TheChek News Flow

Satya Nadella says the attempt to remove Sam Altman from OpenAI was ‘amateur city’

Microsoft chief explains his decision to back AI lab’s boss in 2023 coup attempt during testimony in Elon Musk’s lawsuit

Microsoft chief executive Satya Nadella said the OpenAI board’s abortive attempt to oust Sam Altman in 2023 was “amateur city” as he took the stand in the high-stakes legal battle between the AI lab and Elon Musk.

Nadella told the jury on Monday he never received “a specific answer” about why the board fired Altman, citing concerns about dishonesty, before reinstating him days later.

The episode has cast a shadow over Altman’s reputation — something Musk has exploited, branding him “Swindly Sam” as he pursued the case arguing OpenAI betrayed its charitable mission.

OpenAI’s board said at the time that Altman was removed because he was “not consistently candid”, something Nadella agreed would be grounds to fire a CEO.

But, questioned by Musk’s lawyers about why he supported Altman, Nadella said: “Whenever I asked specifically why Sam was fired, [the board] never gave me a specific answer.”

“It was amateur city as far as I’m concerned,” he said.

Microsoft’s CEO was testifying in a case Musk brought in which the $3tn software giant — OpenAI’s largest investor — is accused of “aiding and abetting” the AI lab and Altman in selling out its non-profit mission to become a for-profit business.

Victory for Musk, who made early financial contributions to OpenAI as a non-profit, could upend the $852bn start-up’s plans to go public this year.

Nadella’s testimony, and internal messages presented in court, also betray his concerns about the software group’s strategic position as OpenAI’s popularity grew.

Ahead of a $10bn investment in OpenAI, Nadella emailed Microsoft executives in 2022 “to figure out how to continue the partnership in a more aligned way”.

“I don’t want to be IBM and OpenAI to be Microsoft,” he wrote.

IBM struck a deal in the 1980s for a fledgling Microsoft to build an operating system for its personal computers that ultimately led the software group eclipsing the computer maker.

Asked what he meant by the email, Nadella said he wanted to avoid “slipping into a situation . . . where Microsoft [did not have] its own self-sufficiency”.

A few months later, Nadella emailed Microsoft’s chief financial officer Amy Hood to flesh out the concerns: “Right now we are a very thin layer on top of Nvidia and all the IP is with OpenAI . . . if we are going to spend this kind of money and not have control of destiny, it makes no sense.”

Microsoft has since struck several deals with OpenAI to restructure their relationship and tried to increase its independence in AI.

The Big Tech company now owns about 27 per cent of OpenAI, valued at just over $200bn.

According to disclosures in the trial, in 2016 Microsoft was offering OpenAI heavily discounted cloud computing access. By 2018, with OpenAI’s computing bills increasing, “we were not comfortable thinking of it as a marketing expense”, Nadella said.

“This was one you started to think ‘enough of the charity, on with the business’?” asked Steven Molo, an attorney for Musk. “Yeah,” Nadella responded.

Microsoft invested $1bn in 2019, $2bn in 2021 and a further $10bn in 2023, helping to support OpenAI’s transition since 2015 from a small research lab into one of the most valuable for-profit companies in the world.

“[The investment] has worked out very, very, very well for Microsoft hasn’t it?” Molo asked Nadella on Monday. Microsoft’s CEO replied: “Because we were the only ones who took the risk.”

Hudson River Trading generates $6.4bn in first-quarter trading revenue

Volatile markets fuelled by Iran war help to propel proprietary trading firm

Hudson River Trading brought in $6.4bn in trading revenues in the first quarter of the year as the ascendant rival to market leader Jane Street seized on turbulent markets.

Hudson River made $4.2bn in net profit on those revenues, according to a person familiar with the matter. Wider volatility in markets contributed to the results at the New York-based firm, which employs about 1,000 people and uses quantitative and algorithmic technology as a market maker across global assets.

The US war in Iran has fuelled large swings in the price of oil as most ships are blocked from passing through the Strait of Hormuz, a crucial transport channel for commodities. The fallout has spread to other financial markets such as Treasuries and currencies as investors grapple with longer-term consequences to the global economy. 

Volatile markets are ripe environments for market-makers, which profit from differences in price as they buy and sell assets. Proprietary firms such as Jane Street and Hudson River have posted record-breaking results over the past year to become dominant players in that business. 

In 2025, Jane Street, Hudson River and Citadel Securities made more than $60bn in trading revenues. 

These secretive firms have often outstripped comparable desks at major banks. Hudson River’s net trading revenue in the first quarter was greater than that of Bank of America or Wells Fargo in the same period.  

Meanwhile, Jane Street made record revenues of $16.1bn in the first three months of the year, with net income of $10.3bn. Those results beat the trading desks at JPMorgan and Goldman Sachs, which also benefited from whipsawing in markets. 

Jane Street’s latest results were also driven in part by a private portfolio of stakes in businesses such as AI lab Anthropic and Thinking Machines Lab, which has contributed to record earnings as the valuations of the AI groups surge.

Hudson River was founded in 2002 by partners with a speciality in mathematics and computer programming. The firm uses automated algorithms to price across many global markets and it has recently made big investments in AI and the computational resources required for trading models. 

Hudson River declined to comment. Bloomberg News first reported the results.

UK to take full ownership of British Steel

Government has been locked in talks with Chinese group Jingye over lossmaking business

The UK is to take full ownership of British Steel, Prime Minister Sir Keir Starmer said on Monday as he set out why the Labour Party should back him as leader as pressure rises for him to quit.

The government last year used emergency legislation to take control of the steelworks amid fears that Chinese owner Jingye was running down its site in Scunthorpe after halting coke imports.

Starmer said the government and Jingye had not been able to agree a commercial sale. “So I can announce that legislation will be brought forward this week to give the government powers, subject to that public interest test, to take full national ownership of British Steel.”

Jingye has continued to have economic control of British Steel, meaning the government has been legally unable to sell all or part of British Steel or make decisions about its strategy. Jingye had initially demanded more than £1bn in compensation from the UK government and rejected a £100mn offer in February.

The government said it now “does not believe an agreement could be reached which would deliver acceptable value for money for taxpayers”.

But public ownership of British Steel still could take some months and there are hurdles to overcome, not least the threat of a potential legal challenge by Jingye.

Peter Ware, head of government at law firm Browne Jacobson, said full nationalisation would require primary legislation addressing compensation for the former owner.

“The compensation question is particularly complex: with the government having already committed over £400mn in working capital, Jingye’s scope to claim substantial compensation may be limited, but legal challenges under bilateral investment treaties or domestic property rights principles cannot be ruled out.”

The nationalisation plan will be included in the King’s Speech on Wednesday alongside new laws to bolster Britain’s relationship with the EU and impose an organisational shake-up in the NHS.

Starmer made the pledge on Monday as part of efforts to save his ailing premiership after Labour’s worst local election results in history last week.

The FT reported in March that the government was considering how to use legislation to take legal control of British Steel.

Starmer said on Monday: “Steel is strategically important to our economy and our national resilience. That’s why we acted last year to avoid a sudden halt to production at Scunthorpe, protecting workers and the community that depend on the site, and why we’re now bringing forward legislation to give us options to protect Britain’s steelmaking capability.

“This is what an activist state looks like, taking decisions in the national interest.”

Unions welcomed a nationalisation of British Steel for giving clarity to workers who have been in limbo for more than one year.

Charlotte Brumpton-Childs, GMB national secretary, said: “Unions have long known Jingye will not negotiate in good faith.

“This legislation will cover the whole steel industry — it isn’t specifically for British Steel but it is what will protect it from foreign owners. British Steel is a nationally strategic asset, it is right the government does everything in its power to secure its long-term future.”

Heavily lossmaking British Steel is costing the government about £1.3mn a day to run, according to the National Audit Office. It has already cost £377mn between last April and the end of January to keep the business operating.

However, the site is seen as important to UK supply chains as it is home to Britain’s last two operating blast furnaces and produces 95 per cent of the steel used on UK rail tracks.

The Community and Unite unions said: “British Steel has a bright future, with a world-class highly skilled workforce making strategically important steels for the UK’s rail and infrastructure. The business also has the potential to expand and become a vital part of the supply chain for other UK steel sites.”

By nationalising British Steel, the government will now be able to take control of company strategy and press ahead with a shift from using blast furnaces to a greener electric-arc furnace.

Business secretary Peter Kyle said: “Revitalising our steel sector is a top priority for this government, and bringing forward this legislation would allow us to explore potential future options for British Steel. The government recognises that securing the long-term future of the UK’s steel sector relies on both public and private investment for modernisation.”

The apparently insatiable demand for AI compute has data center entreprenuers looking to the stars. There’s a key problem: There aren’t enough rockets to put data centers in orbit around the Earth, and they’re too expensive.

Most of the players are hoping that SpaceX’s Starship — expected to make its twelfth test flight as soon as this weekend— will solve the problem. But once the vehicle is operational it may be years before it is commercially available, given SpaceX’s internal satellite business. The same is true for Blue Origin’s New Glenn rocket, which failed to deliver a satellite during its third launch in April.

That leaves space data center schemes either targeting the mid 2030s, like Google Suncatcher, or preparing to start off doing edge processing tasks for space sensors, like Starcloud.

In theory, there’s a third way: “We’re standing up our own rocket program,” Baiju Bhatt, the CEO and founder of Cowboy Space Corporation, told TechCrunch. He expects the first launch before the end of 2028.

Today, the company announced the closure of a $275 million Series B round at a post-money valuation of $2 billion, led by Index Ventures, as a downpayment on that work. Breakthrough Energy Ventures, Construct Capital, IVP, and SAIC also participated.

Bhatt, a co-founder of online stock platform Robinhood, launched this startup in 2024 as Aetherflux, with plans to collect abundant solar energy in space and beam it down to Earth. The idea of space data centers led the company to pivot towards using its electricity while in orbit. The practical realities of that effort, in turn, led him to a rocket development program, and the company’s new name.

Bhatt said he spoke to multiple launch providers to try and find a path where his company would only build satellites, but he couldn’t find enough launch capacity to truly scale an orbital data center business, or do so in a way where the unit economics could compete with terrestrial alternatives.

This Week Only: Buy one pass, get the second at 50% off

“There’s a lot of new rockets that are coming online, but as we look three, four years out, it’s still very, very scarce, and I think that you’re going to see a lot of the first party rocket providers actually specialize into their own payloads,” Bhatt said.

Of course, while bringing the rocket in-house is logical, it’s also nuts. Only a handful of private companies in the West, mainly SpaceX, Rocket Lab and Arianespace, are consistently launching commercial rockets. Two others, Blue Origin and United Launch Alliance, have been struggling to drag their vehicles out of development hell for years. A number of startups, including Stoke Space, Firefly Aerospace, and Relativity Space, have worked for years and are still waiting to deliver operational systems.

This evolution of the company will also bring Cowboy Space Corporation into direct competition with SpaceX and Blue Origin, the most advanced and well-funded players in the market.

“The prize here, and the size of this market, is big enough that there’s room for many players to succeed,” Bhatt said “I see the demand for AI getting more and more acute, and I see the options on Earth getting more and more limited.”

One advantage, Bhatt argues, is the company’s focus on this single market (data centers), and its unique design. Orbital rockets typically have a booster stage that flies the vehicle to the edge of space, and a second stage that carries the payload and delivers it to orbit. Cowboy Space plans to build its data centers directly into the second stage of its rocket. It’s actually a bit of a throw-back: The first US satellite, Explorer 1, was built as the final stage of a rocket, filled with radio equipment and a few scientific instruments.

Making the rocket purpose-built only to launch its data-center satellites should simplify the design process. The company expects each satellite to have a mass of 20,000 to 25,000 kilograms and to generate 1 MW of power for just under 800 onboard GPUs. That means its rocket would be slightly more powerful than the SpaceX’s workhorse Falcon 9, though still smaller than its under-development Starship. Eventually, Bhatt says, he expects the booster to be reusable.

Cowboy Space has hired veterans of the space industry, including former Blue Origin propulsion engineer Warren Lamont and former SpaceX launch director Tyler Grinne. The company also plans to build its own rocket engine, the most complex and expensive part of any launch vehicle. Cowboy Space is still working through key development needs, like facilities to test, manufacture and launch its rockets.

The new vision comes with a new name for the startup, to emphasize its mission to “power humanity from the high frontier,” although Bhatt admits “it gives me a reason to wear a cowboy hat and also grow this sick mustache.”

How many AI chip designers can the public market support? We’ll get a sense of that on Thursday when Cerebras Systems, the biggest of a new generation of chip designers dedicated to AI, is expected to go public at a valuation of $35 billion. If it follows the pattern of other AI-related IPOs, such as cloud firm CoreWeave, Cerebras will be a big hit.

CoreWeave went public a year ago at $40 a share and, despite a lot of ups and downs, it closed on Friday at $114. (Most other tech IPOs from last year, that had no AI exposure, have had a very different experience.) CoreWeave has been well received even though it is burning through billions of dollars in cash right now as it ramps up its network of data centers. Investors’ favorable reaction to it suggests Wall Street will look past the fact that Cerebras is also burning cash and focus on its AI exposure. (Demand for stock in the IPO is also strong, Bloomberg reported on Friday).

Indeed, we are at a moment of peak AI infrastructure optimism, with shortages of computing capacity apparent constantly (hence Anthropic’s deal with SpaceX this past week). Of course, the capacity shortfall is an argument for CoreWeave, whose customers include Anthropic, Microsoft, OpenAI and Meta Platforms, but not necessarily for Cerebras. Is there the same kind of demand for another AI chip supplier?

Buyers of chips—mostly cloud firms and big AI developers—have plenty of choice already. Aside from Nvidia, which dominates the field, there are AI chips from Advanced Micro Devices, Google and Amazon. Meta and Microsoft are developing their own chips. Chinese firms have Huawei and many others. So what’s Cerebras’ pitch?

Cerebras does have OpenAI as a customer (and shareholder), even though OpenAI is also developing its own chip in concert with Broadcom. Cerebras also struck a deal to supply its chip to Amazon Web Services as a supplement to Amazon’s Trainium AI chip. But there’s no getting around the fact that the chip market is crowded. That’s perhaps why some other AI chip startups have effectively sold themselves: Groq licensed its tech to Nvidia, which hired the startup’s top people, for instance. And Meta hired a key group of engineers from another chip startup, Graphcore.

Of course, a takeover of Cerebras might still happen—after the IPO.

Gapping down
In reaction to earnings/guidance:

Other news:

Gapping up
In reaction to earnings/guidance:

Other news: