Former NSC official Chris McGuire on the US-China tech freeze, chip export controls, and the humanoid robot national security gap

Summary

The US-China technology standoff entered a new phase on October 30, 2025, with a freeze-for-freeze agreement reached at a summit in South Korea. Washington suspended its affiliates rule — a regulation closing a major export control loophole that allowed exports to unlisted subsidiaries of blacklisted Chinese entities — while Beijing paused its retaliatory rare earth export restrictions. The deal locks in the status quo and resolves peripheral flashpoints including Chinese soybean purchases, but produces no structural breakthrough. Critically, there was no agreement to ship Blackwell-generation chips to China, and a proposed Chinese offer to build BYD and CATL factories on US soil was rejected.

Chris McGuire, former Deputy Senior Director for Technology and National Security on the National Security Council and senior State Department official through multiple administrations, characterizes the agreement as a tactical deescalation that tactically favors China. Beijing is buying time to circumvent export controls and close the technology gap, and every day the US pauses enforcement, China makes incremental progress. McGuire argues the Chinese rare earth response was disproportionate to the affiliates rule fix — a signal that Beijing will go "over the top" in response to any further technology restrictions — and that the move reflects a deliberate strategy of aggressive deterrence rather than genuine capability.

Chips and the Semiconductor Leverage Gap

On semiconductors, McGuire is direct: China cannot come close to meeting domestic AI demand with indigenously produced chips. He estimates Chinese seven-nanometer production capacity at roughly one twenty-fifth to one one-hundredth of US capacity, with no meaningful production at five nanometers or below. Any posture of rejecting US chips entirely would self-hobble China's AI industry. The public Chinese stance of not wanting NVIDIA products is, in his assessment, a negotiating tactic — likely an attempt to trade H20 access for Blackwell access, which did not succeed.

The more consequential near-term legislative development is the GAIN AI Act, sponsored by Senator Banks (R-Indiana), passed by the Senate as part of the NDAA and currently in House conference. The bill does not take a position on whether chips should be exported to China, but requires American buyers to receive right of first refusal before any controlled chips are authorized for Chinese buyers. A recent provision also exempts certified trusted US persons from export license requirements for chip shipments outside China, provided they retain at least 50% of computing capacity in the United States — a provision designed to unblock pending multi-billion dollar US data center deals in the Middle East that have been stalled by licensing requirements. NVIDIA has publicly opposed the bill, with McGuire characterizing their concern as regulatory complexity and operational constraint, though he argues a supply-constrained market makes US priority access a logical policy position regardless of one's view on China exports.

Rare Earths: Known Risk, No Action Taken

Washington's vulnerability in rare earths was not a surprise — Chinese dominance was widely understood in policy circles for years. The failure was structural: the kind of coordinated, heavily resourced supply chain diversification required is precisely what the US government finds hardest to execute absent a crisis. McGuire draws a direct parallel to Operation Warp Speed, noting the US can move quickly in emergencies but typically requires a visible crisis to mobilize. He expects Capitol Hill legislation to resource rare earth alternatives, though the mechanism — equity investment, loans, Defense Production Act authorities — remains unresolved. The surprise was not that China held the card, but the scale at which they played it.

Humanoid Robots: The Unaddressed Gap

McGuire identifies humanoid robotics as the most urgent unaddressed national security exposure in the consumer technology space. Unitree robots are currently available on Walmart.com. The concern is structural: any Chinese-made robot with mapping capability, audio collection, and physical appendages deployed at scale in US homes or factories represents a potential remote-access vulnerability. He draws a direct analogy to connected electric vehicles — the Biden administration effectively banned Chinese EVs before significant market penetration occurred, supported by domestic industry lobbying from Detroit and a near-zero Chinese market share at the time of the decision.

The window for a clean regulatory intervention in humanoids is closing. The DJI precedent illustrates the cost of delay — drone restrictions are now deeply complicated by existing government and civilian dependency, requiring mitigation steps even if a ban were implemented today. A class-based prohibition on Chinese-origin robots meeting certain capability thresholds — mapping, audio, physical manipulation — is McGuire's preferred approach, modeled on the vehicle framework. He is explicit that without protection, Chinese manufacturers will sustain losses to capture market share long enough to prevent US companies from reaching the scale needed to compete, replicating the DJI and TikTok dynamic in a significantly more consequential hardware category.

Strategic Framing

McGuire frames the broader US-China technology contest as a shift that began in the first Trump administration around 2015 to 2017, was substantively deepened by the Biden administration in semiconductors and AI — with a conscious decision to separate supply chains and maintain maximum lead time in advanced chips — and has now entered a more volatile phase under the second Trump administration. Xi, in McGuire's read, is managing for a generational timeline with Taiwan as the ultimate objective, while Trump is optimizing over a shorter political horizon. The freeze deal reflects both sides pulling back from an escalation ceiling, but the underlying dynamic remains asymmetric: the status quo benefits China's catch-up trajectory more than it benefits US technology leadership.