TheChek News Flow

Long-serving pharma chief Pascal Soriot hopes research and risk-taking will secure growth in China and US

When the AstraZeneca board rejected a near £70bn bid from Pfizer that would have created one of the world’s biggest pharmaceutical companies, chief executive Sir Pascal Soriot understood the gamble they were taking.

The approach more than a decade ago offered a lifeline for struggling AstraZeneca, which Soriot had then led for less than two years, and shareholders were excited.

Yet Soriot was always confident walking away was the right move. As he rang the opening bell at the New York Stock Exchange last month to celebrate the company’s recent direct listing on the bourse, he said the decision “quite frankly . . . has worked better than we could have imagined”.

Pfizer’s final offer in 2014 was £55 a share. AstraZeneca now trades at about £128 a share and its £207bn market cap makes it more valuable than its would-be buyer. “We thought we had what it took to create more value on our own,” Soriot tells the FT.

Such calculated risk-taking has characterised the 67-year-old’s time at the company he joined in October 2012. To meet ambitious growth targets, he has focused on in-house drug development, US expansion and investing in emerging markets before it became fashionable in the sector, making AstraZeneca the largest western pharma company in China.

A French-Australian national who studied veterinary medicine in Paris before moving to New Zealand and Australia in his twenties, Soriot has credited his appetite for a fight in part to a childhood in Paris’s banlieues, the gritty suburbs outside the city.

AstraZeneca hired him as chief executive from Swiss rival Roche in a period of turmoil. His predecessor resigned under pressure from shareholders unhappy with setbacks in late-stage trials aimed at replacing blockbuster drugs losing patent exclusivity.

Soriot’s response was, and continues to be, investment in research and development, particularly in two countries — China and the US — at the vanguard of pharmaceutical innovation. He has committed billions of dollars in China — in January announcing $15bn of investment by 2030 — to build manufacturing capacity, strike deals with local biotech firms and establish a new R&D site in Beijing. That research pipeline is expected to funnel medicines for sale globally and especially in the US, a growth market that brought in 43 per cent of group revenue last year.

The approach requires walking a geopolitical tightrope that has recently become more treacherous. As tensions rise between China and the west, US politicians are seeking to slow down investment in the rival superpower, proposing legislation to make it more difficult for drugs licensed from Chinese biotechs to be sold and approved in the US.

Industry experts caution that such moves are unlikely to materialise soon, but the possibility is a clear concern. Soriot says the challenge is to “tap into the innovation” in China while “navigating the geopolitical environment and making sure to operate in a way that is acceptable to governments”. Last year the company announced a $50bn investment commitment in the US, and it became the first foreign drugmaker to strike a tariff-exemption deal with the Trump White House.

Soriot describes AstraZeneca’s R&D strategy as delivering on a “fundamental ethical and moral responsibility . . . to identify new treatments and bring them to patients”, who ultimately do not care about the origin of life-saving cancer drugs. “If a drug is invented in China,” he says, “people in the US and UK . . . that can be treated with new medicines will want the new medicines.”

An ambitious revenue target of $80bn by 2030, up from $58.7bn last year, is also contingent on growth in the US. Set in 2024, it is a further stretch from 10 years previously, when Soriot announced a goal of $45bn revenue by 2023.

Back then markets were sceptical, says Sean Conroy, an analyst at Shore Capital. But AstraZeneca delivered, an achievement Soriot credits with walking away from Pfizer and spending on R&D. “If we merged, it was going to disrupt our science,” he says. “Of course, there was a risk . . . But we thought we owed it to our employees, shareholders [and] patients to say no and to focus on our own pipeline.”

He says the company’s “defining focus” remains on “science and recruiting the best people”, in R&D but also “the commercial side . . . Our medicines only make a difference if they are brought to physicians.”

AstraZeneca is now among the largest spenders in the industry, ploughing $14.2bn, 24 per cent of revenue, into R&D last year. Last month it edged closer to joining the competitive field of weight-loss treatments after positive data from a mid-stage clinical trial of a pill the company is developing.

It has also made some well-timed acquisitions, notably the 2021 purchase of Alexion for $39bn, which allowed the company to launch a business in rare disease medicine, earning $9.1bn, 16 per cent of last year’s revenue.

The plan has not always delivered straightforwardly. When a combination drug for lung cancer failed a key trial in 2017, it wiped £10bn off the company’s value and led to questions from investors about the chief executive’s future. But the drug has since proved effective in treating other tumours, and last year brought in $6.1bn in revenue.

Soriot admits AstraZeneca would have been in “trouble” if such bets had failed to deliver. And the company needs its investments to keep paying off.

One setback came this week as its share price fell sharply on Thursday after a nerve disease drug failed to meet its target in a late-stage trial to treat a heart condition.

Three of AstraZeneca’s most valuable drugs lose patent exclusivity after 2030, opening nearly $20bn of sales to generic competition. The group is halfway to a target of launching 20 new medicines by 2030 to hedge against this. “Pharma companies constantly have to reinvent themselves,” says Conroy. “They will have to keep investing in R&D.”

Described as “mercurial” and as having an “edge” by one executive who previously worked for him, Soriot’s work ethic — one he describes as “casual intensity” — is well known. “He hates being at the office,” says the former executive. “He likes being on the manufacturing sites and understanding how things work from the ground.”

In the past year his pragmatism has sometimes challenged a relationship with the UK, where the company is based and regarded as something of a national asset. Soriot was knighted in 2022 for services to UK life sciences, and particularly its Covid-19 vaccine development work, which made it a household name.

But the decision last year to bring AstraZeneca’s US listing on a par with London and Stockholm sparked concerns a company considered a British crown jewel could eventually leave London. That compounded worries that investment in the UK was not keeping up with China and the US. The company ditched a vaccine plant expansion near Liverpool, after failing to reach an agreement with the government over the level of state support, and last year paused expansion in Cambridge and the development of a new lab in the northern English town of Macclesfield.

It said in April it would revive the Cambridge and Macclesfield projects following a UK-US trade deal, which Soriot described as “progress” and which included measures to increase NHS spending on medicines. He is adamant any AstraZeneca success in the US is a net positive for the UK, where it employs thousands of people. He has also been vocal about the UK — and the rest of Europe — needing to increase spending to compete with China and the US for investment, and to attract companies.

“There are many places around the world, including emerging markets, that want to do clinical trials and R&D, so you have a lot of competition,” he says, adding AstraZeneca works with almost every country globally. “If you do this, you have to find a way to work constructively with whoever is leading the country . . . From country to country, the political systems are different. At any time you can have a government that is left- or right-leaning.”

As for the leadership of AstraZeneca, Soriot acknowledges scrutiny over who might replace him, though shows no sign of slowing down.

“Everyone focuses on when am I going to retire,” he says. “I’m not running out of excitement . . . But of course biology will catch up with me at some point. That affects all of us and we have a great focus on succession planning like any good company should.”

ByteDance later became a global giant through the success of its short video app, Douyin, in China and TikTok’s popularity overseas. (Last year ByteDance sold part of TikTok’s U.S. subsidiary in charge of data security to a venture controlled by U.S. investors to resolve longstanding national security concerns about TikTok in the U.S. ByteDance still controls TikTok’s commercial U.S. operations.)

SIG China owned 15% of the TikTok parent, The Information reported in 2020, although it’s unclear how many shares it owns now. But every sign suggests it remains a major shareholder. Arthur Dantchik, one of SIG’s co-founders, still sits on the board of ByteDance. Jeff Yass, another co-founder, is a major donor to President Donald Trump and the Republican party.

Moreover, a Susquehanna affiliate is an investor in the new TikTok USDS joint venture, and a Susquehanna executive, Mark Dooley, sits on the joint venture’s board. At least one of the SIG funds that holds ByteDance shares was due to expire last year, before SIG extended its lifetime for another year, according to one of the people.

The regulatory and geopolitical challenges ByteDance has faced, in both China and the U.S., helps make the company’s path to an initial public offering uncertain.

SIG was founded in 1987 by a group of six college friends who met at the State University of New York at Binghamton. The Bala Cynwyd, Pa.–based company is low profile and subdued compared to the glamorous blue chip firms on Wall Street or in Silicon Valley. After making a fortune from high-frequency and quantitative trading, SIG started doing venture investments, beginning in China in 2005.

The U.S. venture arm, Susquehanna Growth Equity, was established one year after the China arm. In 2014, SIG also set up a fund dedicated to Japan. The private equity and venture arms are fully funded by SIG itself, with no outside backers.

Aside from the early ByteDance investment, other notable deals by SIG China include film studio Bona Film Group, music platform Ximalaya, online game company Giant Network and restaurant chain Country Style Cooking.

Gong, who has a doctorate in electrical engineering from Princeton University, worked for telecom firm UTStarCom and also founded his own startup before joining SIG China. He is already looking for deals for his new fund and has already made one investment, in an Asian prediction market app called Bagel, which is led by a Chinese founder, according to one of the people with knowledge of Gong’s new fund.

Worries the president’s new plane lacked defensive capabilities as well as an Israeli warning about a possible threat

A lack of defensive capabilities aboard President Trump’s new luxury jet led him to fly home from Turkey aboard an older Air Force One plane this week after U.S. officials received Israeli intelligence about a possible plot to kill him, according to U.S. officials.

That threat was deemed not to be entirely credible, according to some of those officials.

But the information, combined with Trump’s proximity to Iran as hostilities were resuming, was enough that the Secret Service, the White House’s military office and some national security staff determined Trump’s preferred plane—a jet gifted by the Qatari government and overhauled by the U.S. Air Force—wasn’t adequate to fly the president home, the officials said.

Trump ordered new military strikes on Iran earlier this week after Iran attacked ships in the Strait of Hormuz, raising concerns that Tehran would retaliate against U.S. allies across the Middle East.

The president flew the newer, overhauled jet to attend a North Atlantic Treaty Organization summit in Ankara, Turkey, earlier this week. But he switched planes en route to Washington, flying the older version of Air Force One to the U.K. before re-boarding the new plane to fly home.

Trump denied Wednesday there was a security concern that prompted the decision to switch planes, insisting the new Qatari-donated plane was headed to the U.K. so troops stationed there could see it. When asked whether he was aware of any credible threat from Iran against Air Force One, the president replied that he faces threats “all the time” and that he is “number one on their list.”

White House communications director Steven Cheung said the luxury plane “has been fitted with high-level security protocols” to ensure the president’s safety.

The Wall Street Journal reported Thursday that Israel shared new intelligence with the U.S. that it said indicated a fresh Iranian plan to kill Trump. The New York Times reported earlier that Trump flew aboard his older plane as a security precaution.

Trump said on social media on Friday night that should Iran act on the threat to assassinate him, “Orders have already been given, and the U.S. Military is ready, willing, and able, for a one year period of time, subject to extension, to completely decimate and destroy all areas of Iran.”

Some U.S. officials said they were concerned that Israel shared the intelligence to influence U.S. decision-making toward a return to all-out war with Iran, noting that one piece of intelligence in isolation often provides a fragmentary look at a possible threat. Iran knows that targeting Trump could lead to a more robust military response, officials said.

While Trump has shown a preference for diplomacy, Israel would like to see a return to the bombing campaign to continue degrading Iran, according to people familiar with the matter. Israeli military officials have continued to plan for additional targets in Iran since the end of the campaign, according to military officials, even as Trump has sought the diplomatic track.

A spokeswoman for the Israeli Embassy said Israel shares intelligence with the U.S. because of the close partnership between the two countries and not to advocate for a hidden agenda like advocating for war. Israeli Prime Minister Benjamin Netanyahu’s office didn’t respond to a request for comment.

“The Cease Fire is OVER!” Trump wrote Friday on Truth Social while noting the U.S. had agreed to further negotiations with Iran.

Air Force Gen. Dale White, who oversaw the U.S. military effort to overhaul the Qatari plane into one that could transport the president, said in an interview earlier this year that the service didn’t take any shortcuts when it comes to security aboard the plane. But the Air Force also said in a June statement that the government “made trades on some of the less commonly used mission sets.”

The precise defensive capabilities of Air Force One are classified.

The jet was gifted to the U.S. by the Qatari government last year after Trump complained that the planes he was required to fly on as president were too old and small. Former officials said there are significant engineering challenges to overhauling a plane to be capable enough to carry a president in complicated threat environments.

Frank Kendall, who served as U.S. Air Force secretary during the Biden administration, said it should take three to four years to retrofit a 747 plane with a full suite of defensive capabilities required to keep the president safe.

“It’s a lot of major modifications to the airplane—it’s not simple,” Kendall said. “You have to think of the airplane as just a box into which you put an awful lot of things to make it Air Force One. The airplane cost about $250 million or so. Air Force One cost about $2 billion, so there’s a huge difference.”

Current and former government officials predict Trump will likely use the Qatari plane less in overseas travel. “I would have expected them to only use it in the United States, so I was a little surprised when they took it to Turkey,” Kendall said.

The expected Iranian response and broader conflict with Iran was also a consideration for the plane switch, the officials said.

But tensions between the companies soon began to mount as OpenAI embarked on an effort to build its own family of AI-powered hardware devices. It didn’t help that OpenAI did so by hiring high-profile former Apple leaders, including Tan and Jony Ive, Apple’s former chief design office. Last year, OpenAI paid $6.5 billion in an all-stock deal to acquire io Products, which was co-founded by Tan and Ive (Ive works as an independent contractor for OpenAI).

Signs of a rupture between OpenAI and Apple began to appear in May when OpenAI threatened a lawsuit of its own against Apple over alleged breach of contract with its disappointing 2024 ChatGPT integration into Apple’s AI platform. An OpenAI official told The Information at the time that the company was considering legal action. A suit hasn’t materialized yet.

Apple’s lawsuit could cast a shadow over OpenAI’s plans to launch its first hardware device, which could happen as early as February. It could also cloud the company’s plans to go public by raising uncertainty among potential investors. The company has filed paperwork confidentially with regulators for an IPO, but the exact timing of an offering is still up in the air. In the past, Apple’s court battles with other power tech companies like Samsung and Qualcomm have taken years to resolve.

In a statement an Apple spokesperson said: “We will always defend our teams’ hard work and innovations, and we are taking all appropriate steps to do so.”

In its lawsuit, Apple, one of the most secretive companies on the planet, chronicles in vivid detail how its former employees allegedly violated their confidentiality agreements with the company.

When Liu left in January to join OpenAI, for example, Apple said that it contacted Liu to remind him of the confidentiality agreement, schedule an exit interview and confirm he had returned his company devices. Lui didn’t respond to Apple, according to the lawsuit. Apple said it discovered that Liu had held onto a work laptop after leaving the company.

While at OpenAI, the lawsuit says, Liu exploited a previously unknown authentication bug with Apple’s corporate systems to access and download dozens of confidential hardware-related documents to his Apple work laptop. Apple said Liu then coached Apple employees, whom he was recruiting to OpenAI, how to avoid Apple’s security team when copying Apple files.

While investigating Liu, Apple said in its complaint that it uncovered a broader pattern of theft by alumni who went to work for OpenAI. Apple accused some of its former employees of emailing themselves confidential Apple material before leaving Apple. And it accused others of using knowledge of Apple’s trade secrets to assist in OpenAI’s hardware development.

The case could draw comparisons to a 2017 lawsuit Google filed against Uber after Anthony Levandowski, a longtime leader in Google’s self-driving car team, downloaded files before departing to set up a competing effort at Uber. The companies settled the case, though the U.S. Department of Justice prosecuted Levandowski, who eventually pleaded guilty. President Donald Trump later pardoned him at the end of his first term in office.

Apple has a long track record of protecting its intellectual property by monitoring and taking legal action against current and former employees. In 2019, Apple sued Gerard Williams, a former chief architect for Apple’s chips after he co-founded chip startup Nuvia. Apple alleged that he had breached his Apple employment contract by jumpstarting the concept for his company and recruiting employees while he was still working at Apple.

In legal filings, Williams countered Apple with allegations that Apple was intimidating any Apple employee who might consider leaving the company by monitoring its employees’ phone and text messages “in a stunning and disquieting invasion of privacy.” Williams and Apple settled in 2023.

In the OpenAI lawsuit, Apple said that in the early stages of its investigation in February it wrote to OpenAI raising concern about the potential of confidential information going to the company. Apple asked OpenAI what steps it was taking to avoid this. OpenAI never responded, according to the lawsuit.

Tan has been the primary recruiter for OpenAI’s hardware project, which got a boost after the acquisition of io Products in 2025, the lawsuit says. Since that deal, OpenAI has hired hundreds of Apple hardware employees who’ve worked on everything from product design, to displays and antennas, to supply chain management and hardware procurement.

OpenAI has yet to announce any of its hardware devices. One of the products OpenAI has talked to suppliers about making resembles a smart speaker without a display, The Information has previously reported. OpenAI has also considered building glasses, a digital voice recorder and a wearable pin, and is targeting late 2026 or early 2027 for the release of its first devices. The effort is part of OpenAI CEO Sam Altman’s longtime quest to develop a highly responsive AI akin to the virtual assistant in the Spike Jonze film “Her.”

In conversations with current Apple employees, Tan has promised some people that they will encounter a less bureaucratic and more collaborative environment at OpenAI than Apple, The Information has previously reported.

At Apple, Tan had been a key member of Apple’s hardware engineering team, including overseeing iPhone development, for around 25 years before leaving in early 2024. Apple claimed that before he left Apple, Tan met with “OpenAI or its collaborators” to discuss meetings with an important Apple supplier. He then emailed himself information about Apple’s suppliers.

Last year, The Information reported that Luxshare, a major assembler of iPhones and AirPods in China, had secured a contract to assemble at least one of OpenAI’s devices and that the AI startup had approached Goertek, which assembles AirPods, HomePods and Apple Watches, to supply components for OpenAI’s future products.

Apple’s lawsuit shows how protective it is over its supplier relationships and the techniques it develops with them. It accused OpenAI of tricking an unnamed Apple partner to perform a confidential Apple metal finishing process for OpenAI’s benefit. “OpenAI and its cohorts misled the partner to believe they had Apple’s permission to have the partner carry out the confidential metal-finishing technique for OpenAI’s benefit,” the lawsuit said.

AI once again dominated venture funding this week, claiming five of the 10 largest announced rounds, including a pair of billion-dollar financings for AI infrastructure and cybersecurity that led the pack. Investors also continued to back quantum computing, geothermal energy, crypto infrastructure and aerospace startups with large checks. Let’s take a look.

1. (tied) Keyfactor, $1B, cybersecurity: Keyfactor raised a $1 billion private equity round led by Summit Partners. Other investors in the private equity round for the Independence, Ohio-based company included Insight Partners and Sixth Street Growth. Keyfactor provides digital identity and machine identity management software that helps enterprises secure certificates, encryption keys and connected devices. It has now raised $1.21 billion to date, per Crunchbase.

1. (tied) SambaNova, $1B, AI infrastructure: Palo Alto, California-based SambaNova officially announced a long-awaited $1 billion Series F deal at an $11 billion post-money valuation led by General Atlantic. A very long list of other investors joined the round, including Battery Ventures, BlackRock, Capital Group, Intel Capital, Qatar Investment Authority, T. Rowe Price and Vista Equity Partners. SambaNova develops AI chips and enterprise AI infrastructure for training and inference workloads. The company has raised nearly $2.5 billion to date, per Crunchbase.

3. Oratomic, $300M, quantum computing: Arch Venture Partners, Khosla Ventures and Spark Capital co-led a sizable $300 million Series A for South Pasadena, California-based quantum startup Oratomic. A total of 16 investors participated in the round, including Bezos Expeditions, General Catalyst, Index Ventures, Robinhood co-founder Baiju Bhatt, and computer scientist Scott Aaronson. Oratomic is developing neutral-atom quantum hardware and fault-tolerant architectures designed to accelerate the commercialization of quantum computing, an area that has seen robust venture investment in recent years.

4. Quaise Energy, $134M, clean energy: Houston-based Quaise Energy raised a $134 million Series B led by Prelude Ventures. Additional investors included Jera Ventures, Idemitsu Americas Holdings Corp. and Safar Partners. Quaise is developing millimeter-wave drilling technology to unlock deep geothermal energy, an emerging source of carbon-free power. To date, the company has raised $225 million.

5. Prime Intellect, $130M, artificial intelligence: San Francisco-based Prime Intellect raised a $130 million Series A led by Radical Ventures. A long list of investors — many of them prominent Silicon Valley figures — joined, including Box CEO Aaron Levie, Perplexity CEO Aravind Srinivas, OpenAI co-founder John Schulman, Cloudflare CEO Matthew Prince and Ramp co-CEO Karim Atiyeh. Corporate investors Dell Technologies Capital, Intel Capital and NVentures also backed the round. Prime is building an open platform for training and deploying AI models across distributed compute networks. It has now raised $200.4 million total, per Crunchbase.

6. Gauntlet, $125M, crypto infrastructure: New York-based Gauntlet reportedly raised a $125 million Series B, with Japan’s SBI Group as the sole investor. The company develops simulation, risk management and optimization software for decentralized finance protocols.

7. Norm AI, $120M, artificial intelligence: New York-based Norm AI secured a $120 million Series C led by Khosla Ventures at a reported $1.2 billion valuation to expand its AI-powered regulatory compliance platform. The company develops AI systems that translate complex laws and regulations into software to help enterprises automate their compliance workflows. The latest funding included a long list of other venture, corporate and individual backers including Bain Capital Ventures, Fenwick & West, Craft Ventures, Coatue, New York Life Insurance, Vanguard and Hamilton James, the chairman of Costco and former president of Blackstone Group, which also participated in Norm AI’s deal. The startup has now raised just over $256 million, per Crunchbase.

8. Venus Aerospace, $91M, aerospace and defense: Aerospace continues to draw substantial investor attention, as was the case this week with Houston-based Venus Aerospace’s $91 million Series B. Mercury backed the round, which will be used to advance development of Venus’ hypersonic propulsion technology. The company is building engines and aircraft designed to dramatically reduce long-distance flight times while supporting future defense applications. It has now raised $197 million total. An array of investors joined in its Series B, including Trousdale Ventures, Prime Movers Lab, Lockheed Martin Ventures, Airbus Ventures and Draper Associates.

9. EDX Markets, $76M, fintech: Digital asset exchange EDX Markets raised $76 million as institutional interest in crypto trading infrastructure continues to grow. The deal was backed by sole investor SBI Group, marking the second large crypto funding deal for the Japanese firm this week, along with Gauntlet’s aforementioned round. EDX operates a marketplace designed specifically for institutional investors. It’s not clear how much it raised in previous rounds.

10. Fore Biotherapeutics, $67.4M, biotechnology: Philadelphia-based Fore Biotherapeutics (previously known as NovellusDx) raised $67.4 million in Series D funding to advance its precision oncology therapies targeting rare cancer mutations. The company is developing targeted treatments for patients whose tumors are driven by specific genetic alterations. SR One led the latest round, which brings its total to date to just over $274 million. Windham Venture Partners, Wellington Management, Primer Ventures, Samsung Securities and other investors also joined.

Apple sues OpenAI alleging theft of top-secret information

Lawsuit marks collapse of relationship between two of the biggest names in Silicon Valley

Apple has accused OpenAI and two of its employees of stealing top-secret information in a lawsuit filed in a California federal court on Friday, as the relationship between two of the biggest names in Silicon Valley unravels.

Apple claimed that the ChatGPT maker has used its former and current employees to steal hardware designs as the start-up prepares to launch its own AI-focused devices. It alleged that this was part of a pattern of misconduct normalised by OpenAI’s top leadership.

“Significant evidence has emerged suggesting individuals employed by OpenAI wrongfully took Apple’s secret and confidential information regarding our unreleased technologies, processes and products,” Apple said. “We will always defend our teams’ hard work and innovations, and we are taking all appropriate steps to do so.”

A spokesman for OpenAI said: “We have no interest in other companies’ trade secrets. We remain focused on building innovative technology that empowers people everywhere.”

The lawsuit marks a breakdown in the relationship between Apple and OpenAI as each encroaches on the other’s territory: Apple, with the launch of its new ChatGPT-like Siri in June, and OpenAI with its plan to take on the iPhone.

OpenAI signalled its plan to compete with Apple in the consumer device space when it acquired io, the studio led by Apple’s former design chief Jony Ive, for $6.4bn in May last year.

OpenAI has since aggressively poached employees from Apple, part of a broader trend of AI talent leaving the company. Apple claims that over 400 of its former employees now work there.

While OpenAI in 2024 became Apple’s first big partner in the AI space, integrating ChatGPT into Siri, Apple has since partnered with Google for its newest features.

Apple said in its lawsuit it launched an internal investigation into OpenAI’s potential theft of trade secrets in February and warned the ChatGPT maker but received no response.

Apple alleged that the evidence detailed in the complaint was just the “tip of the iceberg” at OpenAI, “where such misconduct is normalised and exemplified by leadership”.

“As a natural result, OpenAI’s nascent hardware business now rests on the shakiest of foundations, rotten to its core by its illegal reliance on misappropriated trade secrets,” the lawsuit said.

Tang Tan, OpenAI’s chief hardware officer, left Apple in early 2024 to join Ive’s outfit before it was acquired by OpenAI. Tan had a 24-year career at Apple which included a spell as vice-president of product design for the iPhone and Apple Watch.

The lawsuit alleged that Tan shared with OpenAI confidential information about Apple’s suppliers in the months before he left the iPhone maker. Now, “when interviewing Apple employees for jobs at OpenAI, Mr Tan uses Apple’s confidential information to gain access to even more insider knowledge”, it said.

Apple claimed that Tan and OpenAI had instructed Apple employees to bring digital designs and prototypes from the highly secretive tech company to interviews at the AI start-up for “show and tell sessions”.

Chang Liu, a former electrical engineer at Apple, worked on some of the company’s “most sensitive product development programs” before he joined OpenAI in January of this year, according to the lawsuit.  

Apple alleged that Liu failed to return at least one work device and did not respond to queries about whether he had complied with other “exit procedures”.

The lawsuit also claimed that Liu used an Apple employee’s computer to access trade secrets and coached her on how to copy files while evading Apple’s security before she departed for OpenAI in April.

Liu further exploited a “vulnerability” in Apple’s network storage to steal confidential files after he left the company, the lawsuit said. Apple said Liu texted the employee, “LOL, I found out I can access the [network storage], so funny.” 

Other former Apple employees have similarly taken confidential information with them “on their way out the door”, according to the lawsuit.

Apple is seeking a legal injunction that will prevent OpenAI, Tan and Liu from destroying evidence related to its claims and “return all copies of Apple’s trade secrets and confidential information”. It is also asking for damages or royalties from any misappropriated trade secrets.

The start-up, which is preparing for a potential public listing next year, already faced a legal dispute this year with Elon Musk, who sued the company and founder Sam Altman for transforming it into a for-profit company.

The case, thrown out by a jury in May, proved a costly and embarrassing distraction for OpenAI, which has pushed beyond developing AI models to building its own chips, AI infrastructure and consumer tech products.

The FT has reported that OpenAI investors have increasingly scrutinised its $852bn valuation as it has been leapfrogged by Anthropic. Apple said in its lawsuit that the alleged attempts to steal its trade secrets stemmed in part from OpenAI’s need to find new lines of revenue quickly.

Jeff Bezos-backed fusion start-up to become first to go public

General Fusion’s ‘steampunk’ approach to the technology is under scrutiny ahead of Nasdaq debut

General Fusion is set to become the first publicly listed fusion company, even as experts question whether recent scientific results from its “steampunk” technology constitute meaningful progress towards commercialisation. 

Chief executive Greg Twinney told the FT that in emerging industries first movers into the public market tended to “dominate the narrative” around what investors should expect from a commercially focused fusion company.

He added that General Fusion would have access to a “much bigger pool” of investors for future fundraising due to the lack of publicly listed rivals.

“There’s only one, that’s us.”

The Canadian company, backed by Amazon founder Jeff Bezos, who has participated in several funding rounds over the past 15 years, completed a Spac merger on Friday with Spring Valley Acquisition Corp III, valuing it at an enterprise value of $724mn and raising as much as $338mn.

Trading on the Nasdaq is due to begin on Monday.

Fusion aims to recreate the reaction that powers the Sun by fusing atomic nuclei together in superheated plasma, unlike nuclear fission, which generates energy by splitting atoms.

It has long been considered a “holy grail” of the energy industry, providing a source of low-carbon, almost limitless power, and more than 50 start-ups have been scrambling for funding.

Private fusion companies have yet to show that their fusion systems can produce more energy than is required to sustain the plasma, with the long-running joke being that fusion is “20 years away, and always will be”. But scientific breakthroughs by a US laboratory and investments by tech giants in recent years have boosted hopes among believers.

General Fusion is undergoing tests for its Lawson Machine 26 prototype in Vancouver to prove its economic viability and aims to have an operational commercial fusion plant in the mid-2030s.

The company’s listing comes as fusion companies race to secure funding for increasingly expensive machines. The number of private fusion companies has more than doubled over the past five years, intensifying competition.

A recent scientific paper published by General Fusion has raised fresh questions over whether its unconventional design is making sufficient progress towards commercial fusion, though the company says it remains confident it is on track.

Dan Brunner, at consultancy Future Tech Partners and former chief technology officer at Commonwealth Fusion Systems, said General Fusion’s recent results were “so far behind where they actually need to go” to build a commercially viable machine, adding that the company’s commercial timeline was “very hard to believe”.

Brunner said the paper showed that the company’s plasma compression failed to raise ion temperatures sufficiently, suggesting energy was still being lost through heat leakage — one of the central challenges in achieving commercial fusion power.

Most fusion start-ups are concentrating on so-called tokamak designs, which trap plasma continuously with powerful superconducting magnets. General Fusion’s process compresses magnetised plasma using a rapidly collapsing liquid-metal cavity driven by mechanical pistons — an approach Brunner described as “steampunk”.

He added that the approach posed greater uncertainty than the more established tokamaks.

Tony Donné, chair of General Fusion’s technology advisory committee and former chief executive of the European research consortium EUROfusion, disputed that assessment.

He argued that the low ion temperatures were not inherent to the design of the prototype, but instead reflected what he believed was the premature publication of the paper, as the company faced “pressure” to disclose progress ahead of its public listing.

Donné said the temperature issue had already been “investigated” and that he expected the results to improve in future tests.

This is not the first setback for one of the oldest private fusion companies. General Fusion announced in 2025 that it would lay off a quarter of its workforce after running short of funding, although Twinney said most of those employees had since been rehired following fresh investment. Multiple investors, executives and market observers also questioned whether the company chose to go public because private fundraising had become increasingly difficult.

Twinney argued that General Fusion had deliberately avoided “raising billions of dollars to build large science machines” like some rivals, instead pursuing what he described as a more capital-efficient path to commercialising fusion. “Scientific milestones are just steps along the way. They are not the end goal.”

Fatih Birol says EU should have moved faster to achieve energy independence after 2022 gas crisis

Europe has made a “major mistake” by failing to wean its economy off imported fossil fuels quickly enough since the 2022 energy crunch, the head of the International Energy Agency has warned, as the EU prepares to roll out measures to increase electrification next week.

Fatih Birol said that Europe’s low electrification rate — electricity’s share of the energy consumed in the EU — of around 23 per cent is holding back the bloc’s competitiveness and economic “sovereignty”. The EU’s rate is similar to major oil producers such as the US despite its heavy reliance on imports of hydrocarbons.

“This is in my view a major mistake for Europe,” Birol told the FT. “In general, I would have hoped and expected that Europe would have been more responsive to this crisis.”

In a joint interview with Europe’s energy commissioner Dan Jørgensen, both leaders said that Europe needed to electrify its economy faster after facing two energy crises in less than five years. Europe should seek to emulate countries such as China, Japan and South Korea, which have an electrification rate of more than 30 per cent, Birol said.

The EU has pledged to raise this rate to 32 per cent by 2030 but the comments come as Jørgensen prepares to lay out a longer-term target to boost electrification by 2040. “Our electrification rate has stagnated for the last decade . . . We need to electrify and we need to do it much faster,” he said.

The Danish commissioner argued that the EU had deployed more renewables, increased energy efficiency and lowered gas consumption by 20 per cent in 2022, when Russia dramatically reduced pipeline supplies of gas — but the bloc’s heating, transport and industrial sectors remain reliant on fossil fuels.

Jørgensen acknowledged this continued reliance on imports meant the region had still been hit hard by the global disruption to oil and gas supplies caused by the Middle East conflict.

The Commission will lay out plans next week to require countries to lower taxes on electricity and offer support to encourage households to adopt heat pumps, electric cars and other green technologies.

A draft document of the proposal, seen by the FT, underlines that there are only two member states where electricity is less than twice as expensive as gas for industries: Sweden and Finland.

The plan will seek to boost electrification by introducing incentives so that electricity costs are no more than 2.5 times gas prices for households and 2 times for industry by 2030. This would be achieved partly by mandating that electricity is taxed less heavily than fossil fuels.

This should encourage individuals to install heat pumps and buy electric vehicles, while also providing industries with a stronger economic case to decarbonise, the document states.

But the measures could prove expensive for countries that are heavily reliant on taxation stemming from electricity bills. Greece, Italy, Hungary and Ireland have some of the highest electricity to gas price ratios.

Levies to fund grids and other charges are often added to electricity bills.

Birol also warned that ongoing issues with grid capacity were holding back the EU’s electrification push. Hailing a “record” year of 85GW of renewable installation last year as “very good news”, he said that 600 gigawatts of renewables were already finished and in the queue for a grid connection.

Much of the grid congestion in Europe stems from problems at a national or regional level, as countries struggle to adapt to lots of renewable power projects, sometimes in remote locations, generating electricity rather than a smaller number of very large fossil fuel plants that are often closer to industrial centres and households.

“Member states already tomorrow can speed up the process of expanding their grids and using the ones that we have more efficiently,” Jørgensen added.