The Trump-Xi summit in Beijing brought together the most powerful business figures in America alongside a geopolitical confrontation that could reshape global AI, trade, and semiconductor supply chains. Elon Musk, Jensen Huang of Nvidia, Tim Cook of Apple, and a roster of Wall Street heavyweights flew to China with President Trump — and the lineup alone tells you everything about what's really being negotiated.
Who Flew to Beijing for the Trump Xi Summit?
The delegation that touched down in Beijing was unlike anything seen in modern diplomatic history. On the business side, Trump brought:
- Elon Musk — Tesla and SpaceX
- Jensen Huang — Nvidia
- Tim Cook — Apple
- Kelly Ortberg — Boeing
- David Solomon — Goldman Sachs
- Steven Schwarzman — Blackstone
- Larry Fink — BlackRock
- Jane Fraser — Citi
- Dina Pal McCormack — Meta
Notably absent were the leaders of the top AI labs — no Sam Altman from OpenAI, no Dario Amodei from Anthropic, no Demis Hassabis or Sundar Pichai from Google DeepMind. Each of those figures has a complicated and often tense relationship with the current administration, so their absence isn't shocking. But it does underscore a real disconnect between what Silicon Valley is obsessing over and what's being discussed at the highest levels of global diplomacy.
Why Was Jensen Huang on Air Force One?
One of the more surprising details to emerge was that Jensen Huang boarded Air Force One — reportedly joining in Alaska — and flew with Trump and Musk to Beijing. There was earlier reporting suggesting Jensen wasn't invited, which Trump quickly corrected via Truth Social, stating he had personally extended the invitation. The confusion likely came from the fact that these schedules are incredibly fluid, especially when you're the CEO of the most valuable company in America.
The reason Jensen's presence matters so much is simple: Nvidia sits at the absolute center of the US-China tech conflict. Chip export controls, the H20 saga, restrictions on what Nvidia can sell to Chinese buyers — all of it runs through Jensen's company. While Apple does enormous business in China, it hasn't been the focal point of semiconductor sanctions the way Nvidia has. Having Jensen in the room signals that AI chips were very much on the agenda, even if nothing definitive has been reported yet.
The visual of Jensen, Musk, and Trump stepping off Air Force One together — captured in grainy but unmistakable footage — is itself a statement. And yes, Jensen was wearing a leather jacket. The uniform endures.
What Is Actually at Stake in the US-China Summit?
The Wall Street Journal's editorial board laid out the stakes bluntly. The US enters this summit looking for stability — a word that, as was noted, is a bit ironic given who's doing the asking. The agenda covers three major areas:
Trade and Rare Earths
The best realistic outcome on trade may simply be ratifying the status quo — a tariff truce, and a commitment from Beijing not to hold the world's rare earth supply hostage again. China has made promises to buy more US farm goods and aircraft before and never followed through, so expectations are tempered. But even a stable, predictable tariff environment is something markets and business leaders can plan around.
Taiwan
The most dangerous topic on the table is Taiwan. Xi Jinping is reportedly pressing for veto power over US arms sales to Taiwan and wants the US to formally oppose Taiwanese independence — a significant step beyond the current US posture of simply not supporting it. Xi will likely frame this as a small semantic tweak. It is not. Decades of carefully balanced diplomatic language have preserved peace in the Taiwan Strait, and any crack in that foundation sends a signal to Japan, South Korea, and every other US ally in the region.
Iran and Bad Actors
The US wants China's help — or at minimum, China's non-interference — on Iran. China remains the primary financier and industrial base for what the Journal describes as the world's bad actors: Russia, Iran, and North Korea. A direct ask from Trump to Xi about whether China is actively providing intelligence support to Iran would be a historic moment. Whether it happens privately or not, we may never know.
What Is Claude Mythos and Who Can Access It?
In a related development that underscores just how sensitive AI has become at the geopolitical level, Anthropic's Claude Mythos model — previously restricted to roughly 50 organizations worldwide — is being extended to Japan's three mega-banks: MUFG, Sumitomo Mitsui, and Mizuho Bank. The expansion was reportedly facilitated by US Treasury Secretary Scott Bessent during meetings in Tokyo.
Mythos is described as capable of discovering and exploiting software security vulnerabilities far faster than previous models. The UK was the first international recipient of access. Japan is now the second. The significance here is that the US Treasury Secretary is personally vetting which organizations worldwide get access to the most powerful AI model in existence — because the national security risks of misuse are considered that serious.
The irony that some have pointed out: while Mythos access is being rationed out one country at a time, Jensen Huang is simultaneously flying to Beijing to discuss chip sales that could accelerate China's own path to a Mythos-equivalent model. The administration's AI policy, by any honest reading, contains real tensions that haven't been resolved.
How Close Is China to Matching US AI Capabilities?
By most credible estimates, China is approximately six months behind the US frontier in AI model capability. Dario Amodei has said as much publicly, and that figure aligns with what other researchers are seeing. The open-source community is also increasingly falling behind frontier closed models, which complicates China's ability to simply fine-tune publicly available weights into a Mythos competitor.
The deeper question is what China's own "Mythos moment" looks like when it arrives. In America, that moment involved congressional hearings, a near-designation of Anthropic as a critical infrastructure supplier, and an intense public debate about who controls the most powerful AI systems — government or private sector. America handled it messily, democratically, and without a clean resolution.
In China, the dynamic is entirely different. DeepSeek's founder holds what appears to be majority control of a company that has emerged as China's most credible frontier AI lab — with two billion dollars of his own money deployed at low dilution. When the CCP decides it wants to consolidate compute resources or redirect a company's research priorities, the mechanism for doing that is very different than anything we'd see in the US. What that moment looks like from the outside — and what we learn from it — could be one of the most consequential events of the next few years in AI.
Why Are SpaceX and Google Building Space Data Centers?
Separate from the Beijing summit, SpaceX and Google are reportedly in talks for a rocket launch deal that would support orbital data centers — a genuinely novel infrastructure play that both companies are betting on ahead of SpaceX's anticipated IPO, which could be the largest in history. Google, which already owns 6.1% of SpaceX, announced its own Project Suncatcher initiative last year, aiming to have prototype satellites operational by 2027.
Sundar Pichai's take: "There's no doubt to me that a decade or so away, we'll be viewing it as a more normal way to build data centers." A decade is a long timeline in an industry that moves this fast, but the logic is straightforward — space offers land-free, regulation-light real estate with line-of-sight power options. The engineering challenges are enormous, but so is the potential upside if latency and energy problems can be solved.
Is US Inflation Getting Worse in 2025?
Rounding out the week's big stories: US producer prices rose 6% year-over-year in April, well above the 4.8% estimate. Consumer price data has also come in hot, pushing investor inflation expectations to multi-year highs. The break-even rate on TIPS versus ordinary treasuries — a key market signal — hit its highest level since October 2022, suggesting markets expect around 2.7% average annual inflation over the next five years.
The concerning backdrop here is that the AI and data center boom is driving GDP numbers that look healthy in aggregate, but the real economy — the part most Americans actually live in — grew at just 0.1%. Layering 2.7% inflation on top of stagnant real growth is the classic recipe for stagflation, a combination that's historically very difficult for any administration to manage. Jerome Powell navigated a soft landing with extraordinary precision. Whoever sits in that chair next will have a completely different set of challenges — including a Middle East conflict that's already tightening energy markets and supply chains in ways that show up at the register.
The Trump-Xi summit, the Mythos expansion to Japan, the space data center race, and spiking inflation aren't separate stories. They're all facets of the same underlying reality: the world is in the middle of a technological and geopolitical transition that nobody fully controls, and the decisions being made this week will echo for a long time.
