TheChek News Flow

Eric Schmidt warns west to focus on open-source AI in competition with China

Former Google chief’s caution comes following Chinese start-up DeepSeek’s breakthrough

Former Google chief Eric Schmidt has warned that western countries need to focus on building open-source artificial intelligence models or risk losing out to China in the global race to develop the cutting-edge technology.

The warning comes after Chinese start-up DeepSeek shocked the world last month with the launch of R1, its powerful-reasoning open large language model, which was built in a more efficient way than its US rivals such as OpenAI.

Schmidt, who has become a significant tech investor and philanthropist, said the majority of the top US LLMs are closed — meaning not freely accessible to all — which includes Google’s Gemini, Anthropic’s Claude and OpenAI’s GPT-4, with the exception being Meta’s Llama.

“If we don’t do something about that, China will ultimately become the open-source leader and the rest of the world will become closed-source,” Schmidt told the Financial Times.

The billionaire said a failure to invest in open-source technologies would prevent scientific discovery from happening in western universities, which might not be able to afford costly closed models.

Schmidt was speaking at the AI Action Summit in Paris this week, which also saw US vice-president JD Vance vow that the US would remain the dominant force in the technology.

OpenAI chief Sam Altman said last month he had been on the “wrong side of history” when it came to open-source models, suggesting the company needed to develop a new strategy that moved away from expensive closed models.

However, the group has taken early strides in developing the technology, leading to the start-up holding talks with SoftBank over a huge funding round at a $260bn pre-money valuation. Other US Big Tech giants, including Google and Amazon, have also invested billions on closed models, believing it will be the best way to secure returns on significant investment.

“I think that [Altman] was being a little too quick,” said Schmidt, adding that the future of AI dominance lay in combining both open- and closed- source AI models.

While the US had dominated the first phase of AI development by building powerful AI models such as GPT-4, Europe had a chance to reap economic gains by thinking about how to apply AI to businesses through building “applications” on top of the technology, said Schmidt.

“Europe’s got to get its act together,” he said. “The application layer is profoundly powerful, and it will make your Europe more efficient.”

On Wednesday, Schmidt announced he was investing $10mn to create a new AI Safety Science programme through Schmidt Sciences, a non-profit he set up with his wife, Wendy.

The programme will include 27 projects working on fundamental research into safety problems in AI. The awardees include Yoshua Bengio, a Turing prizewinning computer scientist, who will develop a risk mitigation technology for AI systems, and Zico Kolter, an OpenAI board member and professor at Carnegie Mellon University, who will explore AI attacks.

Schmidt called for the west to collaborate with the Chinese on AI safety, as the country would face the same issues around the powerful technology. He compared this to the kind of information sharing done between militaries when they test rockets, for example.

“How could it possibly be bad for us to give them information that they could use to make their models more safe?” he said.

Will a private tech company reach a $1 trillion valuation in the coming years?

It’s not a ridiculous question. A couple of companies seem like potential candidates. OpenAI is closing in on $300 billion in its financing with SoftBank, and SpaceX recently shot to $350 billion.

Other private tech heavyweights are reaching impressive fresh new valuations, without going public. My colleague Sri Muppidi broke the news Monday that Stripe is talking to investors about an employee share sale that would value the company at $85 billion, more than Capital One or Barclays. Waymo’s $45 billion valuation last year put it above Ford and Honda.

What’s more, there are few rumblings that any of these companies are going to go public anytime soon. Top-tier private companies have plenty of access to capital. Even Goldman Sachs’s CEO says delaying an IPO seems logical for big private companies. It’s becoming more likely that name-brand Silicon Valley firms might end up like other privately held, globally important conglomerates like Koch Industries, Bloomberg or Hearst.

A handful of tech companies are expected to go public in the first half of this year. Most will have strong businesses, but they are not seen as industry defining. Investors aren’t as excited about Chime or Klarna as they would be about Stripe. Data center operator CoreWeave is sure to get plenty of hype, but it isn’t OpenAI or Anthropic.

Another question I’ve been mulling: Is this a problem?

One obvious downside is that everyday stock market investors can’t invest in the most successful recent crop of startups without resorting to sketchier investment vehicles.

Another issue is more esoteric—the potential costs to society when some of its most valuable companies keep a tight lid on their financials, governance structures and risks. Reporting on these companies, as well as disclosures to investors during private fundraising events, only go so far. Companies have to open up a lot more as they prepare to go public.

Jeff Nykun, a former senior analyst at the asset manager Artisan Partners, put it to me this way: “If the world stays the way it is, wouldn’t it be weird if we didn’t know what’s going on with the biggest companies in the world? I don’t know if that’s good for anybody.”

Nykun is one of the many tech investors optimistic that the IPO drought has to end within the next year or two. It’s crucial to the venture capital industry that the dam breaks. He is a general partner at tech investment firm WndrCo, which has invested in private software companies like Databricks, Deel and Figma—hot companies that IPO watchers expect to begin marching to public listings by next year.

He thought what might finally push top startups over the edge comes down to brass tacks—that they will be able to sell stock for a higher price in the public markets than to private investors. Private software companies that are growing faster than 30% a year, a quicker pace than nearly all their publicly traded peers, are in high demand, he added. “I think people will be shocked how well they trade,” he said.

That could drive another classic tech-industry phenomenon: fear of missing out.

Right now, tech founders appear to be taking cues from the same elite crew that have proven allergic to going public, like Elon Musk at SpaceX or Patrick Collison at Stripe. Cool kids just don’t IPO. Their companies are too precious for public hands.

Another crop of celebrity founders, at least by Silicon Valley standards, might make IPOs trendy again. I’d bet Dylan Field of Figma or Ali Ghodsi of Databricks will ring the IPO bell at Nasdaq or NYSE by sometime next year. They might be in for a wild ride but their employees, investors and customers will thank them for their new transparency.

Would Musk Be Better Than Altman at Running OpenAI?

Conventional wisdom is that a Musk-led group’s bid for OpenAI on Monday night isn’t a serious offer but more of a way to complicate Altman’s effort to restructure the artificial intelligence firm. For the sake of argument, let’s assume it is real. Then the issue becomes one of price and, potentially, who is the best person to generate value from OpenAI. Given Musk’s accomplishments as an entrepreneur (you know the list: Tesla, SpaceX, xAI and so on), you would be hard pressed to discard him in favor of Altman.

That might be one reason why Altman is rejecting the bid by claiming, as we reported on Monday night, that it wasn’t in the “interest of OpenAI’s mission to develop advanced artificial intelligence for all of humanity.” (Musk has made a similarly pious argument in his lawsuit against Altman and OpenAI.) But this battle is essentially an extension of a longstanding argument between the two OpenAI co-founders, who split several years ago, well before the company became a juggernaut.

There’s plenty of precedent for disgruntled co-founders suing their former colleagues once their business becomes rich and famous (Reggie Brown at Snap and Eduardo Saverin at Facebook are just two examples). One difference in this situation is that Musk isn’t an also-ran—he’s even more successful as an entrepreneur than Altman is.

So when Altman pushes back, asserting that OpenAI “is not for sale,” investors in the company should wonder. Might they be better off with Musk? Maybe or maybe not. After all, OpenAI has had a revolving door as AI researchers have left for greener pastures, often firms started by OpenAI alum such as Mira Murati. Would Musk make the talent situation worse—or could he fix it?

Could Trump resurrect the idea for the US to buy Nokia or Ericsson?

The rationale behind the proposal in his first term has not gone away

Returning US President Donald Trump has already revived some of the most striking ideas of his first term. An American takeover of Greenland and radical action in the Middle East are once more the talk of the world.

But what eye-catching proposals has Trump yet to resurrect from his first stint in office? There is one in particular that strikes me as intriguing now as it was then: that the US should buy Nokia or Ericsson, or even both.

William Barr, attorney-general under Trump, suggested in 2020 that the US should actively consider taking a “controlling stake” in either or both of the Finnish and Swedish telecoms equipment makers “either directly or through a consortium of private American and allied companies”.

Like many Trump proposals, the idea was first met by gasps of disbelief. The US government doesn’t tend to buy foreign companies. But as is the case in some of the president’s outlandish schemes, there was a kind of rationale to the proposed purchase — and one that has not gone away in the meantime.

Telecoms equipment manufacturing is one of the very few areas of technology where the US is not just behind but not present at all. Reliable networks are vital for business and consumers alike, as well as becoming increasingly essential in warfare, as Ukraine is demonstrating with its drone warfare.

“They have not solved that issue in the US,” says Anna Wieslander, Northern Europe director at US think-tank the Atlantic Council.

Nokia and Ericsson have an effective duopoly in much of the western world thanks to American pressure on allies not to use Huawei, their main rival, which has close ties to the Chinese state. But they have struggled to draw as much benefit out of that as many might have expected with both experiencing disappointing profitability in recent years.

What is more, Ericsson and Nokia have failed to garner full-blooded support from the EU, all the more strange for being perhaps the one sector where Europe has technology dominance.

It was perhaps no surprise to see the bitter rivals take the unusual step of hosting a joint lobbying event last month in Brussels at which Ericsson and Nokia chief executives warned the EU that they needed to act and quickly if the tech gap to the US would not increase further.

Beneath the surface, there are signs of something of a flirt from Ericsson and Nokia to the US. Börje Ekholm, Ericsson’s Swedish chief executive, lives in the US and has mused out loud about the possibility of moving its headquarters there too, something its main shareholder — the Wallenberg family of industrialists — seems less keen on. Ericsson also went public on its donation to Trump’s recent inauguration.

Pekka Lundmark, Nokia’s departing chief executive, has repeatedly called the US the company’s “second home” even as it lost big American contracts with groups such as Verizon and AT&T in recent years. His replacement at the top of the Finnish company from April will be a US citizen, Intel executive Justin Hotard, with a remit to boost its stagnating sales and share price.

Asked this week about the idea of Trump reviving the idea, Lundmark told the Financial Times: “The US is obviously our most important market, and it is very important that we build good relations with the new administration . . . Nokia is a listed company, and our stock is available on the market.”

He added that Ekholm had used the Brussels event to send a sharp warning to EU policymakers: “Unless Europe finds a way to be a more competitive home for technology companies, there’s a big risk that companies like Ericsson and Nokia could move somewhere else.”

People close to both groups say the problem for each one would be if the other became American. “It would be hugely troublesome. Then they would be a domestic vendor in the US. It would be a headache. We definitely have interest in keeping them European,” says one senior executive of his rival on the other side of the Gulf of Bothnia.

The two companies have different ownership structures. Ericsson has the Wallenbergs and another Swedish industrial holding company as big shareholders, but it also has Cevian Capital, Europe’s largest activist investor, which in 2020 pushed management to consider a deal with Trump. Nokia has no anchor investor — its largest shareholder is the Finnish state investment company Solidum with a 5.8 per cent stake.

There is no guarantee that Trump will revive this idea. But offering again to buy Nokia or Ericsson could prove to be one of Trump’s most beguiling proposals.

Big funding rounds for software developer Helsing and drone maker Tekever defy broader downturn in venture investing

Investment in European start-ups working on defence and related technologies jumped 24 per cent in 2024 to $5.2bn, outpacing growth in venture capital for artificial intelligence in the continent over the past two years.

The figures published on Wednesday by the Nato Innovation Fund and research group Dealroom found that investor appetite for companies such as defence software developer Helsing and drone maker Tekever defied the broader downturn in European VC funding last year.

Kelly Chen, partner at the Nato fund, which has raised €1bn to back defence and “deep tech” start-ups, said: “The VC market is cyclical and over the past year or two, it declined 45 per cent. But if you look at what [sectors] still increased, defence, security and resilience had by far the strongest growth.”

The figures were released ahead of the Munich Security Conference, which begins on Friday, where US vice-president JD Vance is expected to meet Ukraine’s President Volodymyr Zelenskyy.

The Russia-Ukraine war has contributed to a surge of interest from investors in European defence start-ups, which grew 41 per cent over the prior year to $1.5bn in 2024, according to Dealroom.

While that helped narrow the gap with US funding in the sector, American defence tech investment still far outpaces Europe’s, hitting $4.3bn last year, driven by companies such as California-based Anduril.

The Nato fund’s definition of “defence, security and resilience” companies encompasses a range of tech and energy start-ups, including robotics, biotech, quantum computing, earth observation systems and nuclear power, as well as “dual use” AI systems that could be used in military or commercial contexts. Together, those fields now account for 10 per cent of total venture capital investment in Europe, according to the report.

Germany overtook the UK to attract the most VC funding in the space, thanks to a growing cluster in Munich, which accounted for almost $1bn in funding last year.

Helsing, which is developing AI for use on the battlefield, raised €450mn last July in a deal led by US fund General Catalyst valuing the Munich-based start-up at about €5bn.

Despite frenzied US investment in start-ups building AI models and applications, capital flowing into Europe’s AI sector was broadly flat between 2022, when OpenAI’s ChatGPT debuted, and 2024, according to Dealroom.

By contrast, VC investment in defence, security and resilience grew 30 per cent over the same period.

The Nato Innovation Fund was launched in 2022 to help fill a gap in funding for companies working on military and commercial capabilities amid concerns that European start-ups lacked the financial muscle of their US peers. Venture investors, in particular in Europe, had long been wary of backing defence tech companies over ethical concerns but that has begun to change since Russia’s full-scale invasion of Ukraine in February 2022.

Lengthy government procurement cycles in defence are also seen as a challenge for smaller companies. Chen said the fund was trying to “open up procurement channels”.

This is “the chicken and the egg problem”, added Chen. Faster procurement was needed in order for the burgeoning sector to attract more private capital.

“For this industry to grow we need more private capital. It won’t just be government grants and public money,” she added.

Farallon presses Astellas to cut costs and overhaul drugs research in latest shareholder campaign gripping Tokyo

Farallon Capital Management has taken a significant stake in one of Japan’s biggest pharmaceutical companies as activist shareholders, emboldened by a corporate governance reform drive, take aim at the country’s household names.

The San Francisco-based fund with $39bn in assets has built a more than 3 per cent position in Astellas, according to people familiar with the matter, making Farallon a top-three shareholder, based on LSEG data.

The move is part of a wave of shareholder activism in Japan as governance reforms by the Tokyo exchange push companies to take investor demands more seriously.

Astellas shares, which were down 2 per cent in early afternoon trading on Wednesday, gained sharply on the news to trade up 3.5 per cent.

Farallon made its name in the country with a series of high-profile bets, including a long and often contentious engagement with Toshiba, where it successfully installed an outside director shortly before the cash-strapped conglomerate was taken private.

Its involvement in the management change, alongside fellow activist Elliott Management, redefined the role of shareholders in Japanese corporate crises.

Since exiting its roughly $800mn position in Toshiba in 2023, Farallon has left the market speculating about its next target.

Farallon, which calls itself an engagement fund, is hoping to work with the Astellas management and board to quickly and significantly increase its share price, said people familiar with the matter. The company has a market capitalisation of ¥2.6tn ($17bn) after its share price fell almost 40 per cent from a peak in 2023.

The fund has been engaging with Astellas since 2020 and sent its board numerous letters, which have been seen by the Financial Times.

Farallon wants Astellas to cut costs more quickly and overhaul its research and development programmes, which it says are diffuse and have failed to advance a sufficient number of drugs past trial stages. It also wants Astellas to refocus its merger and acquisition strategy on later-stage drugs with a higher chance of getting to market.

Astellas had spent more than ¥1.5tn on M&As since the beginning of 2016, according to Farallon’s letters, and the company said last year that under new chief executive Naoki Okamura, it was the “right time to go on the aggressive to further accelerate growth”.

One of its most successful drugs, Xtandi, which treats prostate cancer, is expected to reach the end of its patent exclusivity period in 2027, giving the company a limited window to invest cash flow from the drug’s sales.

Astellas during its third-quarter results last week said it was making progress in cost-cutting and drug development. It has targeted a 30 per cent core operating margin, compared with its current level of close to 20 per cent.

Farallon declined to comment. Astellas said: “We appreciate and value the discussions we have with individual shareholders; however, we do not comment on matters regarding individual shareholders.”

With Farallon increasing its engagement with the company, Astellas has asked a group of external directors to provide “objective oversight of the execution” of its priorities, which overlap with the activist fund’s suggestions, said people familiar with the matter.

Farallon has indicated to the board that it expects the external directors to conduct a broad review of the company’s cost structure, R&D and M&A plans, as well as other strategic possibilities, including partnerships or an outright sale, the people added.

Some analysts have compared the group to a special committee, which is often tasked with analysing takeover offers but can also advise on strategy.

Since Japan overhauled M&A guidelines in 2019, activist shareholders have used the new rules to show that investors can start a sale process, most recently with IT services company Fuji Soft, in which Farallon was also an investor.

Japanese pharmaceutical companies have taken different approaches to finding the scale and resources to compete against global rivals, including M&As and partnerships with bigger groups, such as Chugai Pharmaceutical’s strategic alliance with Switzerland’s Roche.

Last week, Bain Capital acquired 340-year-old Mitsubishi Tanabe Pharma for $3.3bn, betting that drug development in Japan would benefit from regulators loosening their approval process.

EU weighs temporary gas price cap to counter diverging costs with US

Proposal draws backlash from industry, warning it could damage ‘trust’

Brussels is weighing new powers to temporarily cap EU gas prices, which have recently hit record levels compared with the US.

European natural gas prices traded at the highest in more than two years this week, in part because of low temperatures and a lack of wind that has hampered renewable energy production. They are between three and four times higher than in the US, providing a critical handicap to European companies.

The European Commission is considering a cap as part of discussions about a “clean industrial deal” policy document to be presented next month, said three people with knowledge of the talks.

The strategy paper should outline ways to shore up the EU’s heavy industries as businesses grapple with multiple challenges including US President Donald Trump’s aggressive trade measures and the EU’s own ambitious green transition.

Talks around mechanisms to cap prices, though still at an early stage, have drawn a backlash from industry groups which warn against damaging “trust” in the European market.

Eleven groups including Europex, the association of European energy exchanges, and AFME, the financial markets lobby group, sent a letter to commission president Ursula von der Leyen on Tuesday and seen by the Financial Times. It said: “We believe this measure, if announced, could have far-reaching negative consequences for the stability of European energy markets and the security of supply across the continent.”

A gas price cap would “harm the trust” in Europe’s benchmark Title Transfer Facility, the main centre for trading and settling the price of gas, the letter said. It would also “prompt the global gas community to shift towards other, unrestrained and therefore more representative reference prices, which are primarily located outside of the EU”.

The EU first proposed a similar cap in 2022, at the height of the bloc’s energy crisis that followed Russia’s steady squeeze on gas supplies to its European neighbours following the full-scale invasion of Ukraine. The cap was never enacted, as prices remained below the €180 per megawatt-hour benchmark.

Building on that experience, former European Central Bank president Mario Draghi last year called for the commission to have powers to bring in “dynamic caps” for situations when the EU gas price diverges from global energy prices.

“We are studying in detail Draghi’s recommendations on this specific issue,” an EU official said.

Two senior bloc officials said the plans would also include measures to prevent traders pushing up prices of gas in the summer as European countries stock up on the fuel ahead of the next winter.

One EU diplomat said some member states were likely to be “reluctant” about having a price cap in place. Germany and the Netherlands were among the countries that opposed the previous cap.

The commission declined to comment.

Any intervention could also hamper the bloc’s efforts to stave off Trump’s tariff offensive unless the EU buys more liquefied natural gas from the US.

“The emphasis in Europe should be on acquiring enough energy to run their industries and heat their houses,” said Amund Vik, senior adviser at Eurasia Group and former Norwegian state secretary for energy.

“Fixing a wholesale market price cap is not going to solve that when the underlying problem is the lack of energy.”