Jordan Schneider: US-China Beijing summit was 'prestige on the cheap,' rare earth stalemate has years to run, and Taiwan remains critical beyond chips

Jordan Schneider on the Beijing summit, rare earths, and Taiwan

Jordan Schneider, founder of ChinaTalk, sees the May 2025 US-China Beijing summit as largely theatrical. His headline: "prestige on the cheap." The full cabinet flew over, every major American CEO made the trip, and the result was atmospherics with no substantive movement. The charitable read is that giving China the optics of a high-profile visit is low-cost and may reduce the risk of reckless escalation — the same argument made about not publicly humiliating Putin. But Schneider doesn't think much more came of it than that.

The rare earth stalemate

The backdrop is a shift in leverage that crystallised in April 2025. The US had spent roughly eight years — across Trump, Biden, and into the second Trump term — steadily turning up pressure through tariffs, export controls, and investment restrictions, operating on the assumption that China would absorb the pain. Then Beijing punched back with rare earth restrictions, and the experiment confirmed that China holds real coercive leverage over the American economy. Both sides now recognise the stalemate, which explains the summit's content-free pageantry.

Schneider doesn't expect the rare earth leverage gap to close quickly. Reports he's reviewed suggest the imbalance persists five or more years out. The challenge is structural: decoupling from the world's second-largest economy to the point where it can no longer squeeze you coercively is a generational project, not a policy sprint. The more tractable near-term goal is protecting specific critical inputs — specialised fighter jet components, insulin supply chains — rather than achieving broad insulation.

“Both sides have kind of realized that [rare earth leverage exists] and what we got is prestige on the cheap... The percentage of chips that are going to be manufactured in Taiwan eighteen months from now is still gonna be like 85% or 90%... China, I think, is deeply skeptical of these [AI safety] sorts of arguments on face because they're kind of sci fi.”

AI and strategic stability

The absence of AI lab leaders from the delegation is notable but probably not surprising. The summit was framed as a trade visit, and the US government's own domestic AI regulatory posture remains unresolved — Trump rescinded an AI safety executive order the same week. Schneider's argument: you can't expect meaningful US-China AI safety dialogue when Washington hasn't figured out what it wants domestically.

He's also sceptical that China's leadership is ready to engage seriously on AI existential risk. The Chinese Communist Party is hardware-minded and views these arguments as science fiction. Historical precedent suggests real strategic-stability talks between adversaries require a crisis of Cuban Missile Crisis magnitude to concentrate minds. An Anthropic paper flagging Claude as dangerous doesn't come close.

Taiwan

Schneider won't engage with Chamath Palihapitiya's claim that Taiwan becomes geopolitically irrelevant within 18 months, beyond refusing to acknowledge it. The chip dependency alone makes that prediction hard to square: roughly 85–90% of leading-edge chips will still be manufactured in Taiwan eighteen months from now, and a disruption in 2028 would still constitute a global economic catastrophe.

Dan Wang's point — that Western technologists over-index on the chip angle — also lands. Taiwan matters to the broader Pacific architecture the US has built over 75 years. If Taiwan falls under PLA control, the first island chain that anchors American alliances across both Southeast and North Asia becomes untenable.

On whether accelerating AI progress makes a Taiwan invasion more likely, Schneider's answer is that AI also improves defensive capabilities, and that Xi Jinping has watched Putin spend five years in a war he couldn't end. The lesson about starting something you don't know how to finish is one Beijing has had to absorb in real time.

Humanoids and hardware asymmetry

The one area where China's manufacturing depth creates a durable edge is physical AI. Western humanoid companies likely have better software today. But once humanoid robots become economically useful, China's ability to manufacture them at orders-of-magnitude greater scale could make the software gap irrelevant. Schneider frames this as the inverse of the chip story: in semiconductors, the US has compute advantage as long as Taiwan holds; in embodied AI, the hardware that deploys the model will almost certainly be made in China at a scale the West can't match near-term.