TheChek News Flow

Chips to be unveiled at developer conference expected to drive growth next year and beyond, but will depend on Big Tech’s spending spigot staying on full blast

Nvidia NVDA 5.27%increase; green up pointing triangle has barely started shipping its latest artificial intelligence chips, and everybody is already looking to hear about the next ones. For the world’s most valuable semiconductor company, that is a high-quality problem.

Nvidia’s annual GTC developers conference in San Jose, Calif., used to be a relatively modest affair, drawing about 9,000 people in its last year before the Covid outbreak. But the event now unofficially dubbed “AI Woodstock” is expected to bring more than 25,000 in-person attendees flocking to the company’s backyard this week, all lining up to hear what the AI kingmaker has up its sleeve for the months and years ahead.

Nvidia’s Blackwell AI chips, the main showcase of last year’s GTC conference, have only recently started shipping in high volume following delays related to the mass production of their complicated design. Blackwell is expected to be the main anchor of Nvidia’s AI business through next year. Analysts expect Nvidia Chief Executive Jensen Huang to showcase a revved-up version of that family called Blackwell Ultra at his keynote address on Tuesday.

But Nvidia watchers are especially eager to hear more about the next generation of AI chips called Rubin, which Nvidia has only teased at in prior events. Ross Seymore of Deutsche Bank expects the Rubin family to show “very impressive performance improvements” over Blackwell. Atif Malik of Citigroup notes that Blackwell provided 30 times faster performance than the company’s previous generation on AI inferencing, which is when trained AI models generate output. “We don’t rule out Rubin seeing similar improvement,” Malik wrote in a note to clients this month.

Rubin products aren’t expected to start shipping until next year. But much is already expected of the lineup; analysts forecast Nvidia’s data-center business will hit about $237 billion in revenue for the fiscal year ending in January of 2027, more than double its current size. The same segment is expected to eclipse $300 billion in annual revenue two years later, according to consensus estimates from Visible Alpha. That would imply an average annual growth rate of 30% over the next four years, for a business that has already exploded more than sevenfold over the last two.

High marks, to be sure. And doable—if the world’s largest tech giants keep dumping significant amounts of capital investments on AI infrastructure. But by that point, such spending will only be justified if there is substantial demand from consumers and businesses for generative AI services, and a willingness to pay a premium for them.

Such an outcome is still a bit iffy, especially with the potential for the global economy to slip into a recession sparked by tariffs, geopolitical instability and inflation. Nvidia has also been haunted by worries about competition with in-house chips designed by its biggest customers like Amazon and Google. Another concern has been the efficiency breakthroughs claimed by Chinese AI startup DeepSeek, which would seemingly lessen the need for the types of AI chip clusters that Nvidia sells for top dollar. Nvidia’s stock has actually picked up some gains against the market’s turbulence in recent days, but the shares are still down more than 9% so far this year, more than double the S&P 500’s decline.

That at least sets a more reasonable stage for this year’s GTC. Nvidia’s stock is now trading for less than 27 times this year’s projected earnings, which is a 23% discount to the multiple it was commanding at last year’s conference and cheaper than the stocks of most other megacap tech companies that aren’t growing nearly as fast. The tech world still expects a lot of Nvidia and its latest chips. But at least the market is no longer pricing in guaranteed blockbusters.

The Met Has Millions in Stolen Art. It’s Not Waiting to Be Asked to Return It.

The museum is taking the lead on re-evaluating its art and artifacts to determine where these works came from in the first place

At the Metropolitan Museum of Art in late February, an increasingly familiar scene played out: A museum was restituting a work of art, giving back something from its collection to the place where it originated—in this case, Greece. The announcement, however, was more celebratory than sheepish. It was being done with the fanfare more often associated with acquiring a great object rather than returning one. That’s because, in this case, the Met itself launched the inquiry into the sculpture’s origin.

The work in question was a small bronze sculpture of the head of a griffin, a mythological creature, made in the 7th century B.C. It had been on display at the Mets since 1972. Max Hollein, the Met’s director and CEO, spoke to the assembled crowd. The piece, he said, “could not have legitimately left” Greece.

Over the last few years, museums have had to respond to inquiries—sometimes as part of legal claims that can come from countries or former owners, or are initiated by U.S. law enforcement—about works that were stolen, illegally excavated, exported or traded improperly.

But with the griffin head, the Met didn’t wait to be asked. Greek authorities weren’t attempting to reclaim it. Instead, in a turn that shows how the process of repatriation and restitution has changed, the Met took the initiative to investigate how the griffin head got into its collection. The museum found it had disappeared from the Archaeological Museum of Olympia in the 1930s.

Last year the Met appointed Lucian Simmons, a lawyer and former Sotheby’s executive, to be its first-ever head of provenance. He leads a team of 10 full-time researchers, which the Met says is the biggest staff dedicated to this task at any museum.

Simmons acts as “air-traffic control,” he said, providing support, supervision and coordination to the provenance research specialists, who are embedded within the museum’s departments.

Simmons’s broader philosophy is that more sunlight is better when it comes to determining pieces’ back stories. “We need to know the truth,” he said. “Sometimes it’s uncomfortable.”

With these claims, the stakes are often high: In 2022, local and federal authorities seized 27 objects worth a reported $13 million from the Met and returned them to Italy and Egypt, their countries of origin.

Provenance research, or finding a comprehensive record of an art object’s previous ownership and history, is “core museum work” these days, said Hollein, certainly when it comes to vetting new acquisitions. But now, more museums are examining the back story of objects they’ve had for decades and then taking action based on what they find.

“Museums used to take a wait-and-see approach,” said William M. Griswold, the director of the Cleveland Museum of Art and the chair of the cultural property task force of the Association of Art Museum Directors. “Now, they are much more proactive about investigating and finding ways to resolve issues that pertain to provenance.”

Hollein noted how much the landscape of looking into a work’s provenance had changed. “Forty, fifty years ago, nobody cared about documents. It’s not because everything was illegal. That’s just not how it worked.”

The Met has more than 1.5 million objects in its collection, six million visitors a year and an endowment of $4.5 billion.

With the griffin head, Met staffers flagged in 2018 that there were issues, and Hollein then reached out to the Directorate of Antiquities and Cultural Heritage of Greece. The mystery was never solved definitively, and hence the museum could not be sure the piece met its standards.

But it’s not gone for good. The bronze is coming back to the Met as a loan next year.

When the billionaire Leonard N. Stern, a real-estate developer and heir of the pet-product company Hartz Mountain, wanted to give the Met a collection of Bronze Age sculptures from the Cycladic Islands, which are now part of Greece, the Met couldn’t determine the art’s origin.

“Ninety percent of Cycladic art comes from unknown contexts, and a lot of it probably comes out of tombs,” said Seán Hemingway, the Met’s chief curator of Greek and Roman art.

Instead of turning down the gift, the Met worked with the Greek government and the Museum of Cycladic Art to establish a long-term partnership. The works will be owned by Greece, but most will be on display at the Met for the next 25 years, and the country will loan other Cycladic art to the museum. Separately, Stern is supporting extensive joint research and conservation projects.

Hollein said the reaction to such an offer had changed completely. “Maybe like 40 years ago, it would have been, ‘We’ll take it,’” he said. “Then maybe 10 years ago, it would have been, ‘It’s impossible for us to get right.’

Sometimes, the outcomes of provenance research don’t result in a piece being returned, as with Jean-Louis Lemoyne’s “La Crainte des Traits de l’Amour” (1739-40), a Rococo marble sculpture in the Met’s Petrie Court depicting a woman startled by a tiny Cupid.

Last year, an associate curator in the European Sculpture and Decorative Arts department flagged a gap in its ownership records. Upon further investigation, the museum confirmed that the sculpture, once owned by a branch of the Rothschild banking family, had been seized during World War II, a fact previously unknown to the museum. After being taken from the family, “it was put on a truck by Nazis in Paris,” Simmons said. “It was taken to the Jeu de Paume and then shipped to Berlin.”

Somehow, the work found its way back to the Rothschilds, who later sold it, and the Met bought it from a dealer, so it didn’t clearly call for restitution. Instead, new information explaining this history will be added to the placard describing the work, part of a new New York state legal requirement that requires New York museums to publicize when an artwork was stolen by the German Army.

Watchdog will begin assessing tech groups’ compliance with new obligations under Online Safety Act

Social media groups, search engines and messaging apps will next week be told to bring in strict measures to remove illegal material quickly and reduce the risk of such content, under new rules from the UK media watchdog.

Ofcom will on Monday begin enforcing the new rules designed to protect internet users from illegal content and harmful activity online. Regulators and lawmakers want additional powers to curb the rise of the kind of extreme and false information that provoked violent unrest last summer following the mass stabbing in Southport.

Under the UK’s Online Safety Act (OSA), tech companies needed to complete compulsory illegal content risk assessments by this weekend to understand how likely it was for users to encounter illegal content on their service. 

In the case of “user-to-user” messaging services, this includes how they could be used to commit or facilitate criminal offences.

So-called priority illegal content covers 17 categories from terrorism, child sexual abuse and encouraging or assisting suicide, to stalking and drugs, crime and fraud offences.

Ofcom will from next week start assessing platforms’ compliance with the new illegal harms obligations under the OSA, and begin enforcement action where there are issues and failures to comply. The OSA was passed by parliament in 2023, but is being implemented in phases this year and next.

Sites and apps will need to start implementing safety measures to mitigate risks, with a senior executive named as the staff member accountable for compliance and better moderation, easier reporting and built-in safety tests. 

Under the new rules, tech companies will have to ensure their moderation teams are resourced and trained, and set performance targets to remove illegal material quickly when they become aware of it. Platforms will need to test algorithms to make illegal content harder to disseminate.

Ofcom will initially prioritise larger sites and apps that may present particular risks of harm from illegal content owing to their size or nature, for example because they have a large number of users in the UK, or because their users may risk encountering some of the most harmful forms of online content and conduct.

Suzanne Cater, enforcement director at Ofcom, said: “Platforms must now act quickly to come into compliance with their legal duties, and our codes are designed to help them do that. But, make no mistake, any provider who fails to introduce the necessary protections can expect to face the full force of our enforcement action.”

Linklaters, the British law firm, described the new rules as “the first big regulatory deadline” under the OSA. Companies can be fined up to £18mn or 10 per cent of their qualifying worldwide revenue, whichever is greater.

Ben Packer, partner at Linklaters, said: “We’re going to find out quite quickly who’s engaged with this properly and done a thorough job of the risk assessments. I suspect there may well be some companies in scope who haven’t done much at all.”

Packer added that the threat by Ofcom of intervening may outweigh the financial imposition, noting that the regulator could order companies “to take additional measures when it comes to content moderation or user reporting or the technology deployed to detect content”.

“Our experience in other sectors is those kinds of interventions in company operations that tend to be the more impactful,” he said.

US president warns Tehran against supporting militant group which has disrupted shipping in the Red Sea

The US on Saturday launched a wave of air strikes against Houthi rebels in Yemen, as Donald Trump warned that “hell will rain down upon” the Iran-backed group if they continue to attack shipping in the Red Sea.

The US president also warned Iran to halt its support for the militant group “immediately”.

The attacks came days after the Houthis, which control most of Yemen’s populous north, warned they would resume attacks on shipping in the Red Sea and Gulf of Aden.

“Today, I have ordered the United States Military to launch decisive and powerful Military action against the Houthi terrorists in Yemen,” Trump said in a post on Truth Social. “They have waged an unrelenting campaign of piracy, violence, and terrorism against American, and other, ships, aircraft, and drones”.

The Houthis launched scores of attacks on merchant vessels and US navy ships in the Red Sea last year, saying they were acting in solidarity with Palestinians as Israel carried out its ferocious offensive against Hamas in Gaza.

The assaults severely disrupted traffic through one of the world’s most important maritime trade routes and continued despite US and UK air strikes on Houthi military targets in Yemen.

The Houthis halted their campaign after Israel and Hamas agreed to a fragile ceasefire in January under which the Palestinian militants agreed to release hostages seized in their October 7, 2023 assault on the Jewish state.

But the Houthis warned this week that they would resume their attacks on shipping after Israel blocked the delivery of all aid into Gaza in a bid to pressure Hamas to accept a revised proposal for the second phase of the ceasefire deal.

The US air strikes on Saturday targeted Houthis’ leaders, bases and missile defences, Trump said.

“The Houthi attack on American vessels will not be tolerated,” Trump said. “We will use overwhelming lethal force until we have achieved our objective”. 

“To all Houthi terrorists, YOUR TIME IS UP, AND YOUR ATTACKS MUST STOP, STARTING TODAY. IF THEY DON’T, HELL WILL RAIN DOWN UPON YOU LIKE NOTHING YOU HAVE EVER SEEN BEFORE!” Trump added.

Trump on Saturday also warned Tehran against supporting the Yemeni rebels, which are part of Iranian-backed militants that make up the so-called “axis of resistance”.

The US and Israel accuse Iran of supplying the Houthis with increasingly sophisticated drone and missile technology, as well as intelligence.

“To Iran: Support for the Houthi terrorists must end IMMEDIATELY!” he said.

“Do NOT threaten the American People, their President, who has received one of the largest mandates in Presidential History, or Worldwide shipping lanes. If you do, BEWARE, because America will hold you fully accountable and, we won’t be nice about it!”

Trump is reimposing his so-called “maximum pressure” campaign against the republic. During his first term, Trump withdrew the US from the 2015 nuclear deal Tehran signed with world powers and imposed hundreds of sanctions on the republic.

He has announced more sanctions on Tehran since returning to the White House, while also saying he wants to negotiate with Iran on its nuclear programme. But he has also held out the threat of military action.

His administration has also designated the Houthis a foreign terrorist organisation as it looks to step up the pressure on the Yemeni rebel group.

The Houthi-run health ministry said that at least nine civilians were killed and nine wounded in the US strikes on Sana'a, Yemen’s capital.

The Houthis have controlled Sana’a and most of the north for a decade since ousting the Yemeni government as the impoverished Arab state descended into civil war.

The conflict drew in Saudi Arabia and the United Arab Emirates, which led an Arab coalition that intervened to fight the Houthis in March 2015.

Yemen’s embassy in Washington and Iran’s mission to the United Nations did not immediately respond to requests for comment.

Saudi Aramco and big oil is on ‘wrong side of history’, says John Kerry

Former US climate envoy says companies have been ‘intimidated’ by Trump presidency into dropping green targets

John Kerry, the former US secretary of state and climate envoy, said companies were being intimated into dropping green targets under Donald Trump’s presidency and accused oil groups of “being on the wrong side of history”, even as the industry doubles down.

Following the CERAWeek industry conference where Saudi Aramco chief executive Amin Nasser said there was “more chance of Elvis speaking” than the shift to renewable energy working — instead highlighting reducing emissions from “conventional energy” — Kerry said this view “could not be more wrong”.

“If the head of a major fossil fuel company wants to pretend it isn’t going to happen, have at it. But they’re on the wrong side of history. And history is not just waiting to prove it. [It’s] proving it right now. This transition is happening.”

Kerry, who has joined Galvanize Climate Solutions, the investment group founded by fellow Democratic presidential contender Tom Steyer, said solar and wind energy was being deployed rapidly even in the US where renewables accounted for almost 90 per cent of new electricity.

The International Energy Agency said last year that spending on clean energy was twice that of fossil fuels, at €2tn globally. However, this was mainly meeting rising energy demand rather than replacing fossil fuels.

The 81-year-old argued some companies had been “intimidated” into dropping or downplaying green efforts, a reference to the pushback from Republicans states over environmental, social and governance issues.

A host of companies have ditched or delayed net zero emissions targets or dropped out of industry coalitions in the face of a political backlash, often citing the slow pace of change and a lack of government support.

But Kerry said that behind the scenes many were still pushing ahead with climate risk planning. “Everybody I’m talking to assures me they are staying on target, but they just don’t want to put a target on themselves in doing so, because of this weaponisation [of climate change],” he said.

Kerry said that companies and states had already changed their operations and productions for a greener future, citing carmakers as an example.

“The CEO of Ford and General Motors and Mercedes and Volkswagen, all of whom have changed their production facilities to produce electric vehicles. None of them is suddenly going to say, oh, let’s go back and make internal combustion engine cars,” he added.

Many companies would push ahead regardless of Trump, he believed. “They’re doing this because there’s money to be made, because there’s investing to be done, because this is the transformation that’s going to affect the world.”

The former US Secretary of State also said Europe’s focus on upping its defence spending was “overdue and appropriate”, but added that this did not have to be at the expense of climate action.

He was speaking in London at the sidelines of an event organised by the Sustainable Markets Initiative, a group established by King Charles to push the private sector to accelerate action on climate change.

Jennifer Jordan-Saifi, chief executive of SMI, added that many chief executives were reluctant to put “their head above the parapet” on climate issues because “they’ll be shot.”

“When you have this kind of headwind and people are worried about their shareholders and often even the CEOs are worried about their own position. They care enough about the [climate] issue but if they don’t respond to their shareholders, they’re in a Catch 22,” she said.

Kerry also played down concerns that other countries would follow the US in pulling out of Paris agreement, the global accord to tackle climate change, after Trump withdrew the world’s biggest historical polluter for the second time.

“No country with wise and thoughtful leadership is going to burn its back on embracing the new energy future for the simple reason it’s better, it’s healthier, it’s cleaner, it’s safer.”

GTC, Nvidia’s biggest conference of the year, begins Monday and runs till Friday in San Jose. TechCrunch will be on the ground covering the news as it happens — and we’re expecting a healthy dose of announcements.

CEO Jensen Huang will give a keynote address at the SAP Center on Tuesday at 10 a.m. Pacific, focusing on — what else? — AI and accelerating computing technologies, according to Nvidia. The company is also teasing reveals related to robotics, sovereign AI, AI agents, and automotive — plus 1,000 sessions with 2,000 speakers and close to 400 exhibitors.

Here’s how to watch the Nvidia GTC 2025 keynote online, along with many other sessions, talks, and panels.

So what do we expect to see at GTC? Well, Nvidia typically reserves a big chunk of the conference for GPU-related debuts. A new, upgraded iteration of the company’s Blackwell chip lineup seems likely.

During Nvidia’s most recent earnings call, Huang confirmed that the upcoming Blackwell B300 series, codenamed Blackwell Ultra, is slated for release in the second half of this year. In addition to higher computing performance, Blackwell Ultra cards pack more memory (288GB), an attractive feature for customers looking to run and train memory-hungry AI models.

Rubin, Nvidia’s next-gen GPU series, is almost certain to get a mention at GTC alongside Blackwell Ultra. Due out in 2026, Rubin promises to deliver what Huang has described as a “big, big, huge step up” in computing power.

Huang said during the aforementioned Nvidia earnings call that he’d talk about post-Rubin products at GTC, as well. That could be Rubin Ultra GPUs, or perhaps the GPU architecture that’ll come after the Rubin family. (The chips are named after Vera Rubin, the astronomer who discovered dark matter.)

Beyond GPUs, Nvidia may illuminate its approach to recent quantum computing advancements. The company has scheduled a “quantum day” for GTC, during which it’ll host execs from prominent companies in the space to “[map] the path toward useful quantum applications.”

One thing’s for sure: Nvidia could use a win.

Early Blackwell cards reportedly suffered from severe overheating issues, causing customers to cut their orders. U.S. export controls and fears of tariffs have massively depressed Nvidia’s stock price in recent months. At the same time, the success of Chinese AI lab DeepSeek, which developed efficient models competitive with models from leading AI labs, has prompted investors to worry about the demand for powerful GPUs like Blackwell.

Huang has asserted that DeepSeek’s rise to prominence will in fact be a net positive for Nvidia because it’ll accelerate the broader adoption of AI technology. He has also pointed to the growth of power-hungry so-called “reasoning” models like OpenAI’s o1 as Nvidia’s next mountain to climb.

To be clear, Nvidia isn’t exactly hurting. The company reported a record-breaking quarter in February, notching $39.3 billion in revenue and projecting $43 billion in revenue for the subsequent quarter. While rivals such as AMD have begun to encroach on the company’s territory, Nvidia still commands an estimated 82% of the GPU market. But Huang is reportedly feeling impatient to see AI applications that matter beyond the tech industry.

• DeepSeek forbids some employees from traveling abroad freely

• Local government is screening potential investors before company meetings

• Some DeepSeek employees handed in passports

Alexandr Wang has positioned the valuable startup to become a key partner for the Trump administration.

Just a day before a major AI summit in Paris commenced last month, the event’s organizers made a sudden addition to the speaker lineup: Alexandr Wang, the 28-year-old billionaire co-founder and CEO of Scale AI, a startup that specializes in fine-tuning and assessing the quality of AI models for both private companies as well as the U.S. government.

Wang spoke at a closed-door session requiring security clearance titled “Fuelling trustworthy AI innovation through collaboration between industry & government,” and during his panel’s discussion, he declared that public-private partnerships like the ones he has keenly pursued represent “the future of AI.”

Later that day, Wang attended a private dinner at the Élysée Palace hosted by French President Emmanuel Macron along with around 150 other guests, including Vice President JD Vance, foreign heads of state and other tech leaders like Sundar Pichai and Demis Hassabis. Only a few attendees were asked to give remarks to the group, said a person who went to the dinner, and Wang was one of them. He talked about growing up near a place forever marked by world-altering technology—he’s from Los Alamos, New Mexico—and also discussed how AI would play a growing role in defense.

Wang, who wouldn’t comment for this story, has reiterated the pitch for combining AI companies and the public sector again and again lately in a whirlwind marketing tour stretching from D.C. to Davos. It seems like his spiel is going over well. Scale is currently negotiating a big new Defense Department contract, according to a source familiar with the discussions, and on the same day of Wang’s panel in Paris, Scale announced a new partnership with the U.S. AI Safety Institute, making it the first private company to evaluate frontier AI models for possible risks alongside the U.S. government.

And in yet another sign of Scale’s growing presence in politics, Michael Kratsios, a former top Scale executive, is set to shortly win congressional confirmation as the new director of the Office of Science and Technology Policy. (Kratsios is an old face in Trump world: He left a job at Thiel Capital in 2017 to spend four years in the first Trump administration.)

As Wang maneuvers closer to Washington, he joins a crowded group of Silicon Valley leaders trying to align themselves with President Donald Trump, influence policy and secure lucrative government contracts. In some ways, it is classic Wang: He has long been skilled at extending his network and tying his fortunes to powerful friends, like OpenAI CEO Sam Altman, and customers like OpenAI as well as Meta Platforms, Microsoft and Cohere. Those customers pay Scale to help refine their AI, and much of that work is done by low-wage contractors overseas.

Wang has been adept at shifting his company’s focus within the AI industry several times, and he seems to see the Trump administration as a potential goldmine. The administration has, after all, expressed a strong interest in downsizing functions traditionally managed by the federal government and transferring them to the private sector. Within hours of entering office, Trump also began dismantling President Joe Biden’s AI-related legislative agenda, including nascent efforts to set up regulatory infrastructure—creating an opening for startups like Wang’s to move in.

Wang knows that the current administration takes a dim view on “AI safety” and has instead worked to reframe the issue as a matter of national security. He has positioned Scale as a leading authority in this space—and an ideal partner to assume these functions if they’re removed from federal oversight.

And Wang is well motivated to increase his government involvement. Scale faces new competition from other startups like Turing and Mercor that could eat away at its private sector contracts. Meanwhile Scale’s massive $107 million deal from the Defense Department, part of the Department’s multibillion-dollar AI warfare initiative Project Maven, ended in the fall after five years. The company is currently in the process of negotiating the renewal of that contract.

At the same time, Scale has attracted top investors, such as Accel and Founders Fund, and fetched steadily higher valuations—most recently $13.8 billion in a $1 billion funding round last spring.

But Scale has been dogged by allegations of labor violations over its treatment of its freelance workers—many of whom live outside of America—and last year, the Department of Labor launched an investigation into Scale over allegations of the misclassification of workers and unpaid wages. Scale lawyers were working toward a settlement with the government prior to the change of administration, according to a person with direct knowledge of the matter. Under Trump, the investigation seems to have stalled, part of a broader apparent push by the administration to cut down on regulations and corporate investigations.

A group of roughly 40 former Scale workers who are unhappy about the investigation languishing have launched a campaign to try to get the government to take action. Meanwhile, Scale has been sued by contractors multiple times in the past year over its alleged treatment of workers.

When Kratsios was asked during the confirmation process about Scale’s labor practices, he repeatedly said he didn’t work in the part of Scale’s business that dealt with the contract workers.

Some experts remain concerned about the risks of companies like Scale getting too involved in the government’s regulatory work. “It will exacerbate corruption,” said international AI policy analyst Ana Brandusescu, who attended the panel in Paris.

In 2020, Scale was awarded its first government contract, a $16.5 million deal to help train machine learning algorithms at the Department of Defense. The next year, the company began hiring lobbyists to push lawmakers for more AI spending in defense and intelligence-focused appropriations bills, according to congressional lobbying disclosures.

More recently, the company has turned its attention to model evaluation services, assessing the performance and safety of the industry’s top large language models. In late 2023, Scale launched its SEAL (Safety, Evaluations and Alignment Lab) research team, and later began publishing its LLM leaderboard, which ranks top AI models based on a variety of performance metrics. Since then, Scale has continued to put its stamp on projects that are shaping AI industry standards. Last year, the company collaborated with the Center for AI Safety to release the Weapons of Mass Destruction Proxy benchmark, which measures models’ hazardous knowledge in biology, chemistry and cybersecurity. In January, the two groups teamed up again to publish Humanity’s Last Exam, a set of 3,000 extremely difficult questions collected from world-class experts across a variety of scientific disciplines intended to raise the bar on AI evaluations as models get smarter.

Establishing these benchmarks—all while Wang says he believes artificial general intelligence will arrive this decade—has helped bolster Scale’s authority, possibly positioning the company to win more government contracts.

Wang’s own personal brand has also evolved to suit the changing political landscape. Last year, Scale hired Lulu Cheng Meservey, a high-profile communications consultant who has cemented her status as a go-to image maker for right-wing tech.

In June, as Trump began to pull ahead in the polls, Wang rolled out Scale’s “MEI” system—merit, excellence and intelligence—as an apparent alternative to diversity, equity and inclusion, which had become a lightning rod among Silicon Valley conservatives.

Wang’s campaigning kicked into higher gear after Trump was elected. At the inauguration in January, he sat next to OpenAI CEO Sam Altman and mingled with influencers like Jake and Logan Paul. He has also grown close with political consultant Alex Bruesewitz, who has worked with Donald Trump and his family members on their media strategies. Bruesewitz accompanied Wang to a podcast recording with right-wing influencer Theo Von, who had scored an attention-grabbing interview with Trump during the presidential campaign.

The day after the inauguration, Wang published a full-page ad in the Washington Post that read, “Dear President Trump, America Must Win the AI War,” a dramatic use of language that sent ripples across the AI industry, according to several sources. He later returned to D.C. where he met with members of the Trump administration as well as Speaker Mike Johnson and Republican senators John Thune and John Barrasso.

By the time the Chinese firm DeepSeek revealed the powerful new chatbot powered by its DeepSeek-R1 model, Wang had already established himself as an early, hawkish voice warning of Chinese AI’s growing capabilities. At Davos, he warned that DeepSeek had access to at least 50,000 coveted H100 GPU chips to train its models, despite export controls intended to keep them out of their hands.

Emphasizing the threat posed by China, while warning that the U.S. government is not investing enough in AI for defense, could of course be a useful narrative for a company hoping to capture some of those federal dollars.

Scale recently beat 170 competitors to win a new contract from the Defense Department’s commercial technology wing, for a project called Thunderforge, which is reportedly a “multimillion dollar deal.” Wang has surely made clear he is eager to move ahead. As he said in an interview this fall with Scale investor Index Partners, “My concern is, we’re just not moving fast enough.”