Technology

➚ MOVE FAST: Chip. Nvidia CEO Jensen Huang got a shoutout from President Donald Trump during the Saudi Arabia leg of his Gulf trip. The company will also sell thousands of chips to the kingdom as part of its AI partnership, aided by the US scrapping limits on such deals for the region.

➘ BREAK THINGS: Old block. Apple CEO Tim Cook’s days as a Trump whisperer may be coming to an end. The president called out his notable absence during his Gulf visit, which included a slew of tech leaders. The region isn’t a critical market for the company, but Trump’s comments don’t bode well as he looks to finalize a raft of trade deals.

Trump’s visit to the Gulf this week has so far shepherded $300 billion in deals with Saudi Arabia and $200 billion with Qatar, many of which focused on developing the region’s AI footprint.

Saudi Arabia will focus much of its AI investments through a new Public Investment Fund subsidiary called HUMAIN — launched on the eve of the US president’s visit — which partnered with American big tech giants to announce many of the deals. This includes a $5 billion partnership with Amazon Web Services to build out the country’s AI infrastructure, adding to AWS’ previously committed $5.3 billion.

Nvidia said it will sell hundreds of thousands of chips to support the country’s data center buildout, with the first set of its newest Blackwell chips going to HUMAIN, which also selected Nvidia competitor Groq for its inference work, Semafor’s Kelsey Warner scooped.

Meanwhile, Qatari-owned Al Rabban Capital partnered with Colorado- and UK-based Quantinuum to invest $1 billion in quantum technologies and a supporting workforce in the US.

New Gulf leaders are “forging a future where the Middle East is defined by commerce, not chaos; where it exports technology, not terrorism,” Trump said in Riyadh on Tuesday. The president is also visiting the UAE this week, where news of more potential deals is swirling.

If it feels like you’ve been seeing less of Elon Musk on X, it’s probably because you have (also, please go touch grass). The Tesla CEO cut his “xeet” count nearly in half in April from his rampant posting pace in the months prior, according to a Semafor analysis. Averaging at 40 posts per day last month (a total of 1,200), it’s far from a true social media detox, though it’s a marked decrease from Musk’s daily average of 68 posts over the six months prior — leading up to the election and through the beginning of Trump’s second stay at the presidency. His busiest social engagement was November, with more than 2,400 posts.

The recent pullback came as Musk’s “special government employee” status to the president came to an end and as Trump’s “Liberation Day” tariffs roiled global markets. Musk reportedly broke from Trump on economic policy, asking him privately to reverse the tariffs that would make his Tesla cars and SpaceX’s rockets more expensive. Meanwhile, Tesla took a quarterly revenue hit, and the $170 billion in savings touted by the Department of Government Efficiency fell short of its trillion-dollar goal.

So far in May, Musk has posted more than 750 times, which puts him on track to land higher than April’s figure, but below the totals of previous hot months.

— Rachyl Jones

Meta released a massive trove of chemistry data Wednesday that it hopes will supercharge scientific research, and is also crucial for the development of more advanced, general-purpose AI systems.

The company used the data set to build a powerful new AI model for scientists that can speed up the time it takes to create new drugs and materials.

The Open Molecules 2025 effort required 6 billion compute hours to create, and is the result of 100 million calculations that simulate the quantum mechanics of atoms and molecules in four key areas chosen for their potential impact on science.

“We’re talking about two orders of magnitude more compute than any kind of academic data set that’s ever been made,” said Sam Blau, a research scientist at the Lawrence Berkeley National Laboratory who worked with Meta on the project. “It’s going to dramatically change how people do computational chemistry.”

Scientists have long created mathematical representations of the way atoms and molecules interact, but it has been prohibitively expensive to do those calculations on a large scale. In recent years, Meta has been putting unused data center capacity to work on the problem, employing complex math known as Density Functional Theory to map out the way atoms physically interact with other atoms and molecules.

Those calculations help researchers come up with ideas for new drugs, battery technology, and other breakthroughs.

Meta’s Fundamental AI Research (FAIR) team also believes that the new data set and AI model will help advance overall AI technology. In order to reach what Meta calls “Advanced Machine Intelligence,” its researchers say AI must build a “world model,” or an understanding of its physical surroundings. That includes atoms, the building blocks of the physical world.

I spoke with Josh Wolfe, co-founder of Lux Capital, about the venture firm’s helpline for American scientists. Lux is among a handful of VCs that make riskier bets on companies trying to commercialize basic research, cutting checks for drug discovery companies like Recursion and on robotics like Physical Intelligence.

As we’ve written several times lately, basic research is under threat in the US. That not only impacts venture firms like Lux Capital, which depend on it to feed the pipeline of innovation, but it could set the US back decades, crippling both its national security and economy.

Lux wants researchers across the US to reach out and inquire about commercializing their discoveries, either by founding a startup or by finding a place within the private sector where they can continue what they were doing.

“We’re hearing from scientist friends that funding is cut, they’re going to lay people off. People might be losing visas. It’s a multidimensional problem,” he says.

Wolfe is quick to say this isn’t political. The current crisis is just accelerating an unfortunate post-Cold War trend of government-funded R&D dropping as a percentage of US GDP. And he says universities now sparring with the Trump administration could have done more to save their research programs, including protecting students.

But like most venture capitalists, Wolfe prefers to look for solutions. “We’re willing to take more science risk than ever,” he says. The firm earmarked $100 million for the project.

It’s a valiant effort, but even Wolfe knows it’s not enough. He hopes it will inspire others with deep pockets to step in, too. And it shouldn’t actually be that difficult. The National Science Foundation’s budget is about $10 billion. The NIH’s is $48 billion — about what Elon Musk paid for Twitter.

There is enough wealth in the US to find a way to limit the damage by funding critical projects in areas like cancer research and materials science.

Google DeepMind unveiled AlphaEvolve Wednesday, a new coding agent powered by its Gemini AI model. That may not sound groundbreaking, but it’s a waypoint on the road toward a new world of AI models that can improve on their own.

AlphaEvolve was built by the same team that created AlphaTensor a few years ago with a similar objective in mind: Creating novel algorithms.

This time around, the system is more autonomous, casting a much wider net of what’s possible and leading to new scientific discoveries. For instance, it suggested changes to the design of Google’s latest Tensor Processing Units, its in-house AI chips, and helped design a more efficient method for matrix multiplication in Gemini’s architecture.

It’s difficult to say how far away we are from: “Hey Google, create the world’s best AI model,” but progress in that direction continues.

As AI continues to evolve at a rapid pace, companies are shifting from experimentation to real-world deployment and practical use within their businesses.

Join Anthropic Co-Founder Jack Clark, World Labs Co-Founder and CEO Dr Fei-Fei Li, Booking Holdings President and CEO Glenn Fogel, Capital G Partner Jill Chase, and Singapore Economic Development Board Executive Vice President Ih-Ming Chan for a discussion on the breakthroughs driving AI. Discussions will dive into how global, national, and regional AI ecosystems are shaping the technology’s future, and why building the policy frameworks governing them is more critical than ever for its potential.

May 21, 2025 | San Francisco, CA | Request Invitation

PayPal announced a partnership with Perplexity Wednesday that will allow users to make purchases directly on the AI search engine using PayPal accounts.

It’s the latest in a slew of new deals to profit from the coming wave of AI-powered commerce.

Visa announced a partnership with Perplexity and other AI providers, and Mastercard is also getting into the game. Meanwhile, Stripe is betting on stablecoins as having a big role in AI commerce.

It’s hard to say whether AI commerce will just be… e-commerce. Anthropic’s Model Context Protocol, which is sweeping the tech industry right now, is really just replacing existing APIs, and there’s reason to believe the new payment companies will become the old payment companies.

But there’s also a possibility that something interesting happens and AI commerce becomes something different altogether. If AI agents are shopping for us, there’s an opening for a whole industry of middleware that routes agents, negotiates pricing, facilitates purchases, and conveys information and data.

That’s where you could see Stripe’s stablecoin vision becoming a reality. Billions of micro transactions don’t really make sense on today’s high-fee payment rails. It might work if it utilizes blockchain technology, as one idea. But we don’t really know where all this is headed yet.

Pope Leo XIV cited AI as one of the deciding factors behind his papal name. In his speech to the College of Cardinals on Saturday, the new pope said his chosen name calls back to Pope Leo XIII, who wrote a famous document on how society should respond to the social and economic problems caused by the Industrial Revolution.

In the translated speech, Pope Leo XIV said, “The Church offers to everyone the treasury of her social teaching in response to another industrial revolution and to developments in the field of artificial intelligence that pose new challenges for the defence of human dignity, justice and labour.”

The statement suggests the new head of the Catholic Church will follow in the steps of the late Pope Francis, who advocated for ethical AI and warned about its use in warfare. In a document published in January, the Vatican said that because humans are created “in the image of God,” even the most advanced AI can’t match human thinking.

China’s domestic and international energy investments are getting greener, but face some big challenges ahead, a pair of new reports shared first with Semafor’s Tim McDonnell argue.

The temporary ceasefire in the US-China trade war agreed over the weekend is good news for Chinese exporters of batteries and other clean tech and their customers in the US. But the two studies, taken together, illustrate how China is navigating a moment of heightened external volatility by falling back on its enormous domestic market, Tim argues.