Remember Anthropic’s AI going rogue? Then OpenAI’s

 Remember Anthropic’s AI going rogue? Then OpenAI’s AI also did something similar and hacked other services? And then Meta said its Muse AI did something similar? Well, now it’s the turn of Chinese AI models that are regularly sending Wall Street into a tizzy.

Moonshot's Kimi K3 is the second Chinese AI model now reported to have escaped its testing environment. At this point, it feels like a race to the “my AI is more dangerously powerful.” Regulators are watching this PR battle rather closely, by the way.

How about a pre-AI blast from the past? Before Google turned the web into a tidy, searchable index, finding anything online meant wandering. This week's tech news pulls in two directions at once, and the gap between them is what makes it strange. On one end sits nostalgia: a look back at the messy, human-curated internet we traded away for convenience.

It was gloriously disorganized before Google rationalized the mess into something you could actually search, and in doing so it quietly retired a version of the internet that ran on human attention instead of ranking algorithms. Of course, Google also injected AI in each corner, and did social media sites. And that brings us to our final story. The purge on AI slop has actually caught social media influencers as collateral damage. What a mess!

Send us feedback at editorial@digitaltrends.com.

Hackers Figure Out a Trick to Steal Bitcoin From Cold Wallets

Hackers Figure Out a Trick to Steal Bitcoin From Cold Wallets, Grab $110 Million

Futurism · 2 hours ago
by Victor Tangermann · Future Society

Hackers made away with more than $100 million worth of Bitcoin from thousands of supposedly secured “cold” wallets hosted by Canada-based company Coinkite.

As Bloomberg reports, the hackers managed to infiltrate the wallets despite clients having a physical hardware key associated with their accounts, which was supposed to provide nearly impenetrable security.

“The moment it loaded I knew I was screwed because I saw red lines for withdrawals,” one of the victims, Johnathan Goodman, told Bloomberg. “Between 9:36 and 9:43 pm on July 29th, all three of my wallets were completely drained.”

Crypto insights company Galaxy Research estimated that around $110 million worth of Bitcoin had been drained from around 5,000 wallets last week, a figure that grew to at least 7,300 by Monday.

The incident highlights persistent lapses in security plaguing the largely unregulated cryptocurrency industry.

In the case of Coinkite, it was a particularly egregious lapse in security. The firm warned its customers on July 30 that hackers were exploiting a software bug that allowed them to reconstruct wallet “seed phrases,” which are sequences of random words that act as a master key to “cold” — or offline — wallets.

Coinkite promised in an email to Bloomberg in a followup story that it was racing to get a full picture of the embarrassing situation, saying it was working on “helping affected customers.” However, the company refused to estimate the scale of the losses, vowing to conduct a “post-mortem” at an unspecified future date.

“We’re not in a position to independently confirm total losses or validate the specific figures being reported by third parties,” the company told Bloomberg. “We won’t speculate on a number we can’t verify directly.”

The company also kicked off an “ongoing ecosystem-wide security audit” which has revealed “numerous critical bugs in key software systems across the ecosystem using frontier AI models.”

The cryptocurrency firm has since gone into full damage control mode as it investigates the major slipup.

In an “update on customer data retention” published on its website today, the company said that “due to legal obligations arising from the security incident, including the preservation of records that may be relevant to ongoing and anticipated legal proceedings, we have temporarily suspended our automated data-blanking process.”

“This means that customer records that would otherwise have been blanked under our standard schedule will be retained until further notice,” the company wrote.

More on crypto: Trump Boasts That He Can Profit Off Presidency as Much as He Wants: “I Found Out That Nobody Cared”

The post Hackers Figure Out a Trick to Steal Bitcoin From Cold Wallets, Grab $110 Million appeared first on Futurism.

OpenAI Tried to Hire Influencers to Spread Love for Its Products

 OpenAI Tried to Hire Influencers to Spread Love for Its Products, But It Backfired Horrendously

"People will literally sell their soul and their planet down the river for a nice free hotel room."

By Frank Landymore

Published Aug 6, 2026 2:58 PM EDT
Add Futurism(opens in a new tab)More information

OpenAI / Futurism

Sign up to see the future, today

Can’t-miss innovations from the bleeding edge of science and tech

Email address Sign Up

OpenAI’s attempt to boost its image by hiring influencers collided against a harsh reality: that it’s already become a fundamentally uncool company.

The Verge details how influencers recruited by the ChatGPT maker were flown out to its “first ever brand trip” at a nature retreat outside New York City. But rather than creating buzz for OpenAI, gaining widespread attention, or at least showcasing its products, the content that came out of its “Summer Club” campaign sparked fierce backlash against the influencers, who were branded “sellouts.”

“People will literally sell their soul and their planet down the river for a nice free hotel room,” one TikTok commenter quoted by the reporting seethed.

“Wow! Beautiful greenery! They should put a data center there!” another jeered.

It’s not uncommon for companies to rely on engagement bait — and even rage bait — to gain attention. But this didn’t appear to be an intentional maneuver on OpenAI’s part.

Most of these posts and videos barely get any attention, racking up only, at best, a few hundred likes and a few thousand views. But videos criticizing the OpenAI influencers — or “creators,” as the company prefers to refer to them — have racked up hundreds of thousands of views, including from other influencers who didn’t take the AI bag.

“They’re all a bunch of f*cking phony sellouts,” Meredith Lynch, a comedian known for her TikTok content, said in one viral call-out video.

As the influencers, most of them women, toted around branded merch and reveled in their bucolic surroundings, many couldn’t help but point out the hypocrisy of laundering the image of an industry whose data centers have perilous effects on the environment. The choice to select mostly women feels noteworthy, given that, as the Verge‘s reporting observes, women tend to view AI more negatively than men do.

What’s particularly baffling was how relatively little content came out of the trip, according to The Verge, with almost none of it actually about OpenAI’s chief product, ChatGPT. Instead it was “largely aspirational lifestyle content focused on the free swag and pretty setting” — with no clear idea on how it’s supposed to reflect on the company, other than the fact it’s willing to comp expensive trips for a small group of extremely vain people.

The company has made a big push recently to gain some cultural cachet. A few weeks ago, it launched a new line of clothing in a minimalist, trendy style — or at least trendy for five years ago, as if its fashion department had a knowledge cutoff date like its AI models. That, just like its influencer brand trip, was also met with mockery.

In a statement to The Verge, OpenAI defended its influencer marketing scheme and said the event was meant to be educational.

“Creators are an important part of our community and how people get information and learn about our products today,” a spokesperson told the outlet. “We welcome healthy debate as AI becomes more prevalent, and we value creators who choose to engage with us, attend our events, ask tough questions, and learn alongside everyone else. We’ll continue to value them, just as we do traditional media and other marketing partners.”

More on OpenAI: Sam Altman’s Parenting Strategy Sounds Low Key Horrifying



Frank Landymore
Contributing Writer


I’m a tech and science correspondent for Futurism, where I’m particularly interested in astrophysics, the business and ethics of artificial intelligence and automation, and the environment.

Grokipedia Appears to Have Completely Broken

Grokipedia Appears to Have Completely Broken and Nobody Noticed
"As far as we can tell, no entry has changed in more than three months."

By Victor Tangermann

Published Aug 6, 2026 1:54 PM EDT
Add Futurism(opens in a new tab)More information


Shutterstock / Futurism

Sign up to see the future, today

Can’t-miss innovations from the bleeding edge of science and tech

Email addressSign Up

We certainly wouldn’t blame you for having completely forgotten that last year, Elon Musk tried to reinvent Wikipedia because he thought it was too woke.

In October, his AI startup xAI launched Grokipedia, named after its AI chatbot. The encyclopedia stole human-written and edited pages from Wikipedia wholesale and ran them through a large language model, which also fielded user-submitted edits, to better reflect Musk’s contorted worldviews.

The site was hardly an exemplary of objectivity. It was quickly found to heavily cite an infamous neo-Nazi forum as a source of information, desperately tried to paint Tesla’s Cybertruck as a successful and desirable electric pickup truck, and validated Musk’s favorite conspiracy theories. But it was a major cultural moment, garnering countless stories in the media and discussions on podcasts over Musk’s attempt to control the spread of information.

Perhaps unsurprisingly, the desperate attempt to twist the largest online encyclopedia in human history into an extension of an almost-trillionaire’s bigoted mind appears to have ground to a screeching halt — and, damningly, barely anybody seems to have noticed.

According to a new investigation by Lawfare, Grokipedia appears to have entirely stopped updating itself sometime earlier this year. Reviewing suggested edits from humans has also ground to a halt.

“As far as we can tell, no entry has changed in more than three months,” Lawfare concluded of the site, which is now owned by SpaceX since it merged with Musk’s AI efforts.

Worse yet, the site’s edit log appears to have broken completely as well, making it impossible to track which user-suggested changes were “accepted” or “rejected” by the AI.

By January, the Tow Center for Digital Journalism had already observed that xAI’s Grok was making more edits to its own articles than acting on human-suggested ones, suggesting a bottoming out of interest for the project.

Since then, it has turned into a ghost town, a completely redundant and misinformation-filled copy of Wikipedia, which remains one of the last bastions of human-edited and verified sources of information on the internet.

Even when Lawfare submitted an “uncomplicated factual update” about SpaceX having launched its IPO in June — an irrefutable fact that any allegedly “open source, comprehensive collection of all knowledge” should probably include — the Grokipedia entry remained stagnant.

Even pages that are purportedly drawing millions of views aren’t updating. The most recent approved update Lawfare could locate — to the entry about OpenAI’s ChatGPT — dates back to April 24.

As of today, Lawfare found that that there are 13,002 suggestions that “sit unresolved, trapped ‘in review.'”

“Altogether, what we found was that the ‘live’ change feed is gone, past decisions were retroactively reverted without explanation, new decisions are not being issued, and article text appears frozen (and, in some cases, outdated),” the publication concluded. “We do not know whether xAI is preparing a major revision or has altogether abandoned active development of the project.”

It’s a precarious situation. Despite having been abandoned, Similarweb data suggests Grokipedia received over six million visits in June alone. Hundreds of thousands of sites are also still citing Musk’s encyclopedia, which could spread misinformation even further.

Futurism has reached out to SpaceX for an update on the project, but considering the reputation of Musk’s companies and their refusal to interact with the media, we’re not holding our breath for a response.

More on Grokipedia: Elon Musk Is Not Beating the Allegations: Grokipedia Cites a Hardcore Nazi Website 42 Times



Victor Tangermann
Senior Editor


I’m a senior editor at Futurism, where I edit and write about NASA and the private space sector, as well as topics ranging from SETI and artificial intelligence to tech and medical policy.

Walmart reports slowest sales growth in 6 years

  

Walmart reports slowest sales growth in 6 years

Niv Bavarsky

Walmart’s same-store sales growth wobbled out of last quarter with the power of the last remaining shopping cart during a Sunday afternoon grocery trip. US sales growth inched up just 2.6%, the company’s slowest growth in almost six years in what was otherwise a positive earnings report. With Walmart often seen as a bellwether for the broader US economy, economists are worrying if consumer spending might finally be starting to dwindle.

In response to the news, the retail giant’s stock dropped over 9% yesterday. Much of the blame was placed on Walmart’s health and wellness business, which suffered from new drug price caps, including on GLP-1s. But, even excluding those losses, sales growth clocked in at 3.4%, just shy of Wall Street’s expected 3.5%. Last year, same-store sales grew 4.6% during the same time.

So what’s going on?

• CFO John David Rainey told investors that the consumer environment has softened in the last quarter, pointing to gas prices jumping past $4 a gallon in July.
• Average hourly earnings have slowed, consumer sentiment sank this month, and folks’ personal savings rates are at their lowest levels since 2022.
• Spring also brought higher-than-usual tax refund checks that have likely dried up.

Some relief might be on the way. Walmart said that it’s investing its tariff refund—a record $2.9 billion—back into lowering prices. Last quarter, the retailer said it eased prices on 11,000 items, an increase from 7,200 products in Q1.

What about the competition? Retail sales across the board were down 0.6% in July compared with June, according to the most recent Census Bureau data. But some areas are doing better than others: Target said its back-to-school sales were up by almost 20% compared with last year, while Home Depot said that sales of items needed for small projects, like painting and gardening, have replaced large-scale remodels.

You're about to get a major Dolby Vision upgrade

 Hisense TV owners, listen up! You're about to get a major Dolby Vision upgrade

News
By Lewis Empson
Published 3 hours ago

Dolby Vision 2 is finally launching after a year of speculation

(Image credit: Dolby / Hisense)

Share this article
0
Join the conversation
Add us as a preferred source on Google
Newsletter
Subscribe to our newsletter

After almost a year of anticipation and uncertainty, Dolby Vision 2 is finally launching, although you'll need one of the supported TVs to use it.

Dolby surprised home cinema fans with the announcement of Dolby Vision 2, the latest iteration of its dynamic HDR format, just before IFA 2025. With it came a buzz of excitement, but also many questions, most of which have been periodically answered throughout the year.

Now, Dolby has officially announced which TVs will be the first to access the new HDR format, and they all come from launch partner Hisense. We've also seen Dolby Vision 2 running on TVs from TCL and Philips at past press demos, although this launch is strictly Hisense for now.

Supported models include the UX, UR9 and UR8 RGB Mini LED TVs, and U7 Pro (also known as the U7 in the US) Mini LED TV. Interestingly, these models all support the higher-tier Dolby Vision 2 Max format.

The update is only coming to the Vidaa and Google TV versions of these models. Neither Hisense nor Dolby have confirmed if Dolby Vision 2 will come to older TVs yet.

LAPD Abandons Flock Contract After Making a Horrifying Discovery

 LAPD Abandons Flock Contract After Making a Horrifying Discovery

"This contract is not being renewed because of serious concerns around civil liberties and civil rights issues."


By Joe Wilkins


Published Jul 13, 2026 4:48 PM EDT
Add Futurism(opens in a new tab)More information


Illustration by Tag Hartman-Simkins / Futurism. Source: Shutterstock




Sign up to see the future, today


Can’t-miss innovations from the bleeding edge of science and tech

Email addressSign Up

Rarely has the Los Angeles Police Department found a police toy it doesn’t love. The third-largest police force in the US has major Palantir contracts in place, uses sketchy apps to track minority communities, and routinely sends drones to spy on peaceful protestors.

So when the LAPD ends a contract with a surveillance company because it thinks the technology is too dangerous, you know there’s a problem.

Over the weekend, the police department announced it was abandoning its contract with Flock Safety, the controversial company blanketing the United States in AI-integrated license plate cameras (ALPRs). Though its three-year contract with Flock is just about up, the LAPD says it won’t renew the partnership, citing residents’ concerns over privacy and civil rights, ABC reported.

“This contract is not being renewed because of serious concerns around civil liberties and civil rights issues, particularly around privacy and the data that is being collected from these cameras,” LAPD’s chief information officer Dean Gialamas told ABC in a statement. “The LAPD had to make a difficult decision, in this case discontinuing using Flock services until we can get those data, privacy, security and sharing concerns ironed out through a contractual relationship.”

The news comes just days after our sister publication The Drive‘s Joel Feder reported how he was suddenly swarmed by armed cops while test driving a Range Rover with his wife in Minnesota. Officers said they had been tracking him for days using Flock’s AI-integrated cameras, which had erroneously tagged his car as stolen.

After the latest Flock contract news broke, 404 Media revealed that a July 10 audit by the LAPD Office of the Inspector General caught the department’s ALPR cameras generating 161 false stolen-vehicle alerts in just two months — each one ending with officers pulling over an innocent driver. Factoring in 337 alerts which “resulted in the recovery of stolen vehicles,” the LAPD’s cameras carry an error rate of 32.3 percent, effectively giving officers a one-in-three chance at pulling an innocent person over.

“During the review period, officers acknowledged 161 alerts as accurate license plate matches; however, subsequent investigations determined the vehicles were not stolen,” the report reads. “In addition to creating an inconvenience for vehicle owners, these inaccuracies can affect individual liberty interests, erode public trust, and potentially create substantial legal and financial liability concerns.”

The OIG report notes that officers are required to “attempt to verify all records prior to conducting an investigative stop,” though it didn’t disclose how many of the 161 wrongful stops were the result of an officer’s failure to do so. Instead, investigators blamed license plate records it says were “not updated in a timely manner,” which “may” lead cops to “act on inaccurate or outdated information, increasing the risk of unnecessary enforcement actions, including vehicle stops and wrongful detentions, or a confrontation with serious consequences.”

As 404 flagged, the LAPD approaches potentially stolen vehicles as “high-risk” stops. For the LAPD, these stops require drastic shows of force, including “calling for back up, air support, and a supervisor and ordering the suspect out of their vehicle,” the report reads.

It doesn’t take a large language model to figure out that this is an incredibly dangerous scenario for Black people in Los Angeles, who account for 19.5 percent of those killed by the LAPD, despite making up only nine percent of the city’s population.

Whatever way you slice it, it seems one of the most notorious police departments in the US has finally met its match.

US Companies Are Realizing That Chinese AI Models Are Way Cheaper

 US Companies Are Realizing That Chinese AI Models Are Way Cheaper, Ditch American Ones

"'Hey, we don't need the best model, we can use the faster, cheaper models.'"


By Frank Landymore


Published Jul 13, 2026 4:01 PM EDT
Add Futurism(opens in a new tab)More information


Illustration by Tag Hartman-Simkins / Futurism. Source: Shutterstock

Sign up to see the future, today

Can’t-miss innovations from the bleeding edge of science and tech

Email address Sign Up

As corporate AI bills spiral out of control, many companies are beginning to ask themselves a simple question: why pay a pretty penny for the US’s leading AI models when Chinese ones are far cheaper?

Major companies like DoorDash, Airbnb, and Siemens are adopting Chinese AI tools, the Financial Times reports, attracted not only by their lower costs but their “open-weight” approach that allows them to be molded to each company’s particular needs.

According to data from OpenRouter, a platform that provides all-in-one access to major AI models and tracks their usage, leading Chinese models from DeepSeek and Z.ai have overtaken US equivalents like Anthropic’s Claude and OpenAI’s ChatGPT. Cost-cutting, it seems, trumps all geopolitical rivalries.

Chinese models are “the elephant in the room,” Eugene Cheah, CEO of the AI platform Featherless AI, told the FT. “Enterprises are starting to realize, ‘Hey, we don’t need the best model, we can use the faster, cheaper models.'”

US-based AI models have frequently been seen as the most advanced, but that perception is shifting. The release of GLM-5.2 from the Chinese startup Z.ai last month caused a stir in Western tech circles, as major Silicon Valley figures hailed it as capable or nearly as capable as US systems despite being significantly cheaper to use.

Cheap Chinese AI couldn’t be coming at a more opportune moment. The corporate world, wooed by AI companies’ promises of supercharging their productivity, has spent the past year deploying AI across its workforces, and many are being put off by the costs. One organization reportedly blew through $500 million in a month on Claude usage fees. That’s an extreme outlier, but recent research from the Ramp AI Index found that the businesses most dedicated to AI are spending around $7,500 per employee every month on AI.

Considering the culture around AI, it’s no surprise why: some companies like Meta mandate their employees use AI systems as much as possible, factoring it into their performance reviews. Software engineers, now expected to produce more work than ever, often run multiple AI agents at the same time to complete tasks in the background.

If companies are unwilling to crank back the AI knob, then the next best choice is to look for cheaper models. DoorDash cofounder Andy Fang said on X last week that it was saving a lot of money by having “lower-level work” performed by a model from the Chinese startup Moonshot AI. San Francisco startup Lindy has completely ditched Anthropic’s AI tools in favor of DeepSeek’s latest V4 models.

“Enterprises have an incentive to shift some of their workload to cheaper models. Why would you pay a premium for Anthropic, OpenAI models when for a lot of the workloads you need, the Chinese models are generally workable?” Sam Bresnick, a research fellow at Georgetown University’s Center for Security and Emerging Technology, told the FT.

But cost isn’t the only consideration: many Chinese models are “open-weight,” meaning their parameters or values are entirely visible to the user. That allows an organization to mold a model to its specific needs — and from a cybersecurity perspective, have more control and insight into how it might process sensitive company data.

For foreign companies disillusioned by US leadership, the choice is even easier to make. There’s less faith in the US as stewards of AI, especially after the Trump administration suspended access to Anthropic’s Mythos model overseas.

“The Mythos ban was certainly the most tangible event, and people having their access revoked,” Aidan Gomez, CEO of the Canadian AI group Cohere, told the FT. “It exposes the risk of relying on any one single entity for any of your workloads.”

More on AI: OpenAI Is Shutting Down Its Browser That Was Supposed to Change Everything

The EU’s answer to Starlink moves one step closer to launch

 NextTech News

The EU’s answer to Starlink moves one step closer to launch


Copyright European Space Agency
By Indrabati Lahiri
Published on 10/08/2026 - 14:45 GMT+2
Share CommentsAdd Euronews on Google
Rising geopolitical volatility and conflicts such as the Russia-Ukraine and Middle East war have made it crucial for Europe to invest heavily in its own home-grown satellite network.

The European Union’s attempt to launch its own satellite program have achieved a significant milestone.

The SpaceRISE consortium and the European Commission have now joined forces, signing an implementation agreement for IRIS².

This will be the EU’s flagship secure satellite programme, which will provide government-backed and encrypted communication services for EU institutions, critical infrastructure and emergency responders. It will also provide quantum-secured defence layers and high-reliability routing.

This move comes as several industries continue to rely on Starlink significantly for services like airplane Wi-Fi and drone operations in Ukraine. This could potentially allow the EU to benefit considerably from having its own domestic satellite programme to counter Starlink’s widespread influence.

The recent agreement adds another 66 satellites to the programme, bringing the total number 348. Out of these, 330 will be in Low Earth Orbit and 18 will be in Medium Earth Orbit.
RelatedSpaceX posts smaller-than-expected loss despite scrutiny over AI spending
SpaceX rocket stage believed to have slammed into Moon, NASA to collect data

The programme’s first launches are expected to begin in 2029, on a progressive basis.

Its primary users will be European governments and key players from surveillance, defence operations, and critical commercial sectors like transport and energy.

A European satellite network

Rising geopolitical volatility and ongoing conflicts such as the Russia-Ukraine and Middle Eastern wars in recent years have made it crucial for Europe to invest more in digital sovereignty and domestic resilience.

As such, instead of trying to be yet another consumer broadband service, like Starlink attempting to compete with networks like T-Mobile and AT&T, IRIS² is built for defence, governments, critical services and security users.

These are the industries that can best help the continent build true resilience. This is particularly important during times of conflict or natural disasters, when cell towers and internet services can be impacted, leaving satellite connectivity the only real option for communication.

However, overreliance on Starlink, or any other foreign company can also cause a host of geopolitical and governance issues for the EU. This is especially following rising EU-US tariff and other tensions in recent years, following the second Trump administration.

SpaceRISE’s members Eutelsat, SES and Hispasat will work with domestic manufacturers on the satellites and ground infrastructure for IRIS², while the European Space Agency will take care of the technical development and validation.

Can open-source beat OpenAI?

Can open-source beat OpenAI?

Former Hugging Face executive Tiezhen Wang explains how China’s open-source strategy is reshaping the AI race.


By Kinling Lo
+

By KINLING LO
+
15 JUNE 2026
TRANSLATE




As the U.S. and China battle for artificial intelligence supremacy, a fundamental divide in engineering philosophy could determine the winner.

While American pioneers like OpenAI and Anthropic favor a closed-source approach — keeping their proprietary model code locked behind a commercial interface — Chinese AI labs are aggressively releasing open-source models. This strategy allows developers to download, inspect, and deeply customize the underlying code for free, rather than remaining dependent on an American tech giant’s ecosystem.

Tiezhen Wang, former head of the Asia-Pacific ecosystem at AI community collaboration platform Hugging Face, has observed this trend over the years. Before he moved on from Hugging Face in May, he assisted AI labs in the region in launching open-source models, and helped researchers make their models easier for developers to use.

At a Rest of World virtual event, Wang spoke about the history of open-source models, how Chinese AI labs monetize despite not charging for their models, and the debate over model distillation and intellectual property.
Tech reporting from beyond Silicon ValleySee the full picture of how technology is shaping the world. Twice a week, straight to your inbox.


The conversation has been edited for length and clarity. A recording of the live Rest of World event is available here.
What is the role of open-source models in the China-U.S. AI competition?

A lot of people talk about the AI competition. But in the open-source world, we hold a collaborative mindset.

Many open-source releases from the Chinese labs are helping U.S. labs. For example, the reinforcement learning training algorithm from DeepSeek is becoming the default setting for many U.S. research labs. Many Chinese open-source weights are running on U.S. hardware.

It is like helping each other, not competing like a zero-sum game. We can both be winners if we are growing the pie together.
OpenAI and Anthropic have accused Chinese AI companies of distilling their models. How do you view this as an open-source advocate? Is this a challenge for Chinese AI labs?

Distillation is a neutral word in the research world. It’s like I’ve read a book, and I’m telling someone what the book is about. The other person also understands what the book is saying. That’s basically distillation.

It has been done a lot in research labs. I don’t think distillation by itself is anything wrong, and I believe people in the U.S. are distilling each other as well. Recently, we have seen Elon Musk admit that xAI was distilled from OpenAI.

It is well known that Anthropic and ChatGPT are crawling the internet, getting all kinds of information. So it’s interesting to see that those who did not generate knowledge are trying to stop others from reusing that knowledge.

All AI-generated content should have zero copyright, otherwise people who have computers can abuse and generate all kinds of combinations and copyright everything.
How does monetization differ for open-source and closed-source models?

If you open-source your model, it’s hard to make money directly. But you can still make money.

For example, China’s Kimi released their model for free, but their application programming interface and subscription are still in huge demand because they have the best infrastructure support.

Once a model is released as open-source, people have to spend engineering hours to run it. So, on day zero, the research lab that launched the model has an advantage. That’s one way they can make money.

Another way is, for example, Kimi could release a fine-tuned model as open-source, but keep the base models to themselves, which they can sell.

If we want to think about why a research lab would open-source its models, it is so that they can build their branding. When you first start a lab, why would researchers want to work for you? It’s very hard to acquire top talent. But if you have good open-source models, everyone knows you did some great work.
Why are some Chinese AI labs pulling back from open-source releases?

Some models are changing their licenses. For example, Minimax changed the license to basically say that if you use this model to make money, you have to pay. This is a very common practice in the open-source world, especially if you want to prevent free riders, like cloud users.
Related Stories

China builds a rival satellite constellation as SpaceX goes public

Big Tech, big cons: Scammers are hiding in the apps that make your life easy

The Great AI Divide: Navigating U.S. and Chinese dominance

As the world embraces EVs, the U.S. hits the brakes

Cloud providers could run an open-source model for free and generate profit without needing to share profit with the research lab. This is unfair.

So basically, the trend of changes is that if you are an individual user, you can use my model for free forever, but if you are a cloud provider generating revenue by serving my model, you have to give me a portion of your profit. I think this is fair and actually a sustained way of supporting open-source.
Do you think more Chinese AI labs will go closed-source because of monetization problems?

I’m concerned. If they do not find a way to monetize their research, this is a real risk. If they can have money and can keep doing open-source, it’ll be great for everyone.

I think the capital market is coming to help. If you look at the stock price of China’s Zhipu, it is at 10 times growth already. This will definitely help them acquire more compute, talent and data, and generate better models. These investments can keep these labs on the table for a longer time.
If you were advising a U.S. startup today, would you tell them to build on American models, Chinese models, or whichever open-source model performs best?

It’s a well-known strategy that nowadays in the U.S., if you want to launch a startup, the first step is to find a model that has the best product-market fit. That is more important than your technical decisions. If you don’t have a product that deeply satisfies strong market demand, your startup will fail. These models would be closed-source.

But once you have this closed-source model in place, you start to have the first group of users, and you are accumulating data. You know how your product is being used. You can later use the data you generate to train your model. Then companies will start to consider switching to an open-source model. Eventually, you will save maybe a hundred times on tokens.
You just returned from a trip to China. Do you have any new observations about the country’s AI development?

I had a very interesting trip talking to a few research labs and companies generating tokens. I do feel that the whole Chinese market is quickly maturing.

The U.S. is definitely doing a great job with exploring what AI can do, but the token in their way is more expensive even for big companies to afford. For example, Uber burned their entire year’s tokens within four months. Microsoft also said it felt the tokens were more expensive than they’d expected.

In China, it’s kind of the other way around because it has a bunch of open-source models that are not that expensive to use.

As soon as Chinese open-source models cross a usable point, usage will grow exponentially. What I’ve seen is that all Chinese internet companies started tokenmaxxing, which is basically giving their employees unlimited tokens just to see what they can do. I have heard that a few big Chinese tech companies are forcing their employees to be AI-native and stopping them from doing normal things like writing a document.

This kind of approach to AI adoption is way faster than what the U.S. has done. I won’t be surprised if over the next year or two, you see a lot of very interesting AI use and adoption in China.


Kinling Lo
+ is a reporter at Rest of World covering China.

A Canadian payment giant is being sold to U.S. private equity - No thank you

 A Canadian payment giant is being sold to U.S. private equity. Is your digital privacy at risk? | CBC News

A Canadian payment giant is being sold to U.S. private equity. Is your digital privacy at risk?
Moneris handles roughly 1 in 3 payment transactions across Canada


Brock Wilson · CBC News · Posted: Aug 14, 2026 1:00 AM PDT | Last Updated: 1 hour ago


Listen to this article
Estimated 6 minutes

Moneris payment terminals, like this one seen in 2018, are used widely across Canada. The company is being sold to a U.S. private equity firm, pending regulatory approval. (Anis Heydari/CBC)

A payment processing company responsible for approximately one in three payment transactions across Canada will soon be owned by an American private equity firm.

The Royal Bank of Canada and Bank of Montreal announced earlier this week they would be selling jointly owned Moneris, one of Canada's largest commerce solutions providers, to Francisco Partners for $2 billion.

So far, it's been positive for both RBC and BMO, with the companies' shares jumping after the deal's announcement. RBC expects to gain about $475 million from the sale after tax, and BMO $600 million.


However, some industry analysts are raising concerns the move could have negative repercussions for Canada's digital sovereignty amid the country's ongoing trade war with the U.S.
Wide-reaching implications

Broadly speaking, digital sovereignty refers to the power of a country, or individual, to retain control over its own digital assets.

In September, AI Minister Evan Solomon said that Canada needs to create a sovereign digital economy that is "free from coercion."
WATCH | Solomon spoke in September about why the government is acting:




'The moment is here' to modernize Canada's privacy laws: AI minister
June 15|
Duration1:29Artificial Intelligence Minister Evan Solomon says the Protecting Privacy and Consumer Data Act will help keep Canadians safe in the digital age. 'I think it's a common-sense approach to what Canadians and what average folks want from their digital media,' Solomon said on Monday about the new bill. 'They want some control over their personal information.'

That same month, dozens of experts and academics penned an open letter urging Prime Minister Mark Carney to swiftly "defend Canada's digital sovereignty" and protect the country from the whims of the Trump administration.

It's a sentiment Sharon Polsky, president of the Privacy and Access Council of Canada, shares.

"Canadians should be concerned because this is their information," she said.

"We're now going to have what has been a Canadian data-processing company that has access to everybody's purchasing habits ... available to other governments."

Thousands of businesses in Canada use Moneris. The processor services more than 325,000 points of commerce and processes more than five billion transactions every year, according to a press release from the company.

Polsky said the deal means Canadians' data could be available, not only to foreign governments, but foreign law enforcement agencies.

She gave an example of U.S. border agents potentially checking an individual's transaction history before allowing them to enter the country.


"Will you be stopped at the border because your purchase records indicate that you bought something with THC?" she said, referring to a chemical compound found in cannabis, which remains illegal in the U.S. under federal jurisdiction.

"It's possible these days, much more possible than ever before."
Transaction data as leverage?

That the deal is happening as the countries grapple with a trade war only heightens Polsky's concerns.

"It's easy to foresee that the rich data that comes from millions of Canadians' purchases could easily be leveraged to inform trade negotiations," said Polsky, who has advised corporations and governments for more than 30 years on privacy and access matters.

Colin Deacon, an Independent Canadian senator, is also concerned about how the U.S. government could use Canadians' data.

"[Data] under force and under request from the U.S. government could be then shared on an individual basis," said Deacon during an interview with CBC's Power and Politics.
WATCH | Deacon raises concerns over Moneris deal:




Canadian senator concerned by U.S. purchase of BMO, RBC payment processor
August 13|
Duration8:39Nova Scotia Sen. Colin Deacon is warning against further financial integration with the United States after one of Canada's biggest payment processing platforms was bought by an American bidder. Deacon joins Power & Politics to discuss why he is so concerned about this acquisition.

"There's a number of ways in which [U.S. President Donald Trump] has absolute control over services that Canadians rely on every day."

Both BMO and RBC pointed to press releases announcing the deal, refusing to provide further comment when contacted by CBC News.

"As ownership transitions to FP, Moneris's commitment to serving Canadian businesses will remain unchanged," read a separate press release from Moneris.
Privacy legislation in the works

Canada's legislation meant to protect digital privacy isn't where it needs to be, according to Polsky, who said this comes at Canadians' "collective peril."


"Canada isn't ready," she said.

Polsky argues there's a clear outcome if a company based in Canada is compelled by the U.S. to supply data as a result of this deal.

"If the choice is either to comply with the American law or go to jail, or comply with Canadian legislation that really is nowhere near as stringent … it's a no-brainer. They will comply with the American legislation."

The Canadian government has taken steps to address digital sovereignty. In June, the government introduced Bill C-36, the Protecting Privacy and Consumer Data Act.

New privacy bill would give Canadians right to request companies delete AI deepfakes

It proposes a major overhaul of Canada's private sector privacy framework, updating language to establish privacy as a "fundamental right" and replace parts of the Personal Information Protection and Electronic Documents Act (PIPEDA), which sets out rules for how businesses, including tech, airlines and banks, collect and use customers' data.

The bill also requires companies to conduct a "privacy impact assessment" before transferring personal data outside of Canada.

At the time, government officials said the legislation would cover any company that has a "substantial connection" to Canada.

But Polsky said C-36, and the Liberals' other attempts to update Canada's digital privacy laws, only "dance around the issues."

"They don't really speak to data being retained in Canada as a matter of national security or as a matter of data sovereignty," she said.

Bill C-36 represents the Liberal government's third attempt to update privacy regulations, following failed attempts in 2020 and 2023. It underwent its first reading in June, but has a number of steps to go before becoming law.


Meanwhile, the sale of Moneris still requires regulatory approvals, including clearance under the Competition Act, and is expected to close by the end of the banks' fiscal first quarter in 2027.

Where does that leave Canada?

"Behind the eight ball," said Polsky.

"[Canada is] trying to look tough, but we have a ways to go."

ClarificationsAn earlier version of this story said Moneris processes more than $5 billion in transactions every year. In fact, Moneris processes more than 5 billion transactions every year.
Aug 14, 2026 7:41 AM PDT

ABOUT THE AUTHOR


Brock Wilson

Journalist

India's automobile brands are outranking Tesla and BYD in EV efficiency

  In an unexpected twist, India's automobile brands are outranking Tesla and BYD in EV efficiency

Indian automakers Tata Motors and Mahindra have claimed the top two spots in a global electric vehicle efficiency ranking, beating out household names like Tesla and BYD. According to the International Council on Clean Transportation's 2025 report, which evaluated 22 of the world's largest EV manufacturers, Tata Motors led the field with an average energy consumption of just 106 watt-hours per kilometer, followed by Mahindra at 113 Wh/km. Tesla and BYD came in third and fourth respectively, while the overall industry average held nearly flat year on year at 131 Wh/km.

The results arrive at a significant moment for India's auto industry. EVs currently account for less than 5% of new passenger vehicle sales in the country, well below the global average of 25%, but the Indian government has set an ambitious target of electrifying 30% by 2030, a goal framed around cutting emissions, reducing oil imports, and expanding domestic manufacturing. India is the world's fourth-largest petroleum consumer, and officials are already shaping the third phase of tightened standards for automakers covering 2027 to 2032.

Efficiency gains alone, however, will not be enough to drive mass adoption. The ICCT ranking also revealed that Tata Motors ranked last among the evaluated manufacturers for charging speed and sixth from the bottom for driving range, two factors that weigh heavily on consumer decisions and will need to improve significantly if India is to meet its electrification ambitions.

Claude chats were showing up in Google searches

 

Claude chats were showing up in Google searches

Niv Bavarsky

Conversations with Claude that were intended to be private—some involving medical data and children’s phone numbers—were discovered in search engine results by Reddit users over the weekend.

How it happened: Users can create a shareable public link for Claude chats (much like you can with Google Docs), so friends or coworkers can view them. While a public Google Doc with your 80-page screenplay no one wants to read isn’t indexed for search, the Claude creations were.

Claude creator Anthropic told TechCrunch that those links only become public if they are posted where web crawlers can find them, like on a social media platform. Search engine companies had other explanations:

• A Google spokesperson told Wired that it gives Anthropic and any website owner “clear controls” over whether web pages are crawled or indexed.
• Microsoft’s Bing (it still exists) says that developers can use tags to block indexing. But when Wired reviewed the exposed Claude chats, it did not find the recommended “noindex” tag on those pages.

Chat, are we cooked? The Claude links were still publicly available as of Monday, but appear to have been erased from search results as of yesterday. However, it’s possible that third parties have already gotten their hands on the data.

Amazon Is Gutting Its AI Division

Amazon Is Gutting Its AI Division After Sustained Failure
Wait, Amazon has AI models?

By Joe Wilkins

Published Jul 29, 2026 1:57 PM EDT
Add Futurism(opens in a new tab)More information


Getty Images / Futurism




Sign up to see the future, today


Can’t-miss innovations from the bleeding edge of science and tech

Email addressSign Up

By now, you’d be hard pressed to find anyone who isn’t broadly familiar with the AI heavyweights and their flagship models. OpenAI has ChatGPT, Anthropic has Claude, Google has Gemini. But what does Amazon have?

Not a whole lot, it turns out. Unlike its competitors in the tech megacorp scene, Amazon’s standout AI model — organized under the name “Nova” — is about the farthest thing from a household name. And that’s not likely to change anytime soon, because the ecommerce giant is now gutting its in-house AI labs, Bloomberg reports.

Per the outlet, Amazon is scaling back the ambitions of most of its Nova AI models and reorganizing its AI development teams as part of an effort to narrow its focus on “highest priority” goals, after its power-hungry large language models turned out to be a bust.

Accordingly, the company has put most of its AI models on ice, Amazon insiders told Bloomberg. This includes its text-based Nova models, as well as video- and image-generating models, all of which have been placed in a state of “keep the lights on,” meaning they’re technically still supported, but only receive the bare minimum resources needed to sustain them.

Instead, labor and computing power are being diverted toward a singular “frontier-model effort” led by Pieter Abbeel, director of the Berkeley Robot Learning Lab whose robotics company Covariant was hoovered up by Amazon in 2024. At the time Abbeel joined the tech giant, Covariant’s founders were working on AI models for robots, which lines up pretty well with Amazon’s push to automate everything it possibly can.

News of the shift in Amazon’s strategic focus comes a week after it closed one of its key AI offices in San Francisco, an 80-person site specializing in research on artificial general intelligence, the supposed next-level of AI development where the tech obtains human-level intelligence and reasoning skills.

With urgent questions swirling over whether LLMs can ever usher in a return on investment — let alone AGI — it may be that Amazon is jumping out of the AI hype cycle while it still can.

More on Amazon: Amazon’s AI-Generated Animated Series Canceled After Relentless Derision



Joe Wilkins
Correspondent


I’m a tech and labor correspondent for Futurism, where my beat includes the role of emerging technologies in governance, surveillance, and labor.

LinkedIn Adds Button on Every Post to Report

LinkedIn Adds Button on Every Post to Report “Seems Like AI Slop,” Which Is Definitely Going to Get a Real Workout
A button that marks posts as not being AI slop would probably be far more useful.


By Victor Tangermann


Published Jul 30, 2026 12:54 PM EDT
Add Futurism(opens in a new tab)More information


Shutterstock / Futurism

Sign up to see the future, today

Perhaps more so than any other social media platform, LinkedIn has been inundated with lazy AI-generated brainrot content. It has become the AI slop “punching bag,” a repository of painful engagement bait being spewed out by an army of bots who seem at times to mainly be lobbing business platitudes at one another.

Just last week, Substack CEO Chris Best took aim at the Microsoft-owned career site when he announced the launch of an AI detector feature “because we’re sick of slop and we don’t want Substack to turn into LinkedIn.”

According to AI detection company Pangram, a shocking two thirds of LinkedIn posts are showing up as being AI generated by users of its browser extension. More than 40 percent of longform posts were flagged as “fully AI-generated,” according to the company.

Credit where credit’s due: LinkedIn has seemingly caught on to the abysmal state of its service, quietly introducing a new button that allows users to flag a post that “seems like AI slop.”

“Genuinely thought it was a joke until I clicked on a few posts,” one user wrote in a Bluesky post after discovering the feature.

It’s a baffling move, but probably well intentioned. Whether LinkedIn will act on what will likely be a massive influx of flagged posts remains to be seen, however. Once flagged, the post becomes hidden to the user, but it’s unclear whether somebody at LinkedIn will review it.

As 404 Media points out, the inclusion of the new button is particularly surprising, because LinkedIn is simultaneously encouraging users to generate AI slop through an AI-powered writing assistant.

“LinkedIn’s AI-powered writing assistant will share personalized suggestions for your profile, to help you stand out and get noticed,” the company’s website reads. “The AI-powered suggestions are personalized based on your profile information and incorporate insights from analyzing millions of profiles.”

The company also started scraping user data without permission to train so-called AI “coworkers” last year, a dubious experiment that was shut down almost immediately.

It’s clear that the subject of AI is proving divisive in the LinkedIn c-suite. In a May 20 update, LinkedIn promised that it was looking to keep “conversations real” by keeping “AI slop” at bay.

“When AI is overused, especially at scale and in an automated way, it dilutes the valuable insights that real human conversations can spark,” LinkedIn executive editor Laura Lorenzetti wrote in the update. “It’s ok to use AI to help you write, but your posts and comments need to represent your voice and your perspectives. The ultimate value comes from the human behind the tool.”

To allegedly “strengthen authenticity,” LinkedIn is working on “technology systems” that “have been trained to recognize signals of AI slop,” according to Lorenzetti.

Judging by the latest news, it’s looking to leverage its own frustrated user base to stop LinkedIn from drowning in a bot-infested sea of slop. Only time will tell if it’s too late.

Ksi Lisims LNG, German utility Uniper Sign 20-Year Contract

Ksi Lisims LNG, German utility Uniper Sign 20-Year Contract
July 30, 2026
Reading time: 3 minutes

Full Story: The Canadian Press
Author: Lauren Krugel


Σ64/Wikimedia Commons

German utility Uniper has locked down a formal long-term agreement to purchase liquefied natural gas from the $30-billion Ksi Lisims project planned for the northern British Columbia coast.

The Düsseldorf-based company will purchase two million tonnes per year of LNG for up to 20 years, with first deliveries expected in 2032, Uniper and Ksi Lisims said in a joint news release. They signed a letter of interest last month outlining key commercial terms for the binding supply and purchase agreement announced Wednesday.

The deal marks the first major long-term LNG supply agreement between Canada and Germany, they said.

“As a trusted partner, Canada helps diversify Europe’s energy supply and strengthens resilience against future disruptions,” said Uniper chief executive Michael Lewis.

Uniper has 18.5 gigawatts of power generating capacity and is a major LNG importer in northwestern Europe. The German government took it over amid the 2022 energy crisis following Russia’s invasion of Ukraine, but is now in the process of privatizing it.

Houston-based Western LNG is the lead developer and future operator of Ksi Lisims alongside Rockies LNG, a consortium of Canadian natural gas producers, and the Nisga’a Nation, on whose lands the project would be located.

“Canada is naturally endowed with a world-class natural gas resource, which, through responsible development, is one of the lowest emissions production complexes in the world,” said David Thames, head of Western LNG.

The floating plant is meant to export up to 12 million tonnes of LNG per year from the site on Pearse Island, by the Alaska border, but has been having trouble lining up contracts for that much output.

B.C. Premier David Eby was asked about balancing environmental concerns with growth in the LNG industry during an unrelated news conference on Wednesday.

He said the Nisga’a Nation and its partners had “work to do with other nations and with local communities just like any other project proponent to bring people along.”

Eby touted B.C.’s approach to major resource projects, saying the province was “willing to put in the time and the energy to work with communities, to work with First Nations, to deliver these projects in partnership.”

A federal government news release Wednesday said the $30-billion Ksi Lisims project would be Canada’s second-largest LNG facility with the potential to represent 13% of the country’s total natural gas exports by the early-to-mid 2030s.

A report by the B.C. Environmental Assessment Office in August 2025 had pegged initial capital cost for Ksi Lisims at $10 billion to $12 billion.

The $30-billion figure is an updated capital cost estimate and does not include the Prince Rupert Gas Transmission pipeline that would feed the plant, the office of Natural Resources Minister Tim Hodgson said Wednesday.

Richard Brooks, climate finance director with Stand.earth called the deal “a major step backwards for climate action” for both Germany and Canada.

“Moving forward with a massive new methane gas project as devastating wildfires rage across Canada and Europe is irresponsible and an abdication of leadership,” he said.

“Morally, it’s a slap in the face to all of us who are impacted by deadly heat waves and smoke.”

The main body of this report was first published by The Canadian Press on July 29, 2026.
This story is part of The Energy Mix’s partnership with Small Change Fund.

City Travel Books – RG Richardson

City Travel Books – RG Richardson

Interactive City and Travel Books

Author: R.G.Richardson
This is a live interactive search guidebook with 12,300 presets that searches for everything about your city. Pick and click on the icon, never goes out of date!
You can search for events, restaurants, banks, hotels, shopping, apartments and sports. Find everything that is happening in the city!
In the guidebook, you look in the index for what you want to search, and then you click on the button next to it, and you instantly have your search items displayed.
All guides search in 10 languages and feature the Qwant, DuckDuckgo, Startpage Private Browsers.

Since 2003
eComTechnology/RGRichardson©2026
Assign Centre, ISBN Division
Library and Archives Canada
Author R.G. Richardson
Victoria, BC. Canada V8R 5G9

Updated 7/2026.

France tightens checks on foreign purchases of 'sensitive' firms

 France tightens checks on foreign purchases of 'sensitive' firms

Copyright Ian Langsdon/Pool
By Nathan Joubioux
Published on 03/08/2026 - 13:06 GMT+2•Updated 13:09
ShareCommentsAdd Euronews on Google
Until now, the Economy Ministry had a say whenever a foreign company bought 25% or more of a French business; that threshold has now been cut to 10%.


The prime minister's office at Matignon has decided to tighten the rules. In a decree published this Sunday 2 August (source in French), Prime Minister Sébastien Lecornu has strengthened state control over the purchase, by foreign investors, of companies in sensitive sectors.


Until now, the Ministry of the Economy had a right of scrutiny whenever a non-European investor acquired 25% or more of a French company listed on a regulated market outside the European Union. "From now on, any stake of at least 10% taken by a non-European investor in a French company listed on a regulated market outside the European Union will have to be authorised by the state", the prime minister wrote on X.

"Protecting our strategic companies means protecting our sovereignty. Our responsibility is twofold: to support the development of our companies while safeguarding our strategic interests", Sébastien Lecornu said.



The aim of this decision is to "guard against opportunistic non-European shareholdings in French companies listed outside the EU that could pose threats to national security", the prime minister's office said in a statement.

"To avoid unduly undermining companies' ability to raise funds on the markets, this strengthened oversight will be carried out under a fast-track procedure", the statement also said.

Investors will now have to notify their transaction to the Treasury Directorate-General. The economy minister will then have ten days to decide whether it requires an in-depth review.
RelatedPresident Emmanuel Macron announces €93 billion in 'Choose France' investments
"Protecting our companies is not enough"

This decree is in line with several recommendations set out in a parliamentary report calling for a "radical shift in posture" in order to strengthen the country's economic security. Drafted by Christophe Plassard (Horizons MP for Charente-Maritime), Jean-Louis Thiériot (Right-wing Republican MP for Seine-et-Marne) and Charles Rodwell (Ensemble pour la République MP for Yvelines), the report stressed that the French framework needed to be reinforced at a time when foreign investment "is taking on greater significance because of the geopolitical context".

"France has one of the most robust foreign investment screening mechanisms in Europe. But our sovereignty also depends on our ability to finance our strategic companies", Jean-Louis Thiériot said on X (source in French). "Protecting our companies is not enough. We must also give them the means to grow."

In the balance of power between France and investors, Christophe Plassard is calling for us to "shed our naivety". "We need to be defensive if our capital and our companies come under attack. But we must also be on the offensive and clear-eyed about our ability to retain our know-how", the MP said on France 24 (source in French).

Google has promised to pay $44bn of rent on buildings it will never own

Google has promised to pay $44bn of rent on buildings it will never own

Google has promised to cover up to $44bn of rent on data centres it does not own, if the tenants stop paying. Nine months ago that promise was worth $6.5bn. The guarantee is how Google sells chips against Nvidia, and it is the clearest measure yet of an AI build-out being financed by moving risk off the books rather than putting money down. Three more giants report earnings this week.

July 27, 2026 - 11:14 am


Image by: The Pancake of Heaven!


Start with the number, because the number moved fast.

Google has agreed to cover as much as $44bn of lease payments on data centres owned by other people, should the tenant default. It disclosed the figure last week. At the end of September the same commitment stood at $6.5bn, The Information reported.

That is close to seven times larger in nine months.

The promise is simple to state and hard to price. Google does not own the building. It does not occupy the building. It has told the lenders that if the company renting the building stops paying, Google will.
Why a search company is underwriting other people’s rent


TNW City Coworking space - Where your best work happens

A workspace designed for growth, collaboration, and endless networking opportunities in the heart of tech.


The answer is chips. Google wants customers for its tensor processing units, the in-house silicon it is pitching as an alternative to Nvidia.

A guarantee from a company with Google’s balance sheet lets a data-centre developer borrow more cheaply. Cheaper debt makes the project work. The project buys TPUs. Anthropic rents them.

Readers here have seen the mechanism. In June we reported how Google borrowed Nvidia’s own playbook, including a $3.2bn guarantee at Lake Mariner in western New York, $7bn at River Bend in Louisiana and $1.4bn in Texas.

What is new is the total, and the speed. Those individual guarantees have now consolidated into a single disclosed figure that has grown almost sevenfold since September.
A different kind of balance sheet

Until a year or two ago the giants had simple accounts. Cash covered debt several times over. That is no longer the shape of the thing.

Now they use techniques long familiar on Wall Street, expanding the business without carrying all of the risk themselves. The guarantee is the purest example. It commits nothing today and everything in the wrong scenario.

We have covered the aggregate before. Five US giants carry $1.65tn of off-balance-sheet AI debt, more than they report outright. This is the same story told through one company and one line item.
The machinery, and why it is legal

Moody’s set out the mechanics earlier this year. The five biggest US hyperscalers had amassed $969bn of future lease commitments by the end of 2025. Some $662bn of that had not yet commenced, so none of it sat on a balance sheet, Fortune reported. That hidden portion equals 113% of the five firms’ adjusted debt.

The cause is the kit. Data-centre leases used to run 10 to 15 years. AI hardware is useful for four to six. So tenants now demand short leases with options to renew, and landlords demand security before they will build.

The security is usually a residual value guarantee. If the tenant walks away and the building is worth less than an agreed threshold, the tenant pays the difference.

Accounting rules let that sit off the books. A renewal counts as a liability only if it is “reasonably certain”, a test above 70%. Since nobody can say what AI hardware will need in 2031, firms can argue the renewal is likely without being certain, and the guarantee stays in a footnote.

Meta shows the scale. It disclosed data-centre leases starting in 2029 worth about $12.3bn, alongside a residual value guarantee with a $28bn threshold. It judged a payout not probable, so no liability was recorded. Its $50bn Hyperion campus in Louisiana runs on a version of the same structure.
‘Every nook and cranny’

The reason for all this creativity is that the money required has outgrown the ordinary ways of raising it.

John Greenwood, Goldman Sachs’s global head of infrastructure and real asset finance, said he is “looking for capital in every nook and cranny” to support an expected $7.5tn of spending on chips, data centres and power over the next five years, The Information reported.

The hunt does not stop there. Much of that spending goes on chips that need replacing every few years, so the bill arrives again.
The thing money cannot buy quickly

Capital is not the only bottleneck. Nvidia says the harder constraint is physical.

“What’s limiting the access to compute capacity is the fact that it’s very hard to find powered data centers, powered land,” said Raj Mirpuri, Nvidia’s vice president of global AI clouds and infrastructure.

That is why the guarantees exist. A promise from a trillion-dollar balance sheet is the fastest way to turn a plot of land with a grid connection into a building full of chips. Nvidia is now doing the same thing, using its own balance sheet to help customers afford chips and help partners finance the halls to put them in.
Why it matters on Wednesday

Meta and Microsoft report on Wednesday. Amazon and Apple follow on Thursday. Alphabet went first last week and lifted its capital-expenditure guidance to $205bn.

The reported debt will look manageable. The guarantees will not be in the headline numbers.

Investors are already uneasy. Meta is down about 10% this year and Microsoft about 21%, while Apple, which barely spends on AI, is up 23%. Microsoft is short of compute and rationing it. Amazon has guided to roughly $200bn of capital spending against about $185bn of expected operating cash flow, according to S&P Global Market Intelligence, which means dipping into reserves it has topped up with bond sales.

It is in talks to backstop about $250bn of OpenAI’s financing, and it just put $1bn into Korea’s Naver alongside $9bn from Brookfield. The guarantee has become standard equipment.

None of it is improper. The disclosures exist. Moody’s analysts David Gonzales and Alastair Drake made the narrower point that these obligations are not missing, only early. They have not yet been triggered, but they will be.

Their warning was about what the accounts cannot show. “The accounting liability is unlikely to reflect certain plausible future scenarios,” they wrote.

A guarantee costs nothing at all, right up until the moment it costs everything. Google has written $44bn of them in nine months, on buildings it will never own.

OpenAI to triple workforce at Dublin European headquarters to 350

OpenAI to triple workforce at Dublin European headquarters to 350


Published by Global Banking & Finance Review


Posted on July 27, 20262 min read

· Last updated: July 27, 2026Add as preferred source on Google
FinancetechnologyAIIrelandWorkforce
Quick Summary

OpenAI will expand its Dublin European headquarters from just over 100 employees to 350 by leasing 8,000 m² in Silicon Docks and adding 250 roles over two years, strengthening its footprint in a key European tech hub.

Table of ContentsOpenAI's Growth Strategy and Impact on Dublin's Tech Sector
OpenAI's Expansion Plans
Current Workforce and Future Hiring
Dublin as a Hub for Tech Giants
Factors Attracting Multinationals to Ireland
Recent Trends in Tech Employment
Article Credits

OpenAI Plans Major Expansion at Dublin European Headquarters, Hiring 250 New Staff
OpenAI's Growth Strategy and Impact on Dublin's Tech Sector
OpenAI's Expansion Plans

DUBLIN, July 27 (Reuters) - OpenAI is to more than triple its headcount at its European headquarters in Dublin to 350, the ChatGPT maker said on Monday as it announced it was leasing 8,000 square meters (88,000 square feet) of office space in the city's Silicon Docks area.
Current Workforce and Future Hiring

The AI pioneer said it currently employs over 100 people at its Dublin office, which opened in 2023, and that it plans to hire 250 more over the next two years in both engineering and support operations.
Dublin as a Hub for Tech Giants

Dublin is home to the European headquarters of a number of U.S. tech giants including Google owner Alphabet, Facebook owner Meta and Microsoft.
Factors Attracting Multinationals to Ireland

NewsletterStay ahead of the marketsThe day's essential banking & finance news, free to your inbox.Subscribe
Advertisement

Foreign multinationals, mainly in the technology and pharmaceutical sectors, employ about 11% of Irish workers and often cite Ireland's highly educated workforce as a factor in their decision to locate there. Many are also attracted by the country's tax regime.
Recent Trends in Tech Employment

Some large technology companies such as Meta and TikTok have announced job cuts in their Irish operations during the past year and the government has pointed to AI as a priority for attracting new investment.
Article Credits

(Writing by Conor Humphries; editing by Sarah Young)
Key TakeawaysOpenAI currently employs just over 100 staff in Dublin and will add 250 more jobs over the next two years across engineering and support functions, bringing the total to 350 employees (rte.ie)
The company has leased 8,000 m² (88,000 sq ft) of office space in Dublin’s Silicon Docks—specifically the Tropical Fruit Warehouse—to serve as its new EU headquarters (rte.ie)
The expansion reflects Ireland’s continued appeal for tech and pharmaceutical multinationals, buoyed by its skilled workforce, supportive AI strategy, and strong foreign direct investment environment (gov.ie)