2025. Globally, the company had paid $43 billion, which means Ireland accounted this one time for roughly 40% of Apple’s corporate tax bill for the year.

The payment was bolstered by the European Union’s back-tax ruling. The ruling found that Ireland had given Apple “unlawful aid,” resulting in a tax rate of less than 1%.

For years, Ireland opposed the EU’s conclusion.

Timeline of the Apple tax case

The case is the largest corporate tax recovery in EU history. It centers on the “Double Irish” strategy, which let Apple route profits through subsidiaries with no tax residency.

• 1991 and 2007: Ireland’s tax rulings allowed Apple’s Irish subsidiaries to attribute most profits to head offices that were not tax-resident anywhere.
• 2013-2014: EU began investigating.
• August 2016: Against Ireland’s wishes and testimony, the European Commission ruled Apple received illegal state aid, ordering repayment of about $14.2 billion (13.1 billion euros) plus interest.
• 2018: Ireland collected the full amount into escrow while appeals proceeded.
• July 2020: EU General Court annulled the Commission’s decision.
• September 2024: European Court of Justice overturned that ruling, siding with the Commission.
• 2025: Ireland confirmed it received nearly $15.5 billion (14.25 billion euros) from the account’s final closure.

Apple has consistently argued it paid the taxes it owed and complied with Irish and international law. The company has also emphasized that most of its profits were taxed in the United States when repatriated.

Ireland opposed the European Union on this matter for years. The country still markets itself as a low-tax destination but must navigate growing pressure to follow global tax reform rules as part of the EU.

The xTool M2 is priced at the entry level of the company’s fully enclosed machines, but as ever there’s no mistaking it for anything other than an xTool laser engraver. As I got into using the machine, I could see just how finely tuned it is. Whilst it’s lighter weight than many of its more expensive siblings, the laser engraving power, even from the 10W laser in my review sample, was exceptionally good.

Having reviewed these laser engravers for a few years now, there are several essentials I look for in any laser engraver for TechRadar Pro, and one top priority is that it’s fully enclosed. While I might’ve looked at fully open designs earlier on, these days there’s really no excuse for an open-frame design other than cost. When it comes to running a business, safety is obviously paramount, so it makes sense to look for a fully enclosed rather than open-frame design.

Inherent, a London AI lab founded by Google DeepMind alumni, says its AI agent just outperformed much larger models from Anthropic and OpenAI using a fraction of the size.

Of all the startups launched by Google DeepMind alumni, Inherent has gotten relatively little attention. But while better-funded rivals have yet to show the world anything concrete, the London-based team is starting to share what it’s been building.

Just weeks after emerging from stealth with a $50 million seed round, the British startup says its newly released AI agent, Faraday, has outperformed larger, better-known models at a specific task: independently reproducing the findings of published scientific papers without being told the answer in advance.

That may sound like a mere party trick given Inherent’s much loftier goal — building AI that can discover new scientific knowledge and not just verify old results. But paper replication is a standard training exercise for human scientists, too, cofounder and chief scientist Edward Hughes said. “Many PhD students actually start by doing this.”

Beating other AI systems at the task wasn’t the point, Hughes told TechCrunch; how they got there was. “What was most interesting to us about this was not so much the result of beating those frontier agents — which of course we liked — but was actually the way we went about building this.”

Here’s the part that should catch an investor’s eye: measured against Anthropic’s Claude Opus 4.8 and OpenAI’s GPT-5.5 — both much larger, frontier-scale systems — Faraday runs on a comparatively tiny model called Qwen 3.6 that has just 27 billion parameters. (Roughly speaking, “parameters” is a proxy for a model’s size and, typically, its training costs, as well.) Inherent’s bar for success was also higher than simply accuracy. Beyond replicating results, it wanted Faraday to demonstrate “research taste” — an instinct for what experiments are worth running and how to design them well.

Teaching something as intangible as taste is hard, which is where reinforcement learning comes in. It’s a training method that rewards an AI system for good outcomes rather than spelling out rules for it to follow. Rather than training its agents primarily on the study of how science itself is conducted, Inherent leans on this reward-based approach, betting it will generalize better to its longer-term goal of agents capable of contributing across many scientific fields.

“We’re always guided by that north star of building an AI scientist agent and imbuing our agents with taste,” Hughes said. That focus has also shaped what Inherent chooses not to build. Rather than developing its own coding tool, it had Faraday use OpenAI’s GPT-5.5 Codex instead, much the way human scientists lean on existing software rather than building everything themselves, according to the company.

Inherent is also trying to avoid building agents that simply tell users what they want to hear. Instead, Hughes said, the goal is modeled on his favorite kind of teammate — the kind who comes back and says: “I got curious about this, and I went off and I did these experiments. What do you think of these results?”

That collaborative instinct extends to how Inherent operates as a company. Its dozen employees all work in person out of an office in King’s Cross — the once-rundown London neighborhood that Google DeepMind’s presence helped turn into one of the world’s top AI hubs. “We believe that London is the place to be,” Hughes said.

Hughes is bullish on London’s density of AI talent, but he has also added his voice to calls to end “garden leave” — the practice, common in the U.K., of barring departing employees from joining or starting a rival company for months after they resign. It’s a restriction American researchers generally don’t face, giving U.S. startups a head start on hiring talent who’ve left a prior role. “This is a personal view rather than a company view, but I was affected by the garden leave problem,” he told TechCrunch.

Hughes eventually got around that constraint and started Inherent alongside two other DeepMind alumni and a fourth cofounder. The startup isn’t slowing down either. It plans to grow its headcount to “about 20 to 25” by the end of the year. Given its ambitions in world models as well, and with Demis Hassabis’s new role leaving some DeepMind staff unsettled, Inherent’s hiring push could make it an appealing landing spot for DeepMind employees weighing a move.

Pictured from left to right: Inherent co-founders Louis Kirsch, Kaloyan Aleksiev, Tantum Collins and Edward Hughes.

There’s a new publicity stunt tapping into widespread opposition to AI data centers, and it’s using bathroom humor to make a statement.

Former NFL player Jason Kelce, who is also the co-owner of Garage Beer and the co-host of the New Heights podcast with his brother Travis (Taylor Swift’s husband), is the unlikely face of a marketing campaign, launched Tuesday, that puts artificial intelligence on notice. And, well, it’s all about the power of pee.

You read that right, and I apologize in advance for the amount of urine jokes in this article. Tread lightly, I guess.

The now-viral ad opens with Kelce using the bathroom. It cuts to him holding a jar of yellow liquid, issuing a call to action to a happy mob of folks frolicking through grassy terrain, singing, “Let’s pee on computers together and save humanity.” It immediately conjures up that iconic 1970s Coke commercial where hippies harmonize a message of peace, love and unity on a sunny hillside. Except instead of teaching the world to sing, Kelce is asking you to bottle up your pee in a collectible mug and send it to the AI data center of your choice.

If you’re not yet up to speed on the matter, AI data centers are large facilities that store the massive amount of hardware needed to train and run AI models. These warehouses are stocked with high-performance graphics processing units, which provide the power for you to use AI for researching, streamlining video production and, well, creating all the slop your heart desires.

Despite the unserious angle of the marketing stunt taking the piss out of AI (I’m sorry), Kelce’s ad carries a deeper message about the threat data centers pose to our neighborhoods, our resources and our future.

AI data centers guzzle resources

There are nearly 4,000 AI data centers in the country, currently. A database created by environmental activist Erin Brockovich shows where they’re located, with the majority being built in rural areas.

To keep data centers running, an excessive amount of fresh water is required to cool servers so they don’t melt down or explode. Evaporative cooling towers use water to absorb heat generated by servers.

The Environmental and Energy Study Institute published a report in 2025 that says large data centers can consume up to 5 million gallons each day, which is equivalent to the amount of water used by a town populated by 10,000 to 50,000 people.

As massive amounts of potable water are being redirected, people who live near these data centers have growing concerns about how this water extraction will affect their daily lives, farming and basic household needs. Another fear: Who will be stuck paying these ginormous utility bills? Consumers are left holding the bag.

The other resource to discuss is electricity. New facilities, like the recently announced data center from OpenAI and Nvidia that would be built in Pike County, Ohio, are projected to use a lot of compute. The Ports-Pike Technology Campus is projected to consume roughly 8 gigawatts of power, equivalent to the annual energy use of 6 million average US households.

Then there’s Meta’s Hyperion, the 4-million-square-foot AI campus being constructed in Louisiana. It’s flattened an area known for its soybean crops and river cane, and when it’s all built and operational, the Institute on Taxation and Economic Policy says the gargantuan data center will consume three times the power New Orleans uses on any given day.

In other words, more than a piss load of energy.

AI backlash across party lines

Resource consumption, uncertainty over costs and environmental concerns have helped fuel a growing AI backlash.

So, antics aside, Kelce’s trolling ad touches a hot-button issue. The collective sentiment against AI data centers is trending downward. A recent Gallup poll indicated 71% of Americans oppose the construction of new AI data centers in their community.

With the midterm elections a few months away, this negative sentiment has blossomed into a major talking point for both Republicans and Democrats in their campaigns. Another major issue with these data centers is a glaring lack of transparency.

As reported by Brockovich in a blog post, a trend of backdoor deals, elusiveness and secrecy has accompanied the facilities as they pop up all over the country, with the trust of community members tanking. “They’re watching their utility bills climb, finding sick animals they can’t explain, and worrying about the long-term impacts on their health and property values,” Brockovich wrote.

This brings me back to Ohio, an important state for the GOP during the midterms if the party wants to maintain a slim Senate majority. A memo circulated by the National Republican Senatorial Committee warned of a potential defeat for incumbent Sen. Jon Husted unless there’s a fix to the public relations strategy toward AI data centers.

“If voters’ perceptions of data centers are not fixed quickly, the campaign against them will expand far beyond Ohio,” the NRSC stated in the memo, as reported by Axios.

Real-life action versus promotional gimmicks

I now return you to the matter of, ahem, the general public sending urine to an AI data center to cool the machinery of the future.

According to the Cleveland Clinic, urine is 91% to 96% excess water. Heck, there’s a reason Saul Goodman survived the desert in that one Better Call Saul episode. Watch any number of programs featuring survivalist Bear Grylls, and he, too, will tell you this. But I digress.

The marketing ploy here is to equate pee with water and, considering the massive water consumption that keeps AI data centers running, put the public disdain into the type of prank (which you’d normally see in a Jackass episode) to make its point loud and clear.

But you could also argue that it makes no point at all. Ordinary people (those without Super Bowl rings) who have protested data centers IRL have gotten arrested for it. There is real work to be done policy-wise. A trolling ad like this fits into the performative and parasocial attention economy, not into genuine activism.

Remember, this is a commercial for Liquid Death and Garage Beer — a product is being sold here. A collectible mason jar, which is currently sold out, is listed on the Liquid Death website for $18. The product is currently rated five stars, with a single review saying, “the perfect product for shipping your locally sourced urine to your friendly neighborhood data center.”

So, are people actually sending their excreta through the mail? A spokesperson from the United States Postal Service didn’t immediately respond to our request for comment.

That said, USPS rules say urine specimens can be sent by mail, but only for medical purposes such as drug and alcohol testing. To do this right, you’d need to carefully package the fluid in a leakproof container, then place it in appropriate packaging to prevent a messy spill.

Truth is it was never meant to be an action at all. Small print at the bottom of the ad says, “The suits want us to tell you to please don’t actually send your pee.” So yeah, Liquid Death and Garage Beer are really just pissing in the wind.

One thing is certain: Jason Kelce’s reach is strong, and his blue-collar persona can help give this campaign — and the message behind it — legs in rural America, the exact place battling against the spread of water-guzzling AI data centers.

Aaron is a writer for the streaming entertainment team. Previously, he wrote about TV and movies for places like Rotten Tomatoes, Inverse, TheWrap, Washington Post and The Hollywood Reporter. Aaron is also an actor and played a nerd on TV a lot in the ’90s. Seriously.

See full bio

Advertisement

SYSTEMS

Mercury Research blames costly memory and scarce GPUs for 20% processor shipment slide

The processor market is sending mixed signals, with server and mobile shipments rising while desktop CPU volumes decline amid higher system prices. Meanwhile, AMD’s House of Zen has taken market share from Intel across every category.

For Q2 2026, Mercury Research says total processor shipments were lower than in the same period a year ago, attributing this to much lower system-on-chip (SoC) and embedded volumes due to AMD’s declining games console business, plus a large drop in desktop CPU volumes.

Mercury associates the decline in desktop chips with weaker demand for high-end gaming PCs. Although the second quarter is not typically strong for consumer sales, it adds: “We believe that higher PC prices and limited GPU supplies are having a significant impact on end demand for desktop PCs, and thus desktop CPUs, as well.”

Those higher PC prices are the result of increases in the cost of memory components due to a shortage caused by chipmakers prioritizing output of more profitable high-bandwidth memory chips used in AI servers, as The Register has been covering for some time. A shortfall in the availability of consumer GPUs appears to have much the same cause.

According to Mercury, desktop CPU shipments fell by more than 20 percent year on year, although AMD’s decline was smaller than Intel’s. As a result, AMD gained market share, taking nearly 35 percent of desktop chips compared with about 32 percent a year ago.

In contrast, shipments of mobile processors for laptops and tablets were up strongly on the previous quarter, running counter to Mercury’s earlier expectations, although there was only a modest increase compared with a year ago.

The growth followed a sharp increase in Intel’s output, particularly of mobile chips, after two heavily supply-constrained quarters. Intel added millions of units of mobile CPU capacity during Q2, significantly narrowing the gap between supply and demand.

However, AMD’s share of this mobile segment is now up to nearly 29 percent, a significant increase from the 20.6 percent it stood at in the same quarter a year ago.

Server processor shipments also rose, increasing 20 percent year on year and more modestly from Q1. According to Mercury, demand was higher for both datacenter-class CPUs (Xeon SP and AMD Epyc) and chips aimed at networking and storage applications.

Once again, AMD gained market share, accounting for 34.5 percent of server processors compared with 27.3 percent a year ago. Mercury adds that if the calculation included only Intel Xeon SP and AMD EPYC chips, AMD’s share would reach 46.4 percent.

The research firm also keeps an eye on the Arm-based CPU market for PCs and servers, with the usual caveat that its estimates have significant uncertainty as there is no centralized reporting of Arm server or client CPU revenues.

It recorded significant growth in Apple’s Mac products, including the new lower-cost Neo line, while Arm-based Chromebooks also posted strong gains.

Mercury estimates that Arm-based systems captured 15.3 percent of the client market in Q2 2026, up 0.9 percentage points to a record high. In servers, it estimates that Arm’s share reached a record 13.6 percent, up 0.5 percentage points. ®

First look: Widely regarded as China’s homegrown YouTube alternative, video-sharing platform Bilibili has rekindled its global ambitions, launching a new mobile app for international markets. The company has also removed the strict identity verification requirements that often prevented global users from accessing its content.

The international Bilibili app is currently available on Android, with an iOS version expected soon. The company is also reportedly working on an English-language website that will offer content from popular creators.

New job listings on LinkedIn and marketing materials distributed to influencers also suggest the company has started hiring in the US, Europe, Japan, and other key markets.

– bilibili Creator Hub (@bilibili_create) August 19, 2026

According to the job listings, Bilibili is hiring community managers in Los Angeles, London, Mexico City, São Paulo, Istanbul, and Tokyo. Another job listing, for a Singapore-based position, revealed that the company is developing “AI-powered content moderation systems” for the international platform.

In a post on a newly created X account, Bilibili confirmed its international expansion plans but did not reveal a timeline for the official launch. In another post, the company stated that both the Chinese and global versions will offer the same content, but noted that the global version will offer “better localization & easier sign ups” for international users.

– bilibili Creator Hub (@bilibili_create) August 19, 2026

The global version of the Bilibili app uses a pink color scheme, with a small globe icon in the top-left corner of the app icon to distinguish it from the Chinese version.

As part of its global ambitions, Bilibili has been urging popular global influencers, including MrBeast, to upload their content to the existing Chinese-language platform, which the company says has more than 376 million monthly active users.

The company has already taken several steps to make its platform more accessible to international visitors. For starters, it has quietly dropped the identity verification process that required users to provide a passport or ID document to upload content.

– bilibili Creator Hub (@bilibili_create) August 19, 2026

According to Semafor, Bilibili has created a Discord server where it’s marketing its international platform as an additional income opportunity for creators and influencers. The company also reportedly shared a document with global creators describing itself as the perfect platform to connect with “Gen Z Coded, affluent and well-educated” users.

The idea of AirPods with cameras has already made some people uneasy, even though Apple has not announced such a product yet. Online forums have highlighted growing skepticism around the possibility of camera-equipped earbuds, with concerns ranging from privacy to whether such devices could eventually be restricted in workplaces and other public spaces.

Those concerns may become harder to dismiss after a separate discovery reported by MacRumors offered a more detailed look at how Apple’s camera-equipped AirPods could potentially work.

According to MacRumors, forum member mactracker uncovered a hidden macOS framework called “AccessorySensorManager,” which appears to contain references to sensors capable of collecting visual and environmental information from connected AirPods. The findings do not confirm a final product, but they offer a glimpse into the kind of technology Apple may be testing.

The privacy question arrives before the product does

The prospect of putting cameras into earbuds has already triggered a skeptical response online. Unlike a smartphone or even a pair of smart glasses, earbuds are small, discreet, and often worn for hours at a time, making it far more difficult for people nearby to know when a camera might be active.

That creates an obvious challenge for Apple. Camera-equipped AirPods could make AI assistants considerably more useful by allowing them to understand what a wearer is seeing or provide information about their surroundings without requiring a phone to be pulled from a pocket.

Convenience, however, may not be enough to overcome the social discomfort surrounding always-worn cameras.

Users on Reddit noted concerns that such devices could face restrictions in offices, schools, or other sensitive environments. Similar unease appeared in discussions surrounding the MacRumors report, where readers questioned whether the potential benefits justified adding cameras to a product designed to disappear into everyday life.

MacRumors’ findings show what Apple may be working on

The framework uncovered by mactracker suggests each earbud could contain a camera sensor capable of capturing synchronized RGB images at up to one megapixel, according to MacRumors. The system appears to focus on still-image capture rather than conventional photography or video recording, potentially feeding information into Visual Intelligence and other AI features.

The report also references active and passive capture modes. One could potentially be triggered through Siri, while another may collect lower-resolution environmental information related to nearby speech, sound changes, posture, head movement, and other contextual signals. The code also references a hardware indicator that could visibly signal when image capture is taking place, addressing at least part of the privacy problem.

None of the findings confirm when, or even whether, Apple will release camera-equipped AirPods. MacRumors notes that the code could relate to an earlier design, while Bloomberg has previously reported that Apple is working on a camera-equipped AirPods model that could arrive in 2027.

Apple may have figured out how to make AirPods see. The harder task will be convincing everyone else that they should be comfortable with them looking.

You can keep Instagram from using your activity outside the app to influence the ads you get.

Turn off personalization from other businesses

Novak Bozovic for Engadget

Restrict Instagram’s access to your phone and accounts

Novak Bozovic for Engadget

Limit what Meta can learn from AI chats

Novak Bozovic for Engadget

Sony is expected to introduce its next Xperia phone in just a couple of days, but a new leak may have already revealed what the upcoming Xperia 10 VIII will look like.

An image shared by Android Headlines shows what is claimed to be the Xperia 10 VIII ahead of Sony’s August 25 event in Japan. Sony has only confirmed that a new “Xperia Product” is coming, though recent leaks have consistently pointed toward the Xperia 10 VIII. If the image is accurate, Sony does not appear to be changing much on the outside this year.

The Xperia 10 VII resemblance is difficult to miss

At first glance, there is very little separating the leaked phone from last year’s Xperia 10 VII. Sony appears to be retaining the horizontal camera module, with two rear cameras, an LED flash, and the company’s branding sitting alongside them.

The phone is shown in a pale green finish, while the right side appears to retain the familiar volume controls and recessed power button. The latter would suggest Sony is once again using a side-mounted fingerprint scanner. Only the rear and part of the side are visible, so the leak does not reveal whether Sony has made any noticeable changes to the display or front-facing design.

The specifications may be familiar too

The design is not the only thing that could carry over from the Xperia 10 VII. A recent Geekbench listing for the Sony XQ-GH54, believed to be the Xperia 10 VIII, showed a Snapdragon 6 Gen 3 processor paired with 8GB of RAM.

That is the same chipset used inside last year’s model. The device recorded 871 points in Geekbench’s single-core test and 3,028 points in multi-core, suggesting there may not be a major performance upgrade this generation.

As we reported earlier, Sony’s event begins at 11 a.m. Japan time on August 25, or 10 p.m. ET on August 24 for U.S. viewers. Xperia phones are still sold in Japan and select markets such as the UK and Germany, but Sony has scaled back its smartphone presence considerably and no longer officially sells Xperia phones in the U.S. A stateside release for the Xperia 10 VIII therefore looks unlikely.

Sony should fill in the remaining gaps around the cameras, battery, pricing, and availability at the event.