FT : Banijay picks London as new HQ for TV maker after $8bn merger UK base for p

Banijay picks London as new HQ for TV maker after $8bn merger
UK base for production group behind ‘The Traitors’ is another boost to country’s creative sector

The world’s largest independent TV maker has picked the UK for its new headquarters, following the $8bn merger of Paris-based Banijay Entertainment and All3Media, which is owned by RedBird IMI.

A London base for the production group behind shows such as Peaky Blinders, MasterChef and The Traitors will be another boost to the UK’s creative sector — after this week’s deal by Sky to buy the broadcasting unit of ITV included plans to spin off the studios division into its own London-listed business.

“London is the capital of the [entertainment] business,” said Marco Bassetti, chief executive of Banijay Entertainment who will remain the boss of the combined group, referencing the future ITV Studios spin-off. “The big streamers have a big office here for running Europe, and the UK for us is part of Europe.”

New York-based RedBird IMI — which last year attempted to buy the Telegraph newspaper — and Banijay are partners in the new business, which spans 25 countries, with the board equally split between the two. 

The two owners on Thursday confirmed the completion of the deal and the full details of the new company, which will be the world’s largest independent TV maker, encompassing more than 170 production and live-events companies.

On a combined basis, Banijay Entertainment would have generated more than €4.3bn in revenues and more than €700mn in adjusted earnings in 2025, with cost savings of €50mn expected to be achieved within a year of the deal.

Former CNN boss Jeff Zucker, who runs Abu Dhabi-backed RedBird IMI, the investment partnership that bought All3Media in 2024, said that the merger would create “a truly global media and entertainment powerhouse” that will “produce premium content, live events, digital content unlike anyone else”.

Zucker will serve as chair and All3Media boss Jane Turton will become deputy CEO.

The combination of the two groups reflects the wider consolidation across the entertainment world, where deals such as Paramount’s acquisition of Warner Bros Discovery in the US and the merger of Sky and ITV’s television business in the UK are set to create new industry giants.

“We are at a time of tremendous change and consolidation and opportunity in the media space, and that’s why we felt that this was both necessary and opportunistic,” said Zucker. “You have to have more scale in the new world, and you can’t just get there with three or four smaller acquisitions.”

He added that “there’s going to be further opportunities to come, and the ability to grow across content production, both scripted and unscripted, live events and experiences in digital and in sports”.

Zucker did not rule out a potential listing of the entertainment company in future, saying that the company would go through the integrated process before deciding “whether we think that the public markets may be an opportunity as well . . . let’s see what happens”.

One option would be to list the business in New York once the integration is complete, according to a person familiar with the matter. A Banijay spokesperson said that “we just closed the merger today so are yet to explore anything, but we will not exclude any option to give the company the best tools and positioning to win”.

Bassetti said that the merger would bring new opportunities to exploit and build out its existing content portfolio, taking older franchises into new territories and developing live events around popular shows, as well as developing new programmes.

He said that the group had more than 260,000 hours of content in its catalogue, a “significant direct-to-consumer business” and a roster of live shows. The company also produces live events for others, including the opening ceremony of the Winter Olympics in Milan and the World Cup.

FT : Suspicious trading before 41% of takeovers sets new UK record Annual FCA da

Suspicious trading before 41% of takeovers sets new UK record
Annual FCA data shows abnormal trades hitting new highs amid deal frenzy for UK companies

Suspicious share trading activity occurred before more than four out of 10 UK takeover announcements last year, a record high that adds to concerns at the financial watchdog about leaks and insider trading.

The Financial Conduct Authority said that abnormal trading activity it detected in share prices in the two days before takeover announcements rose from almost 38 per cent of deals in 2024 to 41 per cent in 2025.

The deterioration in the regulator’s “market cleanliness” data, along with sharp increases in abnormal trading ahead of other sensitive market announcements, is a setback for the FCA, which aimed to reduce financial crime as one of the four key objectives in its new strategy last year.

The FCA has been increasingly worried about deals and other sensitive information being leaked or used for insider trading and last year its co-head of enforcement and market oversight, Therese Chambers, issued a public warning to market participants.

The watchdog said the increase in abnormal trading ahead of UK takeover announcements pushed it well above the five-year average of 33.7 per cent. It said this was “just one indicator of possible insider trading” and there could be other factors behind it.


“We remained in a period of elevated market volatility, driven by global economic and geopolitical conditions,” the FCA said. “This is likely to have influenced price movements and trading volumes, increasing the number of signals detected.”

Suspicious trading is at record levels as UK takeover activity is also reaching new heights as foreign buyers and private equity snap up London-listed companies.

The watchdog said that despite recent adjustments to the metric’s methodology it “continues to have limitations as a broader measure of market cleanliness”.

Some insider trading could happen before the two-day window the FCA monitors for abnormal share price movements, it said, adding: “Price moves could have been caused by financial analysts or the media correctly predicting likely takeover targets.”

Last year, the FCA called in the heads of M&A from big investment banks to discuss how to tackle the issue of deal leaks alongside the Takeover Panel.

In separate data published alongside the FCA’s annual report on Thursday, it said there were abnormal increases in trading volumes ahead of 8.1 per cent of price-sensitive announcements by UK-listed companies last year, up from 5.6 per cent the previous year.

The FCA said it had secured four convictions for insider trading and fined 12 people a total of £1.77mn for insider trading, market manipulation and making misleading statements in the year to April.

Earlier this week, it charged a former deals lawyer at Goodwin Procter, the US law firm, with insider trading using confidential information gained while working on the sale of maternity brand Seraphine Group.

The watchdog also reported a big increase in UK financial fraud, dealing a further blow to its strategic goal of “fighting financial crime”. Investment fraud more than doubled to a record £1.2bn in the year to April, it said.

Losses from authorised push payment fraud, in which people are tricked into sending money to criminals, rose 19 per cent to £576.4mn in the year to April.

It blamed the rise in fraud on the use of technologies such as AI and cryptocurrencies, which it said “allow fraudsters to operate at greater scale and speed”, adding that market volatility also makes consumers more susceptible to fraud.

The watchdog said it was working with “domestic and international partners, for example, through better data sharing, to stay ahead of criminals” and “supporting wider system-priority work with industry and partners, such as combating money mules”. Last year, it launched a “firm checker” to help consumers spot scams by fake companies.

WWD : LVMH’s Bernard Arnault Is Big on Family, Not Fond of ‘Luxury’ and Ready to

LVMH’s Bernard Arnault Is Big on Family, Not Fond of ‘Luxury’ and Ready to Go 250km an Hour
The often quiet CEO opened up in a rare extended interview on the Legend podcast with Guillaume Pley.

FAMILY AFFAIR: Bernard Arnault only rarely speaks at length publicly, preferring instead to stay heads down running LVMH Moët Hennessy Louis Vuitton.

But when Arnault sat down for a rare extended interview with Guillaume Pley on his popular YouTube and podcast channel Legend, the luxury titan was in a mood to open up.

Over an hour and 40 minutes, he dug into his childhood, his parenting philosophy, his long friendship with U.S. President Donald Trump and, of course, luxury, although he doesn’t like the term even as it has defined the billionaire’s work life.

“Luxury is not a word I am fond of particularly,” Arnault said, according to a translation of the interview, which was conducted in French. “It’s often associated somewhat pejoratively with aspects that have nothing to do with our business. Frivolity, superficiality, uselessness, all that flashy bling.”

Instead, he said LVMH leads in the sector of “high-quality products” that come together through a series of “highly specialized artisanal crafts.”

Arnault, who built LVMH into a luxury powerhouse whether he likes the word or not, stressed the importance of family, remembering what he learned about businesses from his grandfather and father and how he passed those lessons on to his own kids and now has grandkids of his own.

He spoke with particular fondness of his relationship with his grandfather.

“He died on June 26, 1959, and he had always paid close attention to my schooling,” Arnault recalled. “He was very happy to see that I was doing well in school. And so I had received my grades, which were good, and I put them in his coffin.

“My grandfather, he taught me a great deal whenever I went along with him to the various construction sites to see the site managers and to talk with the workers,” he said.

In turn, Arnault showed his own children — now executives in the business — “exactly what a company was” at a very young age.

“I would talk to them about how I ran the business and they followed along with all the different paths we took,” he said.

Learning about business from dad has taken the Arnault brood around the world and into some singular situations.

For instance, Arnault and his son Alexandre were due to have dinner with Trump in September 2024, the day the-then presidential candidate’s security thwarted an assassination attempt, the second in just nine weeks.

Arnault was sure the dinner would be canceled, but it wasn’t.

“We sit down with him and say, ‘It’s terrible what happened this afternoon.’ And he replies, ‘Two to zero,’” he recalled, adding that Trump was “totally relaxed” and that they stayed until 11 p.m.

Arnault also opened a window into his many adventures as chief executive officer, from going over past takeover battles to getting up close with Formula 1 racing.

Before a grand prix in Monaco, former world champion Max Verstappen took the 77-year-old for a spin around the track — “the day before the race at 250 kilometers per hour, through the streets of Monaco.”

“I can tell you, it’s something you remember,” he said. “Well, Verstappen was driving, so I told myself the risk is limited.”

WWD : Hermès Sets Date for First Haute Couture Runway Show Nadège Vanhée, artist

Hermès Sets Date for First Haute Couture Runway Show
Nadège Vanhée, artistic director of women’s ready-to-wear, will present her first made-to-measure collection during Paris Couture Week in January.

PARIS — Hermès’ eagerly awaited haute couture line finally has a launch date.

The French luxury house revealed Thursday that Nadège Vanhée, artistic director of women’s ready-to-wear, will present her first made-to-measure collection during Paris Couture Week in January.

The brief statement, shared exclusively with WWD, provided no additional details about what is expected to be a highlight of the spring 2027 haute couture season, given Vanhée’s technical expertise with exclusive materials like leather, silk and cashmere. It will mark her first foray into couture.

Hermès unveiled plans to enter the category last year and has since beefed up its studios with a dedicated atelier. According to sources, recent hires include French designer Léa Peckre, who was previously design director of women’s ready-to-wear at Céline.

Speaking at the company’s annual results presentation in February, Hermès chief executive officer Axel Dumas said work on the collection was underway. “What I saw was superb. I’m really quite excited, and I’m very proud of what the teams have done,” he said.

The maker of Birkin bags and silk scarves has long relied on its leather goods business as its primary growth driver, with tight supply underpinning premium pricing power and exclusivity. But that model may be showing its limits in a slower-growth environment.

The move to upgrade its fashion offering with a made-to-order component comes as Hermès, previously immune to market turbulence, is starting to feel the impact of a prolonged slump in luxury spending. The company’s share price has fallen by more than 25 percent since the start of the year.

Sales rose 6 percent at constant currency in the first quarter, making Hermès the strongest performer among its luxury peers, but missing analysts’ forecasts. Nonetheless, Dumas told the company’s annual shareholders’ meeting in April he was “very happy” with the result.

“The reality is that luxury demand is not particularly strong these days,” said Bernstein analyst Luca Solca, pointing to a combination of weakening Chinese consumption and geopolitical disruption that is testing the sector’s long-held assumptions.

For Hermès, he added, the challenge could soon be more structural. “We may be getting closer to a time when they need to pull a rabbit out of the hat, when they need to launch something new, when they need to excite consumers again,” Solca said.

In tandem with Vanhée’s couture debut, Grace Wales Bonner will be making her runway debut as creative director of menswear at Hermès next January, ensuring the house kicks off 2027 with a bang.

WWD : Luxury Retail Gets Swept Up in Stock Market Uncertainty Shares of high-end

Luxury Retail Gets Swept Up in Stock Market Uncertainty
Shares of high-end brands fell as new turmoil in the war with Iran clouded the outlook, even though CEOs, like Capri's John Idol, remain relatively bullish.

While consumers and retail companies have been holding up surprisingly well through a world of trouble — from war to inflation to political uncertainty — the fashion industry is still vulnerable to shocks on Wall Street.

Think of it as a reminder of how quickly trends can change.

On Wednesday, it was a sell off of luxury and fashion stocks as the U.S.-Iranian cease fire appeared to crumble as well as a warning from the International Monetary Fund that global economic growth would slow to 3 percent this year, down from an average of 3.5 percent over the prior two years.

The Dow Jones Industrial Average fell 1.1 percent , or 576.76 points, to 52,348.39 and fashion went along for the ride.

Among those falling the hardest in the global fashion industry were Capri Holdings, down 7.8 percent to $17.90; Bath & Body Works Inc., 6 percent to $19.39; Kohl’s Corp., 5.2 percent to $15.89; Kering, 5 percent to 239.65 euros; Hermès International, 4.2 percent to 1,570.50 euros; Tapestry Inc., 4.2 percent to $140.14; Inditex, 4 percent to 54.16 euros, and Brunello Cucinelli, 3.8 percent to 78.82 euros.

The war in Iran, what will happen if it really flares back up and how the oil shock that’s already reverberated out of the conflict will translate into the broader economies is almost the definition of an “uncontrollable” for a fashion chief executive officer.

The corporate bigwigs prefer to look past the big macro changes and focus instead on the “controllables” they can influence.

So it was with John Idol, chairman and CEO of Capri Holdings, who was at a Bernstein investor conference in New York talking about all the change at his company.

Capri, which was set to become part of Tapestry Inc. before the deal was hung up by an antitrust challenge, sold Versace last year, dramatically cleaning up its balance sheet and ending the year with just over $200 million in debt.

That has helped Capri zero in on its Michael Kors and Jimmy Choo businesses.

“Brands go through cycles,” Idol said. “They go up and down at various points in time. And when we really looked at our portfolio, we thought that we had an incredible opportunity to focus on Michael Kors. And that’s when we made the decision to sell Versace.”

Idol is looking to get Michael Kors’ sales from about $3 billion back to $4 billion, while Jimmy Choo builds from $600 million to $800 million.

It’s a multipart plan, with investments into stores, a marketing push, a reduction in stock keeping units, more emphasis on trend-focused styles and, crucially, a pricing reset.

“We looked at our pricing architecture and our sell-throughs had gone down, our markdowns had gone up, and we saw it was a result of our raising prices, coming out of COVID-19, some 20 to 25 percent,” Idol said. “So we went back and we got closer to our traditional historical pricing for the brand. And what’s happened subsequently from that — and this is in our full-price channel, both our own stores and wholesale — is that we’ve seen our full-price sell-throughs go up dramatically and our [average unit retail prices] are going up because we’re taking less markdowns.”

And that is all stuff that Idol can control.

The bigger things will just have to sort themselves out, but the CEO said consumers are still buying, even if they need a little something more to get them to spend.

“Whether they’re at the very top echelon of luxury or whether they’re at a different income level that’s lower than that, everyone’s just being a little bit more thoughtful in their purchase,” Idol said. “They’re still buying, but they’re just making sure that obviously something is exciting first from a design standpoint [and] that it has quality.”

He said the North American customer was still “relatively healthy” despite “concerns around rising costs on everything from fuel to groceries to rent.”

But Idol added that Capri has become “a little bit more cautious about where Europe is going” over the past three months, although it’s been “relatively steady.”

“There’s definitely an issue around tourism,” the CEO said. “In particular, the Middle East tourists are not traveling to Europe. We have a substantial business in both Jimmy Choo and at Michael Kors with Middle East tourists. So we’re feeling the impact of that in Europe and we’re clearly seeing people hesitating on shopping in the Middle East area.”

Idol said he was “optimistic about China. We definitely see the consumer starting to improve. And I don’t think that the economy’s going to be rip roaring, but I think that you see consumers who are coming out.”

TechCrunch : Instagram users: Here’s how to stop Meta’s AI from using your photo

Instagram users: Here’s how to stop Meta’s AI from using your photos

On Tuesday, Meta launched “Muse Image,” a new AI image-generation feature that allows users to create original images, edit existing photos, and even generate custom ads directly within its apps.
But one capability has quickly become the center of controversy.


Muse Image allows users to generate AI images using photos from public Instagram accounts. As long as a person’s profile is public, another user can tag that account and use their images as part of an AI-generated creation. (Only private accounts and accounts belonging to users under 18 are automatically excluded from the feature.)
One huge concern is consent. Users may have no idea that their public photos can be incorporated into AI-generated images by strangers, and they aren’t even notified when someone reuses their public content. Plus, making it easy to manipulate people’s images opens the door to misuse, harassment, impersonation, and nonconsensual image editing.
If you’re looking to opt out of this, here’s how you can do it.
How to opt out of Meta’s Muse Image generator
  • Head to your profile and click the three horizontal lines in the top-right corner.
  • Scroll down to “Sharing and reuse.”
  • Look for the option that says, “Allow people to use your content on Instagram with AI features on Meta”
  • Toggle the setting off for both posts and reels.
ScreenshotImage Credits:Instagram screenshot
Muse Image arrives at a time when AI tools are being increasingly integrated into social media platforms. As tech companies race to roll out new generative AI features, many experts argue that stronger privacy protections and greater transparency are needed, so users fully understand how their photos and personal data are being used.
Public skepticism around AI is already high. According to a Pew Research Center survey, 35% of respondents said they’re more concerned than excited about the growing use of artificial intelligence.

Additionally, Meta’s track record on user privacy has also fueled skepticism surrounding its latest AI feature.
In 2019, the U.S. Federal Trade Commission (FTC) imposed a $5 billion fine against Facebook, concluding that the platform had violated a 2012 consent order by misleading users about how much control they had over their personal information. This followed a high-profile scandal where political consulting firm Cambridge Analytica gained access to data from up to 87 million Facebook users through a personality quiz app. Facebook’s platform policies at the time allowed developers to collect information about those users’ friends without their knowledge or explicit consent.

The Information :Khosla-Backed Startup Claims Breakthrough With Largest-Ever AI

Khosla-Backed Startup Claims Breakthrough With Largest-Ever AI Model on an iPhone

The Takeaway
  • PrismML said it has shrunk a 27-billion-parameter AI model to run on an iPhone.
  • Startup says its technique enables complex AI tasks on-device without performance loss.
  • Apple has talked to PrismML about ways it could use its technology.

Apple is on a quest to shrink powerful AI models to run on iPhones, which could cut down on cloud computing costs and enhance user privacy. But a small startup that emerged from stealth mode earlier this year says it recently got an AI model running on an iPhone bigger than any previous mobile model.

The startup, PrismML, said it has shrunk down Qwen 3.6, an open-source large language model developed by Chinese internet giant Alibaba, to run on an iPhone 17 Pro. The model has 27 billion parameters, which are roughly similar to the synapses in a brain and can help determine the complexity of the data a model can process. In contrast, most models that run on mobile phones have only a few billion parameters active at a time.

The largest AI models, which can measure in the trillions of parameters, are still far too big to run on mobile devices. But the model PrismML has working on an iPhone is capable of tasks like complex chat, reasoning, fully autonomous agents and software coding, the startup said. The open-source model will be available for download next week on Tuesday.

The milestone, which hasn’t been previously reported, reflects a broader push to get AI running on devices instead of expensive high-powered servers in data centers. Microsoft, Amazon, Meta Platforms and others are spending hundreds of billions of dollars racing to build those data centers to keep up with the demand they’re anticipating for AI.

Apple, though, has largely stayed on the sidelines of the costly data center race while also being a vocal proponent of making sure as many of the iPhone’s AI functions as possible run on the devices rather than in the cloud. The company believes on-device AI will better allow it to deliver on its privacy and security promises to customers.

In an interview, Babak Hassibi, CEO of PrismML, predicted that the vast majority of AI will eventually be processed on devices.

“Imagine a world, maybe three years from now, where 95% of the intelligence that you need is available to you locally, on your phone, on your laptop, on your appliances, and it’s really on the last maybe 5% of high-end stuff that you’ll need to go to the cloud,” Hassibi said. “I think that’s how people are seeing the way forward.”

Shrinking models to run on devices, he added, “fundamentally changes the economics of AI.”

Some prominent AI investors have bought into the startup’s vision. Khosla Ventures, which was OpenAI’s first venture capital investor, also put money into PrismML’s $16.25 million seed round earlier this year. In an interview, Vinod Khosla, the firm’s founder, said he became interested in PrismML because the company offers a “fundamental breakthrough.”

“We bet heavily on transformer models in 2018 when we invested in OpenAI, but what are the new ways to build AI?” Khosla said. “Our team is always looking for new approaches.”

PrismML uses a mathematical trick to shrink the Qwen 3.6 model to a fraction of its original size. Shrinking models typically results in worse performance, but the company claims its technique for miniaturizing AI model sizes doesn’t hinder their performance. PrismML has compressed the size of Qwen 3.6 to less than 4 gigabytes, down from around 54.

PrismML is a spinoff of the California Institute of Technology, where Hassibi, a professor of electrical engineering at the school, and his co-founders conducted the mathematical research used in the startup’s technology. Caltech owns the patents behind the technology but licenses them exclusively to PrismML.

PrismML plans to continue shrinking larger AI models, even at the scale of a trillion parameters, which will bring it into the realm of cutting-edge models such as OpenAI’s GPT and Anthropic’s Claude, said Hassibi.

PrismML’s approach may particularly appeal to Apple. At the company’s Worldwide Developers Conference in June, it announced its long-delayed Siri overhaul based on Google’s Gemini models. The most advanced parts of Siri are still so big that they require Apple to tap into Nvidia chips running in Google Cloud.

As part of the new Siri announcement, Apple said some of the iPhone’s new AI capabilities would run on devices. One new on-device Apple model has 20 billion parameters but uses a so-called sparse architecture, in which only 1 billion to 4 billion parameters are active at a time. In the case of PrismML’s on-device model, all 27 billion parameters are active at the same time.

In the past, Apple has run into stumbling blocks with its on-device model efforts. Last year, as it was trying to shrink its internal AI models to fit on the iPhone, the company discovered that performance took a dramatic hit, The Information has reported.

Apple is currently on the hunt for acquisitions of companies that can help it run more AI on device, The Information previously reported. Apple has held meetings with PrismML about ways it could use its technology, people familiar with the talks said.

There are other approaches to on-device AI. For example, startup Argmax is focused on deploying AI on device for tasks like processing voice and images, and then sending that processed information up to the cloud for the more complex reasoning work.

One big reason for that hybrid approach is because the large models running in the cloud are still changing so fast, with updates arriving on a weekly basis. Proponents of that approach argue that AI running entirely on devices won’t get the advantage of the latest, most advanced models, which run in the cloud.

The Information : Cursor Is Developing an AI Agent to Compete With Claude Cowork

Cursor Is Developing an AI Agent to Compete With Claude Cowork

The Takeaway
  • Cursor is making a push beyond coding to capture more business customers. The new AI agent, internally referred to as Sand, would act as a general-purpose personalized assistant.

Cursor is developing a general-purpose AI agent meant to compete with popular tools like Anthropic’s Claude Cowork, two people familiar with the project said, part of a broader push by the company to diversify beyond coding-focused tools.

Work on the new agent began after Cursor started leasing compute capacity in April from SpaceX’s AI unit, known as SpaceXAI, the people said, and comes ahead of SpaceX’s planned $60 billion acquisition of Cursor. The new agent could buttress the enterprise business of SpaceX’s AI unit once that deal closes. It follows the two companies’ joint introduction of a new model, Grok 4.5, meant to perform engineering tasks as well as general knowledge work.

The new Cursor product, internally referred to as Sand, is its first product aimed at more casual AI users rather than developers. It would act as a personalized assistant, performing tasks such as responding to emails or texts and organizing spreadsheets in addition to handling engineering work, the two people said.

Part of Cursor’s product team is working on testing and honing the project, which the company rolled out internally in late June, one person said.

It is unclear whether Sand will make it to market. Cursor typically rolls out a new product to staff before deciding whether to launch it, one person said. The company is also facing uncertainty due to its pending merger with SpaceXAI, which could alter Cursor’s product road map.

If it moves ahead the product would join an increasingly crowded field of AI work assistants. OpenAI announced an AI agent, called ChatGPT Work, on Thursday, that it says can automate presentations and perform research.

The SpaceXAI unit had struggled to catch up in coding after rival Anthropic cut it off from its models, prompting SpaceXAI to team up with Cursor on AI work. SpaceXAI has also seen leadership turnover and a string of engineer and researcher departures this year, including a wave of exits after its Cursor collaboration.

Founded in 2022, Cursor initially offered a coding editor powered by Anthropic and other outside companies’ AI models, and has since developed a GitHub competitor. It also released its own coding model, Composer, late last year. Cursor has focused primarily on using the coding data from its developer base to fine-tune open-source models.

Though Cursor was an early leader in coding tools, it has faced mounting competition, including from the same model developers it originally relied on to power its products. Cursor CEO Michael Truell told staff during an all-hands in May that the company sees its next growth opportunity in business users that aren’t developers, and that customers were asking for such products, a person close to the company said.

SpaceX’s pending acquisition of Cursor will give the coding startup a parent with significant resources. Cursor could also meaningfully add to SpaceX’s AI product revenue once the acquisition is complete. Cursor’s revenue had reached an annualized pace of $2.7 billion as of March. SpaceX, meanwhile, reported roughly $470 million—10% of its overall revenue—in AI solutions revenue in the first quarter, consisting of Grok subscriptions and business access to models.

SpaceX has meanwhile been spending heavily on AI infrastructure, with AI capital expenditures tripling to $7.7 billion in the first quarter. In the run-up to its June initial public offering, however, SpaceX inked major compute leasing deals with outside companies, including a $1.25 billion per month one with rival Anthropic. Those deals promise to significantly lift SpaceX’s AI-related revenue, although the company says the arrangements are flexible, and it could decide to take compute capacity back later if it needs it.

SpaceX’s acquisition of Cursor is set to close in the second half of 2026, but the companies haven’t disclosed the exact date. It’s also not clear how deeply Cursor is collaborating with SpaceXAI on the new general-purpose agent.

The two companies did work together on Grok 4.5, released as an update to Grok, SpaceXAI’s flagship model brand. SpaceXAI said in a blog post that the model is tailored for engineering tasks, as well as basic office work.

Truell told developers at a conference last month that the model, Grok 4.5, was trained on tens of thousands of xAI’s graphics processing units and has 1.5 trillion parameters, making it competitive with leading models from Anthropic and OpenAI.

Cursor’s new agent appears to be targeting the same business opportunity xAI tried to tackle starting last year, when it began working on its own general-purpose agent. Musk has said that project, called Macrohard, aims to create a digital white-collar worker.

Work on the Macrohard product largely stalled at xAI earlier this year, according to two people with knowledge of the project. Some of the work shifted to Tesla, which began working on the project in January as part of its $2 billion investment in the AI startup. Some members of Tesla’s Autopilot team have continued to work on the project in the months since under the name Digital Optimus, one person said.

9to5 : OpenAI unveils ChatGPT Work agent, GPT-5.6 models now available

OpenAI unveils ChatGPT Work agent, GPT-5.6 models now available

OpenAI held its second livestream this week today at 10 am PT. The video teased that the company was “introducing the next chapter for ChatGPT” today. OpenAI has openly discussed bringing Codex workflows into the main ChatGPT app for some time now.

Today’s release includes a new ChatGPT Work agent, an upgraded ChatGPT desktop app with Codex included, and a new hosted sites service for OpenAI customers.

GPT-5.6 Sol, Terra, and Luna are available today
Additionally, OpenAI says it will release GPT-5.6 today after first announcing the three new models, Sol, Terra, and Luna at the end of June. Here’s a recap of the new models:
  • Sol, OpenAI’s flagship model
  • Terra, their balanced model for everyday work
  • Luna, their fast and affordable model
The Sol, Terra, and Luna naming scheme is also new with GPT-5.6.
“In this new naming system introduced with GPT‑5.6, the number identifies a model’s generation, while Sol, Terra, and Luna identify durable capability tiers that can advance on their own cadence,” OpenAI said in June. “Together, the family gives people and developers clearer choices across intelligence, speed, and cost.”
The new models are rolling out over the next 24 hours.
GPT-5.6 includes a new ‘ultra’ acceleration mode
Here’s more from OpenAI on today’s GPT-5.6 model release:
GPT‑5.6 Sol sets a new standard for both intelligence and efficiency, achieving state-of-the-art results across coding, knowledge work, cybersecurity, and science while outperforming previous and competing frontier models with fewer tokens and at lower estimated cost. The result is stronger performance per dollar: more successful work for the same spend, or comparable results at a lower total cost. We also introduce a new way to accelerate the most demanding work: ultra is our highest-capability setting, coordinating multiple agents across parallel workstreams to finish complex tasks faster. Stronger computer use and design judgment make GPT‑5.6 Sol our most polished collaborator yet, helping it inspect, refine, and deliver ready-to-use results.
OpenAI highlights GPT-5.6 improvements around design sensibilities.
“GPT‑5.6 delivers a step change in design judgment,” according to the company.
“With only high-level direction, GPT‑5.6 creates tasteful, ergonomic, and functional interfaces. Its stronger computer-use capabilities let it inspect and refine the rendered result—not just generate the underlying code or content—so it can catch visual and functional issues and apply finishing touches before handing the work back.”
Here are availability details for the new models:
  • Chat: Plus, Pro, Business, and Enterprise users access GPT‑5.6 Sol through medium and higher effort settings. Pro and Enterprise users can also select GPT‑5.6 Sol Pro for the highest-quality results on complex tasks.
  • ChatGPT Work and Codex: Free and Go users access GPT‑5.6 Terra. Plus, Pro, Business, and Enterprise users can choose among GPT‑5.6 Sol, Terra, and Luna and set an effort level for each. max is available to all users with access to GPT‑5.6 in ChatGPT Work and Codex and can be toggled on in settings. In ChatGPT Work, ultra is available to Pro and Enterprise users. In Codex, it is available to Plus and higher plans.
  • API: Developers can access Sol, Terra, and Luna through the OpenAI API. In the Responses API, Programmatic Tool Calling lets GPT‑5.6 write and run programs in-memory that coordinate tools and process intermediate results, making it Zero Data Retention (ZDR) compatible. Multi-agent, initially available in beta, lets GPT‑5.6 run concurrent subagents and synthesize their work in a single request.
OpenAI also explains how GPT-5.6 is priced through the API.
“GPT‑5.6 is priced per 1M tokens across three model sizes,” OpenAI says. “Sol is $5 input / $30 output; Terra is $2.50 input / $15 output; and Luna is $1 input / $6 output.”
Future of ChatGPT announcement video
The video description explains what to expect from OpenAI’s announcement today:
Andrew Ambrosino, Jessica Liang, Ed Bayes, Lauren Gordon, Tejal Patwardhan, and Katy Shi join host Thibault Sottiaux to introduce and demo the new ChatGPT and GPT-5.6.


ChatGPT Work, an upgraded ChatGPT desktop app, and hosted sites
OpenAI is releasing ChatGPT Work, an all-new ChatGPT app, and a hosted sites feature.
  • ChatGPT Work is a new agent that’s available on the web, mobile, and desktop.
  • ChatGPT’s desktop app is getting a Codex-powered upgrade. ChatGPT for Mac now includes Chat, Work, and Codex.
  • And OpenAI is releasing a hosted sites feature for all paid users.

Meanwhile, OpenAI just released a new version of Codex with these release notes. This is the promised merging of Codex into ChatGPT. It looks like users will be able to use a single ChatGPT desktop app, but set preferences to favor Codex functionality, including the app icon.


Codex is now part of the ChatGPT desktop app on macOS and Windows. Existing Codex app users can update as usual and keep their projects, settings, and workflows. You can make Codex the default view and, on macOS, keep the Codex app icon.
### New features
– Edit Markdown and code directly in the app, use inline annotations, and ask Codex to revise selected content.
– Review GitHub pull requests in the sidebar, with reviewer feedback alongside the diff, without leaving the app.
– Work across repositories in one project.
### Performance improvements and bug fixes
– Made Computer Use faster with GPT-5.6.
– Made task activity and progress easier to follow while Codex works.
– Simplified plugin management by moving it into Settings.
– Improved mobile connection reliability and fixed video rendering for SSH projects.
– Additional performance improvements and bug fixes.
Additionally, OpenAI says GPT-5.4 will be retired on July 23 following the introduction of GPT-5.6. The existing GPT-5.5 models will remain available.
To summarize today’s desktop app changes:
  • The existing ChatGPT app is now ChatGPT Classic.
  • Codex is now the new ChatGPT desktop app. It still looks like Codex and includes the Codex icon as an option, but it’s now called ChatGPT.
  • ChatGPT for desktop includes ChatGPT Work and ChatGPT Codex, which share plug-ins. ChatGPT Codex mode shows more technical details that ChatGPT Work abstracts away from the user.
  • It’s possible to have ChatGPT Classic, ChatGPT, and Codex installed, but the way forward seems to be just running the new ChatGPT desktop app. Codex users can still use the Codex app icon, but the app will be called ChatGPT.
You can learn more about ChatGPT Work and the other announcements here.

Le Figaro : Casino : pourquoi la partie de poker menteur entre Daniel Kretinsky

Casino : pourquoi la partie de poker menteur entre Daniel Kretinsky et les créanciers va durer tout l’été

Faute d’accord après huit mois de négociations, le sort du distributeur devrait être tranché par le tribunal de commerce de Paris, en octobre. Mais les parties ont, en fait, intérêt à s’entendre, tant les risques sont élevés.

L’avenir de Casino se jouera-t-il au tribunal de commerce de Paris cet automne ? Depuis quelques jours, cette option semble la plus probable. La dette du distributeur stéphanois (Monoprix, Franprix, Vival, Cdiscount…) est devenue insupportable et le groupe sera incapable de rembourser le 1,4 milliard d’euros exigible en 2027. Daniel Kretinsky, qui détient 53 % de Casino depuis la précédente restructuration financière, en mars 2024, n’a pas réussi à trouver un accord avec les principaux créanciers sur un allégement de cette dette, indispensable pour éviter la faillite.

Après huit mois de négociations, Casino, qui ne vaut plus que 85 millions d’euros en Bourse, a officialisé ce désaccord lundi, en publiant les deux propositions de restructuration (celle de Daniel Kretinsky et celle des créanciers) remises le mardi 30 juin, date limite pour un accord. Le groupe a prévenu que son conseil d’administration déciderait laquelle il présenterait au tribunal de commerce. Selon les observateurs et les acteurs de ce dossier, aussi complexe que sensible, il ne fait aucun doute que le board choisira le plan du principal actionnaire, même si le comité ad hoc d’administrateurs, constitué pour émettre une recommandation, ne compte pas de représentant de Daniel Kretinsky.

À lire aussi Casino : les enjeux de la partie de poker menteur entre Daniel Kretinsky et les créanciers

Passer la publicité

Fort du soutien des banques qui financent l’activité quotidienne du groupe à hauteur de 1,2 milliard d’euros, le principal actionnaire propose d’effacer 600 millions de dette, d’apporter 550 millions d’argent frais et de renforcer son contrôle : sa participation passerait à 66,8 %. Les créanciers obligataires proposent, eux, un abandon de la quasi-totalité de la dette en échange de la prise de contrôle du groupe, ce qui évincerait Kretinsky. Et ils assurent avoir une alternative pour assurer le financement opérationnel. Bref, les positions officielles des deux clans sont diamétralement opposées.

En fait, rien n’est figé, et la date butoir du 30 juin était seulement indicative. D’intenses négociations se sont d’ailleurs poursuivies entre le mercredi 1er juillet et le dimanche 5, sous l’égide de Marc Sénéchal, le conciliateur nommé en mars par le tribunal de commerce. Les créanciers obligataires ont proposé une nouvelle offre. Intraitable, Daniel Kretinsky a fait savoir, à la fin du week-end, qu’il la refusait. Ce projet n’est donc plus sur la table et les deux offres soumises au conseil d’administration de Casino sont celles du 30 juin.

«Daniel Kretinsky refuse tout » et « les créanciers sont trop agressifs »
« Daniel Kretinsky refuse tout », s’étrangle-t-on du côté des créanciers, qui incluent les fonds Monarch et Attestor (ils détiennent à eux deux 40 % des créances) ainsi que ReadyState. « Les créanciers sont trop agressifs, rétorque-t-on du côté de l’actionnaire. Il y a un biais cognitif qui va à l’encontre d’un accord consensuel. » Et sans accord consensuel, la justice peut imposer une solution à toutes les parties prenantes. Dans ce cas contraire, le dépôt de bilan est inévitable.

Le conseil d’administration de Casino, présidé par l’ancien ministre Laurent Pietraszewski, devrait choisir entre les deux offres et « formuler une proposition » dans les prochains jours. Celle-ci sera soumise aux votes des actionnaires, des créanciers et des banques, mais pas avant septembre. Le dossier arrivera ensuite sur la table du tribunal de commerce de Paris, probablement mi-octobre. C’est lui qui devra décider de valider le plan, et donc de l’imposer aux créanciers, qui auront très probablement voté contre.

« Il y a encore beaucoup trop d’incertitudes à lever »
À en croire l’entourage de Daniel Kretinsky, la voie est libre pour que son projet soit mis en œuvre. « En réalité, il y a encore beaucoup trop d’incertitudes à lever, et toutes les parties ont intérêt à trouver un accord avant la décision du tribunal de commerce », assure un bon connaisseur du dossier. Pour commencer, il est risqué de spéculer sur le résultat d’une décision de justice. Ensuite, cette dernière est menacée de recours : les fonds anglo-saxons détenteurs d’obligation menacent déjà de faire appel, et même d’aller jusqu’en Cour de cassation. Enfin, le soutien des banques de Casino (Société générale, BNP Paribas, BPCE…) au projet de son principal actionnaire est conditionné à l’accord des deux tiers des créanciers obligataires.

L’entourage du milliardaire est convaincu que les banques renonceront in extremis à cette condition. « À quelques mois de la présidentielle, elles ne prendront pas le risque de mener Casino et ses 28 000 salariés à la faillite », confirme un proche des banques. Pour faire pression sur les autres acteurs, ses proches martèlent que Daniel Kretinsky est désormais un investisseur institutionnel et une figure du capitalisme français bien établie. Premier actionnaire de Fnac Darty, troisième de TotalEnergies, il détient Editis et les hebdomadaires Elle et Marianne. Daniel Kretinsky serait donc l’opposé des hedge funds anglo-saxons qui détiennent la dette, soulignent-ils, rappelant qu’il a déjà injecté, et perdu, 1,2 milliard d’euros dans Casino. Dans ces conditions, pas question de remettre au pot sans renforcer son contrôle. « Le groupe n’est pas menacé. Il est prospère, stable et soutenu par ses banques partenaires. Son bilan sera rééquilibré, soit par un accord soit par une décision du tribunal », prévenait fin mai Daniel Kretinsky dans Le Figaro.

« Ils se détestent tellement qu’ils ne rêvent que d’une chose : écraser leur adversaire »
Toutefois, en voulant tordre le bras aux créanciers avec le soutien de la justice, il fait un pari très risqué. « Si le plan de Daniel Kretinsky passe, cela enverrait un mauvais signal pour l’attractivité de la place de Paris, martèle un proche des fonds, particulièrement remonté. En matière de restructuration, les créanciers sont prioritaires sur les actionnaires. Notre proposition respecte les textes. » S’ils prennent le contrôle, les créanciers obligataires promettent d’apporter 400 millions d’euros d’argent frais et assurent, sans convaincre, avoir une solution pour financer de la dette opérationnelle détenue par les banques. « Ce serait Noël pour le groupe », lâche-t-on dans l’entourage des créanciers.

Tout comme Kretinsky, ces derniers ont aussi intérêt à un compromis avant un jugement du tribunal du commerce. « Quoi qu’ils disent, ils ne veulent pas devenir actionnaires à long terme de Casino, assure un proche du dossier. Leur vocation est de revendre tout de suite leurs actions après une restructuration de la dette, et surtout pas d’en prendre le contrôle et de réinvestir. » Certains brandissent la menace d’une vente à la découpe si ces fonds, qui ont apporté des engagements de non-cession pour deux ans seulement, prenaient le contrôle du distributeur. « Ce serait une catastrophe industrielle », avertit un protagoniste.

Alors que tout indique qu’un accord négocié entre Daniel Kretinsky et les créanciers est meilleur qu’une solution imposée par la justice, reste à savoir si les acteurs du dossier seront raisonnables. « Ils se détestent tellement qu’ils ne rêvent que d’une chose : écraser leur adversaire », assure un acteur du dossier.