TheChek News Flow

Airline said SpaceX’s satellite-internet service generated interference after being installed on a couple dozen regional jets

Key Points

United Airlines started rolling out free Starlink Wi-Fi last month with fanfare. But it has had to disable the service on roughly two dozen planes to address problems with static interference.

United said Friday that it is working with Starlink to address “a small number of reports” about the problem, which it said isn’t a flight-safety issue. While it is being fixed, the Starlink-equipped regional jets have been operating with the Wi-Fi turned off.

“We expect the service to be back up and running on these aircraft soon,” United said. The fix will be deployed during routine maintenance and the airline doesn’t anticipate any impact to flight schedules.

SpaceX didn’t respond to a request for comment. The issue was previously reported by The Points Guy.

Such problems are fairly common with any new in-flight internet provider, United said. Similar static interference issues have cropped up with other new Wi-Fi technology because of the number of antennas on planes.

United plans to outfit its entire fleet with Starlink, a division of Elon Musk’s SpaceX that provides internet connections using a fleet of thousands of satellites located relatively close to earth. United has touted what it said is Starlink’s faster, more reliable service, aiming to allow customers to use devices just as they do on the ground.

Starlink has driven revenue growth at SpaceX and the company has been pushing into the airline industry, striking agreements with Hawaiian Airlines and Air France. SpaceX surprised many satellite-industry executives by winning the fleetwide Starlink deal with United last year.

The problem emerged after Musk this week engaged in a war of words with President Trump, creating new risks for his companies.

United had previously accelerated its timeline to roll out Starlink. It said it remains on track to have Starlink available on all of its two-cabin regional jets by the end of the year.

If all goes as planned in August, a startup called Starcloud from Redmond, Wash., will load a rocket bound for space with a refrigerator-sized satellite, the first to contain an Nvidia H100 chip. Normally, that chip is at the heart of the sprawling data centers back on Earth that run the latest artificial intelligence models.

Compared to existing data centers, the Starcloud satellite will be almost comically under-powered: It may have only enough computing muscle to run less demanding versions of Google’s Gemini and OpenAI’s GPT models, according to Starcloud CEO Philip Johnston. Still, the satellite will be the most powerful computer in orbit, with 100 times more processing power than the International Space Station and other satellites, according to Johnston, whose company raised $21 million out of Y Combinator in 2024, one of the year’s largest rounds from the incubator.

And, ultimately, Starcloud wants to build the first gigawatt-scale data center in space, with bays full of AI chips powered by a massive solar array spanning 4 kilometers by 4 kilometers.

Hot on Starcloud’s tail fins is another space company, Axiom Space, which plans to launch two orbital data center nodes with a mix of CPU and GPU chips capable of “running some low grade AI” by the end of the year as part of a network of communications satellites for military and commercial clients, according to Jason Aspiotis, Axiom’s global director of in-space data and security. And it’s not just startups that want to put cloud computers above the clouds—some of tech’s most influential billionaires buy into the idea too.

In April, when a journalist on X suggested that former Google CEO Eric Schmidt bought a controlling interest in a rocket company, Relativity Space, with the goal of putting data centers into orbit, Schmidt replied with one word: “Yes.”

Meanwhile, Jeff Bezos has long said that he started his rocket company, Blue Origin, to make it possible to move pollution-belching industries off Earth. According to a person who has discussed the subject with him, data centers are among those industries that he believes belong in space long term. That’s telling coming from the guy who did more than anyone to kick off a frenzy of data center building when he ran Amazon, which operates the No. 1 cloud computing provider, Amazon Web Services.

Still, to say that the concept of orbital data centers has its doubters would be a bit of an understatement. There’s a head-spinning array of technical and financial impediments to making them work—from guarding their computers against radiation, to maintaining them, to affording the expense of getting all that hardware into orbit in the first place. Sure, maybe a company can stick a few computers running lightweight AI models in space, but launching and operating AWS-sized data centers in orbit is science fiction, skeptics say.

“Show me the calculations that suggest this is even feasible,” said Oren Etzioni, the former CEO of the Allen Institute for AI and a professor emeritus of computer science at the University of Washington.

It isn’t likely there’d be much talk of sticking data centers in orbit if there wasn’t a borderline freakout back on Earth about how to power them all. In the U.S., data centers are forecast to consume 12% of total U.S. electricity by 2028, up from 4.4% in 2023, according to a U.S. Department of Energy report last year.

Electricity-hungry AI models are expected to be the biggest factor in that surge. Already, power companies are struggling to keep up with the demand, and electricity bills are rising for consumers. Big data center operators like Amazon, Microsoft, Google and Meta Platforms, meanwhile, are plowing money into nuclear power, geothermal and other sources of clean energy to meet their future needs—though it’s too soon to say whether and when those investments will pay off.

Orbital data centers have a tantalizing answer to the power problem: uninterrupted access to solar energy, without the hindrances of weather, night time and seasons. That means they could enjoy dramatically lower operating costs while also combatting climate change by reducing the reliance on fossil fuels, which currently account for 60% of total U.S. energy consumption. That’s the theory, at least.

This isn’t the first seemingly wacky plan to put data centers in an exotic location for environmental and cost reasons. In 2018, Microsoft researchers dropped a data center encased in a steel tube to the seafloor near Scotland’s Orkney Islands, which allowed them to tap into the cooling effect of the ocean’s chilly subsurface waters. After two years of testing, Microsoft plucked the data center from the water and called the experiment a success, but it hasn’t put another one back in down there.

Other countries are also eyeing orbital data centers, which could prompt competition with the U.S. to build them first. Last month, China launched a dozen satellites that will eventually be a part of an orbital computing constellation of 2,800 satellites that can run AI models and perform other computing chores.

Orbital data centers are starting to pick up buzz in Washington. In April, during the Hill & Valley Forum, a gathering of tech and political luminaries in the city, Republican Senator Mike Rounds of South Dakota told an audience that U.S. control of AI won’t happen unless it can solve the energy problem it’s creating—and that it has to consider the option of putting data centers in space.

“I know it sounds crazy,” said Rounds, who serves on the Senate Armed Services Committee. “We’ve got to be talking unique and innovative ways as we move forward into the future.”

Jack Clark, co-founder of the AI startup Anthropic, who was sitting next to Rounds at the event, responded: “You don’t sound crazy at all. We’re big fans of putting the computers in space.”

On Earth, new data centers often meet with fierce opposition from homeowners, preservationists and other groups. In space, there’s a different kind of hostility to them.

Like anything in orbit, data centers will be constantly bombarded by radiation from solar flares, cosmic rays and Earth’s magnetic field. Experts say that radiation could disrupt the performance of the densely packed rows of AI chips inside orbital data centers. This will require backup chips, ample shielding and other robust materials, all of which will add weight to those systems. Every additional ounce will make for a more expensive launch.

That’s not all. There are collisions with satellites and space debris to worry about, requiring tracking and coordination to dodge those objects. And then there’s the challenge of how to keep all those computer chips from overheating, a task that terrestrial data centers normally handle by circulating large quantities of fresh water around their equipment.

While the frigid temperatures of space provide an obvious way to cool those machines, orbital data centers will still require radiators to dissipate waste heat—very, very big radiators.

In one analysis of space-based computing last month, former NASA administrator Dan Goldin estimated that a 100-kilowatt data center would need a radiator larger than a tennis court. (Typical large data centers consume at least 100 megawatts, or 1,000 times more). A one megawatt data center would need something more like a soccer field.

Starcloud, for one, believes it can figure out these issues, even though no one has ever shown it can be done at scale before. In a mock-up video, Starcloud shows a hypothetical 5-gigawatt orbital data center with a stem-and-leaf design featuring a vast array of solar panels. A radiator on the backside of the solar panels will dissipate heat.

In its design, the data center will consist of dozens of containers filled with networking switches and server racks. After being released from the cargo bay of a rocket, the containers will autonomously dock themselves to a stem that runs between the two sections of Starcloud’s solar array. The data center will use lasers to communicate with existing satellite internet constellations—whether SpaceX’s Starlink or eventually Amazon’s Kuiper—that will transmit data to and from Earth.

In the near term, Starcloud says its data centers will be much smaller than the one shown in its movie—and the cost will compare very favorably to terrestrial equivalents. Johnston estimated it will eventually cost the company about $20 million to build and launch a 40-megawatt data center into orbit—still relatively small by terrestrial standards—compared to hundreds of millions of dollars to build a similar-sized facility on earth and purchase the land beneath it.

Operational costs are where Starcloud’s data center will shine, according to the company. In a Starcloud white paper, it estimates that a 40-megawatt data center on Earth will cost $140 million to power over a decade while one in outer space with a solar array would cost about $2 million.

There are a lot of assumptions built into Starcloud’s cost estimates, though, most notably one envisioning a decrease in the price of rocket launches. Currently, SpaceX charges customers around $70 million to book all of the capacity on a Falcon 9 rocket, but that rocket is too small to put Starcloud’s future 100-ton, 40-megawatt data center into orbit.

It may not be long before those launch prices come down and the capacity to send bigger, heavier objects into space goes way up. SpaceX has done multiple test launches of a gargantuan new heavy-lift rocket called Starship which will eventually allow it to bring bigger payloads to space.

While SpaceX dominates the commercial launch market, Bezos’ Blue Origin in January took an important step in introducing a significant new competitor to the market, when it successfully tested its Starship rival called New Glenn. Rob Meyerson, the former president of Blue Origin, believes more competition in the business will make concepts like orbital data centers more practical.

“You can rethink a lot of the science fiction business plans of the past with the new capacity that’s out there now,” said Meyerson, who is co-founder and CEO of Interlune, which plans to harvest Helium-3 from the lunar surface.

Startups like Starcloud are still going to need a lot more capital than they’ve currently raised to realize their long term visions. Johnston said he recently had a call with SoftBank, which has promised to pour billions of dollars into data center projects with OpenAI, and a representative from the Japanese investment firm asked him: “If you had unlimited resources, could you get a 5-gigawatt data center up within five years?”

“Well, if I had unlimited resources, then yes,” replied Johnston, who said the conversation wasn’t a formal fundraising call. (Starcloud is planning to raise another round after its first launch in August.)

SoftBank representatives didn’t respond to requests for comment.

For now, startups are sidestepping the launch cost issue by sending only the most diminutive data centers into orbit. Starcloud’s first data center going up in August will cost it $2.5 million, including a shared ride on a SpaceX rocket, Johnston said. It will consume a measly 1 kilowatt.

Meanwhile at Axiom, the company planning to launch its own orbital data centers on a partner’s communications satellites by the end of the year, Aspiotis said companies building space-based computing facilities will be lucky to scale to 100 kilowatts by 2030. (Axiom is also building a commercial space station.)

“It’s a decadal business play,” said Aspiotis.

With so little data center capacity expected in orbit in the coming years, don’t expect the world’s leading AI companies to shift their computing workloads into the skies anytime soon.

The most likely initial customers for the first orbital data centers are the ones that are already in space, executives and investors say. Companies, governments and researchers that use satellites to track the weather, missing planes and climate change could process data in orbit more quickly using space-based computers before sending it on to Earth. The U.S. military could do the same for data from the Golden Dome, a new space-based weapons system proposed by President Donald Trump that could take decades and hundreds of billions of dollars to build.

Europe is nosing around orbital data centers too. A few years ago, the European Commission was intrigued enough by their potential environmental benefits that it commissioned a feasibility assessment by a French-Italian space contractor, Thales Alenia Space. The study concluded that orbital data centers were indeed feasible.

Damien Dumestier, a lead researcher on the project, said the next steps in the study are to design a proof-of-concept 50-kilowatt orbital data center by 2030, and to figure out how to use robots to deploy a constellation of them. The European Space Agency is also looking at the possibility of constructing an entirely new rocket on par with Starship to bring the data centers to space, which would add far more complexity and cost to the effort.

If and when large orbital data centers do finally take flight, there’s a good chance large companies with capital and expertise in space could end up dominating the category. Two logical players are SpaceX and Amazon, which could add computing capacity to their Starlink and Project Kuiper satellites for delivering wireless internet to Earth. But skeptics believe there’s little demand for orbital data centers now.

“There’s no market now for a lot of this stuff,” said Clayton Swope, a former policy lead for Project Kuiper at Amazon, who is now at the Center for Strategic and International Studies. “In my mind, it’s very aspirational.”

In the meantime, believers in the future of putting data centers in space urge patience. Christopher Stott, a space veteran whose company Lonestar Data Holdings is building storage computers for the moon, compared the current skeptics of orbital data centers to his former Lockheed Martin colleagues who thought the internet was a passing phase in the 1990s.

These doubters “want to stay in the back of the cave,” Stott said, describing orbital data centers as “absolutely inevitable.”

Software may be eating the world, but it was hardware-intensive startups that scooped up this week’s largest funding rounds, including heavily funded defense tech unicorn Anduril Industries, and orbital spacecraft developer Impulse Space.

1. Anduril Industries, $2.5B, defense tech: Defense tech startup Anduril Industries has raised $2.5 billion in a Series G round led by Founders Fund, more than doubling its valuation to $30.5 billion post-money, the company’s chairman, Trae Stephens, told Bloomberg TV on Thursday.

2. Impulse Space, $300M, spacetech: Impulse Space, a startup focused on orbital spacecraft that pull satellites into more precise locations, raised $300 million in a Series C round led by Linse Capital. The financing brings total funding to date for Redondo Beach, California-based Impulse to $525 million, per Crunchbase data.

3. TAE Technologies, $150M, fusion energy: Foothills Ranch, California-based TAE Technologies, a developer of commercial fusion power, raised $150 million in new funding. Chevron, Google and NEA participated in the financing, which brings total equity funding to date to $1.5 billion, per Crunchbase data.

4. (tied) Fever, $100M, events platform: Fever, a technology platform for live experiences, announced that it raised over $100 million in new funding from L Catterton and Point72. Founded in 2014, the New York-based company offers a platform for finding and booking concerts, games and other live events in a selected city.

4. (tied) Infleqtion, $100M, quantum computing: Boulder, Colorado-based Infleqtion, a developer of atom-based quantum systems for sensing and computing applications, raised $100 million in Series C funding. Backers included Glynn Capital Manaagement, Morgan Stanley’s Counterpoint Global, S32 and SAIC.

6. Spyglass Pharma, $75M, ophthalmic therapies: Spyglass Pharma, an Aliso Viejo, California-based developer of therapies for long-term management of glaucoma and other chronic ophthalmic diseases, raised $75 million in a Series D round led by Sands Capital. The financing brings total funding to date for the 6-year-old company to $185 million, per Crunchbase data.

7. Vivrelle, $62M, fashion: Vivrelle, a subscription service for luxury fashion accessories, raised a $62 million Series C funding round backed by Protagonist. Founded in 2018, the New York-based startup offers access to a collection of designer handbags, jewelry and diamonds for a monthly membership fee.

8. Allay Therapeutics, $58M, pain management: San Jose, California-based Allay Therapeutics, developer of medications for post-surgical pain management and recuperation, closed a $57.5 million Series D financing co-led by Lightstone Ventures and ClavystBio. The company is currently carrying out clinical trials for patients undergoing knee replacement surgery.

9. Antheia, $56M, synthetic biology: Antheia, a startup using synthetic biology to sustainably produce medicines, raised $56 million in a Series C led by EDBI and Global Health Investment Corp. The Menlo Park, California-based company plans to use the investment to scale production of thebaine, a key ingredient in Narcan, and to launch new products.

10. Pactum, $50M, artificial intelligence: Pactum, a startup offering an AI-enabled platform for enterprise procurement, raised $50 million in fresh funding, per a securities filing. The financing brings total funding to date for the Mountain View, California-headquartered company to $105 million, per Crunchbase data.

Winklevoss-backed crypto group Gemini files for Wall Street IPO

Filing comes a day after stablecoin operator Circle surged in its public debut

Gemini, the cryptocurrency exchange run by the Winklevoss twins, has filed to list in the US, aiming to capitalise on soaring investor demand for some of the digital asset industry’s biggest names.

The company said on Friday it had filed confidential paperwork with the Securities and Exchange Commission for an initial public offering, a day after stablecoin operator Circle more than doubled on its debut in New York.

A confidential filing with US regulators allows companies to progress their plans to list before publicly unveiling their financials closer to pursuing a flotation.

Investor enthusiasm for crypto assets in the US has been rekindled by the election of President Donald Trump, who promised industry-friendly policies and a reversal of the crackdown under Joe Biden and former SEC chair Gary Gensler. Gemini has been one of the beneficiaries, with US regulators closing an investigation against it in February.

Cameron and Tyler Winklevoss, who founded Gemini in 2014, were among the most vocal critics of the regulatory clampdown on digital assets companies. Both became active supporters of Trump, contributing to his campaign and to a political action committee funded by the industry to support pro-crypto politicians in Washington. 

The two men are also among the largest personal holders of bitcoin, with about 70,000 coins. Bitcoin is trading at about $105,000, just shy of its record high.

Gemini’s announcement comes a day after stablecoin operator Circle Internet soared almost 170 per cent on its Wall Street debut, having increased both the number of shares on offer and their price in response to strong investor demand. The stock rose a further 44 per cent on Friday to give it a fully diluted market capitalisation of more than $31bn.

The US IPO market is flickering back to life after Trump’s aggressive trade war rocked global equity markets over the spring, delaying several large tech listings to the frustration of investment bankers who had expected a rush of deals under the new administration. 

Stocks have rebounded over the past two months as the Trump administration has rowed back from some of its more bombastic tariff plans, while the Vix index, a so-called fear gauge that prices expectations for future stock market volatility, has fallen back below 20 from a high above 50 in early April.

“The IPO market does not love volatility,” said Karen Snow, chief executive of Rose & Co Capital Advisors and Nasdaq’s former global head of listings. “About 90 per cent of [US] IPOs take place when the Vix is at 25 or below.”

On Monday, San Francisco-based fintech Chime Financial said it hoped to raise up to $832mn in an IPO later this month, albeit at a far lower valuation than it sought four years ago. Shares in retail trading platform eToro have climbed more than 30 per cent since it listed on Nasdaq in mid-May, and product design software group Figma filed for an IPO in mid-April.

Asset-backed digital currencies could pose risks to bedrock of global financial system

Crypto has got its claws into the White House and is swiftly growing into a global financial stability risk.

Up to now, what happened in crypto stayed in crypto. If you buy a token of some kind — there are thousands of them out there, linked to everything from dog memes to the US president — and something goes wrong, it is on you. If the site where you store them goes bust or gets hacked, tough luck — you knew the risks.

“Hey, Crypto Mom, where’s my bailout?” is not an appropriate response to suffering losses, as Hester Peirce, the crypto-friendly commissioner at the Securities and Exchange Commission, said in May.

But we are rapidly reaching the point where the crypto ecosystem poses risks to mainstream markets, or what the crypto community would gratingly call “TradFi”. In fact, it is reaching the most trad market of all, US government debt — the bedrock of the global financial system.

The connective tissue here, between dog tokens and real life, is stablecoins — akin to normal currencies, chiefly dollars, but in the crypto ecosystem. It is much easier to hop in and out of bitcoin or other tokens by using a stablecoin such as tether or Circle’s USDC as a base, rather than a clunky real-world currency.

With stablecoins, the promise is that a dollar is a dollar. They are meant to be backed one for one with reserves of equal value. Holders do not receive interest (but the operators often do, to the tune of billions of dollars a year) or any adjustment for inflation. But they do get to shoot something that smells a bit like real money around the cryptosphere with great ease.

For years, this has been an intriguing sideshow on the periphery of global finance. Stablecoin operators have shown varying degrees of willingness to spell out exactly what they own, with Tether executive Paolo Ardoino once telling the Financial Times that was his “secret sauce”.

Back in 2021, warnings were emerging about the risk this poses to normal markets. Rating agency Fitch pointed out that if a stablecoin were to fold for any reason, it could be forced to sell all of its holdings — the dollar assets held in reserve — upsetting the underlying markets.

Last month, a working paper from the Bank for International Settlements, the central bank for central banks, cranked up the volume on that warning. In it, Rashad Ahmed and Iñaki Aldasoro calculate that when stablecoins (of which tether is by far the biggest and most impactful) draw in funds, and churn them into reserves, that has a marked impact on the value of short-term US government debt. 

That is a reassuring sign that stablecoin operators are indeed buying reserves to match their inflows. Still, this is a substantial and little understood market force. According to the researchers, large inflows of $3.5bn over five days can place enough upward pressure on the price of short-term US government debt to pull down yields by up to 0.025 percentage points over 10 days.

That does not sound like much, but the paper says it is “comparable to that of small-scale quantitative easing on long-term yields” — in the same ballpark as central bank efforts to stimulate a flagging economy.

So that covers when stablecoin money comes in. But what happens when it goes out is more important: the impact on short-term government debt prices is two to three times larger. When money comes in, stablecoin operators can exercise some discretion over precisely how and when to buy reserves. When they face redemptions, they have to act faster.

We might cheer the ascent of stablecoins as a side-effect of the relentless crypto boosting from the Trump family, as it helps on the margins to lower borrowing costs. (Although it might be preferable for people to cut out the intermediary, buy short-term debt and enjoy the interest payments.) But if anything were to go wrong in crypto in future — hardly a wild theoretical exercise — we might all feel the ripple effects. And either way, it all adds an additional layer of complication for central banks.

“If the stablecoin sector continues to grow rapidly, it may eventually affect the pass-through of monetary policy to Treasury yields,” the researchers say, adding that the “opacity” of reserve holdings disclosure by tether “complicates” efforts to model its possible impacts.

The financial stability risks embedded in all this, at a time when the US is seeking to foster greater growth in stablecoins, are obvious. Stablecoin operators hold more short-term US debt securities than large foreign investors such as China. Between them, they bought more than $40bn worth of Treasury bills in 2024.

This moment calls for heavier regulatory requirements on stablecoin operators to report — in detail and often — what they are buying and selling. But the gutting of crypto regulation and enforcement under the second Trump administration suggests this is desperately unlikely. None of us should be surprised to see shocks emanating from this space, either while Trump is in office or beyond.

Jonathan Anderson, Dior’s playful new creative director

After creating car-shaped dresses and tomato bags at Loewe, he now faces the challenges of one of the great fashion houses

What do a tomato, a frog, and a pigeon have in common? They have all inspired bags created by Jonathan Anderson. The Northern Ireland born designer, newly appointed as creative director at Dior, is a master of the viral fashion object, whether for his own label JW Anderson or at Loewe, where he spent over a decade.

Anderson’s realistic pigeon clutch appeared on Sex and the City spin off And Just Like That in 2023, but it’s the tomato bag that best expresses his ability to capture attention in the distracted modern age. A photo of an heirloom tomato with sculptural ridges went viral on X last June, when a user posted it with the caption “This tomato is so Loewe I can’t explain it.” Anderson responded on Instagram by presenting a tomato-shaped leather bag with the words “Loewe meme to reality”.

It was the sort of brand-building playfulness that the 40-year-old excels at. A strong identity is what a fashion house needs, especially during a luxury slowdown. Major brands Chanel, Valentino and Gucci have all changed their head designers since early 2024. When Anderson’s appointment as creative director of women’s, men’s and haute couture collections at Dior was announced on Monday, Bernard Arnault, chair and chief executive of owner LVMH, said “his incomparable artistic signature will be a crucial asset in writing the next chapter of the history of the House of Dior.”

Anderson replaces Maria Grazia Chiuri, who has designed womenswear and couture at Dior since 2016, and Kim Jones, who left his position as menswear designer in January. He is the first designer since founder Christian Dior to take on both men’s and women’s collections and will continue with his own label.

His approach to design should help him in writing Dior’s next chapter. In 2010, he said of his work “I’m trying to tell a story. I’m originally from Northern Ireland and storytelling is a massive part of my upbringing.”

Anderson was born in Magherafelt, County Londonderry in 1984. His father is the former Irish rugby captain Willie Anderson and his mother was a teacher. His maternal grandfather was a textile designer and Anderson has said that while he didn’t “consciously think that this was the job for me . . . I was always very obsessed with fashion, and the artistic process in general.”

Diagnosed with dyslexia in primary school, he moved to the US aged 18 to pursue a shortlived passion for acting. After moving back to Ireland he took his first fashion job at Dublin department store Brown Thomas where he met Manuela Pavesi, the visual communications director for Prada windows. She inspired him to enrol at the London College of Fashion, where contemporaries remember him as exceptionally focused and ambitious. After finishing his degree he had a stint as a visual merchandiser for Prada before launching his own menswear collection in 2008. Two years later he added womenswear to almost instant buzz.

In 2012, Kate Phelan, then creative director of Topshop, asked him to design a collection as part of its NEWGEN talent initiative. She recalls that he always had “one eye on the commercial and one eye on the creative which is a rare combination.” His energy to keep making products was boundless, she added. “I think you still see that now in him.”

But Anderson’s biggest break came in 2013 when LVMH took a minority stake in JW Anderson and hired him as creative director for Spanish leather goods brand Loewe.

There, his catwalk shows became a distinctive blend of big conceptual ideas — dresses in the shape of cars and shoes that featured deflated balloons — striking tailoring, and impressive craftsmanship. For the men’s spring/summer 2015 show the brand introduced the origami-like Puzzle bag, one of the few design classic bags of the last decade.

Former YOOX Net-a-Porter president Alison Loehnis says Anderson has the ability to “toggle between the intellectual and an understanding of what his consumer is going to want”.

He now faces the challenges of a bigger fashion house (one that some claim is suffering from brand fatigue). Dior has a huge global reach and a 78-year heritage.

In the background is the cautionary tale of John Galliano, the British designer who started at Dior in a blaze of glory in 1996. Fifteen years later he was fired after being filmed making drunken antisemitic comments. The designer admitted addiction issues but he has also cited the intensity of his workload as a factor in his self-destruction.

Anderson is a calmer character who conveys a mix of the arty and the earthy. After runway shows he seems unfazed by the hectic scrum encircling him as he expounds on the various recherché influences behind the collection.

Curiosity is a quality that comes up frequently. The filmmaker Jack Elliot Hobbs, who worked with him on a short film called Who Is The Painter? confirms that he is “constantly looking for the next person or thing to develop”. This could be the secret sauce to maintaining the “newness” that fashion craves.

Anderson’s appointment will be judged at his debut menswear show for Dior in June. The likelihood is that he will be presented with a dazzling bouquet of LVMH flowers rather than pelted with (meme-like) tomatoes.

U.S. drug companies are increasingly licensing experimental medicines from Chinese firms. The recent deals could be worth up to $25 billion.

The drug industry has spent months going along with the Trump administration’s efforts to move manufacturing and investment into the U.S., and to disentangle the U.S. and Chinese economies. A long list of big pharmaceutical firms have committed tens of billions of dollars to factories and research facilities in the U.S.

In the past few months, though, U.S. pharma companies have simultaneously supercharged their interest in China-based biotechs, announcing what are likely to be the biggest deals ever for the rights to experimental medicines invented by Chinese companies.

So far, the Trump administration has been silent on the deals, which are worth around $25 billion in upfront and potential milestone payments and seem to fly in the face of White House policy.

The deals could one day result in new options for sick patients, but they also pose a major risk to U.S. biotech firms. The domestic drug pipeline relies on capital from Big Pharma, and now those funds may be going to start-ups in Shanghai, rather than Cambridge, Mass.

The recent deals follow a successful trial result last year by U.S. biotech Summit Therapeutics of a new immunotherapy cancer drug it licensed from Chinese firm Akeso. In the trial, the drug outperformed Merck’s top-selling Keytruda, raising hopes that it could prove a major advancement in cancer treatment.

Since then, Big Pharma companies have rushed to get their own drugs to compete with Summit’s product. There was one deal in November, when Merck licensed an experimental drug from a Chinese drugmaker for around $500 million up front and another $2.7 billion in potential milestone payments.

The biggest deals have come in just the past two weeks. In mid May, Pfizer said it would pay the Chinese biotech 3SBio $1.3 billion up front for its own competitor to the Summit drug, plus billions more in potential milestone payments. Pfizer is also making a $100 million equity investment in 3SBio.

This past week, Bristol Myers Squibb said it would pay $3.5 billion over the next three years, plus billions more in potential milestone payments, for half the rights to a similar drug developed by a Chinese company called Biotheus, which German biotech BioNTech acquired a few months ago.

These numbers would be big for any biotech licensing deal—the drugs remain far from approval. For Chinese-developed drugs, the payment sizes are unprecedented.

“These are pretty sophisticated companies allocating major capital here,” says Craig Garthwaite, a professor and director of the program on healthcare at Northwestern University’s Kellogg School of Management, of the latest deals. “It’s demonstrating the validity of the science.”

It’s all happening while Big Pharma is doing its best to show the Trump administration that it can adapt to the president’s agenda, as the industry seeks to head off threatened tariffs and drug price limits. Eli Lilly, Johnson & Johnson, Bristol Myers, and others have all announced tens of billions of dollars in planned U.S. investments in manufacturing and research facilities this year.

The White House didn’t respond to a request for comment.

Washington has been increasingly anxious about the rapid development of China’s biotech ecosystem. The Biosecure Act, which hasn’t passed Congress, targets complex drug manufacturing in China, while a recent report commissioned by Congress called on the U.S. to take “swift action” to compete with the Chinese biotech sector.

Amid those stalled efforts, the pace of Chinese biotech innovation is picking up. Until recently, the U.S. drug industry had largely seen Chinese biotechs as a source for cheaper “me too” assets, which echo but don’t duplicate existing medicines. Now, with the latest cancer drug deals, Chinese companies are innovating a new class of drugs that every Big Pharma firm seems to think it needs to get in on.

The Summit drug that got investors, analysts, and companies excited last year is known as a PD-1/VEGF bispecific antibody, which combines two proven cancer-fighting tools. Daina Graybosch, an analyst at Leerink Partners, says that companies in the U.S. and Europe have been experimenting with similar combinations, but never tested them in humans.

That’s because early-stage human tests are cheaper and easier to run in China than the U.S. All but the largest U.S. biotechs generally develop just one or two experimental medicines at a time, while Chinese biotechs of a similar size “could have a dozen” different drugs in trials, according to Cantor Fitzgerald analyst Li Watsek.

The combination cancer drugs ultimately weren’t exciting enough for the U.S. biotechs to prioritize. But Chinese companies moved forward with different variations. When Summit and its Chinese partner got promising results on an antibody that combined PD-1 and VEGF, there were multiple other Chinese biotechs with their own PD-1/VEGF combinations ready to go.

None of this is great news for the U.S. biotech sector, which has been battered in recent years by declining share prices, cash shortfalls, and other challenges.

Over the past 12 months, the SPDR S&P Biotech exchange-traded fund is down 11.5%, versus a gain of 11.7% for the S&P 500.

Watsek says that part of the problem for investors is that it’s hard to know what Chinese biotechs are actually working on. “Right now it’s a little bit of a black box, because a lot of these Chinese companies, they may have assets, but it’s impossible to track,” she says. “There could be a dozen molecules that are in development that you may not have even heard of.”

Over the long term, the worries get potentially more complex.

“At some point, it’s going to get terminal velocity, and we’re going to have a very dangerous competitor next to us,” says Joseph Grogan, a senior policy official during the first Trump administration who is now a nonresident senior scholar at the USC Schaeffer Institute. “If the Chinese establish the flywheel of the ecosystem that we built—of private-sector companies, research and development, government support, and strong but prudent regulatory foundation—then we’re cooked.”