Sector Intelligence

The Trump-Xi Summit: What It Means for AI Stocks

Rgs 22, 2026

By Ieva Bružaitė | Founder, Orion Capital | 22 September 2026

Markets are approaching another Trump-Xi summit with a familiar combination: semiconductor momentum, renewed AI optimism and growing hope that diplomacy between Washington and Beijing could reduce some of the geopolitical risk surrounding global technology markets.

But after the experience of May, the most useful question is not whether the September summit will be “good” or “bad” for equities. It is this:

How much of a constructive outcome is already priced in – and what would actually surprise the market?

That distinction may matter considerably for Nvidia, AMD, Broadcom, Micron and the broader semiconductor complex. And, as the evidence below suggests, the answer this time may be different from May’s.

The May Precedent

Donald Trump and Xi Jinping last met in Beijing during Trump’s 13-15 May state visit. Markets entered the Trump-Xi summit with considerable optimism, and on 14 May that optimism appeared to be rewarded: Reuters reported that Washington had cleared sales of Nvidia’s H200 AI chips to roughly ten Chinese companies, including Alibaba, Tencent, ByteDance and JD.com. Nvidia closed at a record $235.74 that day.

Then the detail arrived. Washington had issued the licences – but Beijing had not approved the purchases. Speaking aboard Air Force One on 15 May, Trump said Chinese buyers “choose not to buy because they want to develop their own,” and US Trade Representative Jamieson Greer indicated that semiconductor controls had not been on the formal bilateral agenda. The breakthrough investors had priced was, in practice, a permission slip without a customer.

On 15 May, Nvidia fell 4.4% and AMD declined 5.7%, while Micron lost 6.6% and Intel more than 6%. The Nasdaq Composite fell 1.5%. It is worth being precise about the cause: the sell-off also coincided with rising Treasury yields – the 30-year had crossed 5% that week – and with profit-taking after an exceptionally sharp run in chip stocks. The summit’s outcome was the trigger, not the only force.

Expectation → Rally → Summit → Reassessment

It was not a change in the underlying AI investment thesis. It was a repricing of expectations that had moved ahead of the policy outcome – and, critically, ahead of Beijing’s own willingness to buy.

September Starts With AI Momentum Already Elevated

This week’s summit arrives after another powerful move in AI-related equities. On Monday, 21 September, the Nasdaq Composite rose 2.3% to a record close of 27,122.09 – its first record finish since June – while the S&P 500 gained 1.5% and the Philadelphia Semiconductor Index climbed more than 4%.

AMD’s move was particularly striking: the stock rose nearly 10% to an all-time high above $610, briefly pushing its market value above $1 trillion for the first time. Intel and Arm posted double-digit gains, while Nvidia, Broadcom and Micron each rose between roughly 1.5% and 3%.

When prices have already moved sharply before a catalyst, the question becomes less about whether the news is positive and more about whether it is more positive than the market already assumes. This is where September becomes interesting – because the composition of Monday’s rally tells a different story from May’s.

What Is Different This Time

In May, the stock that led the rally into the Trump-Xi summit was Nvidia – the company with the most direct exposure to Chinese AI-chip access. The market was explicitly pricing a China deal.

In September, the leadership looks different. Monday’s biggest gains came from AMD, Intel and Arm, driven primarily by enthusiasm around Meta’s new Muse AI agent and the server-CPU demand that agentic AI may require, alongside falling oil prices and easing Treasury yields. Summit optimism was one supporting factor – Treasury Secretary Scott Bessent described his weekend talks with Vice Premier He Lifeng as very successful – but it was not the headline driver. Nvidia rose a more modest 2.3% and remains below its May record.

	Into the May summit	Into the September summit
Leading stocks	Nvidia – the most China-exposed name – hit a record close of $235.74 on 14 May	AMD (+~10%), Intel (+~12%) and Arm led on 21 September; Nvidia rose 2.3% and remains below its May record
Stated catalyst	H200 export licences for ~10 Chinese firms	Meta’s Muse AI agent, falling oil, easing yields – with summit optimism as one factor among several
What the rally implies	Investors priced China chip access directly	Investors are pricing agentic-AI compute demand more than a China chip deal
Rate backdrop	30-year Treasury yield crossing 5%	Fed hiked on 16 September; 10-year yield eased back from near 5% on 21 September

This matters for the question in this article’s title. If the current rally is anchored mainly in agentic-AI compute demand rather than China access, then a summit that produces no chip breakthrough may be less damaging to semiconductor prices than it was in May – because less of the China outcome appears to be priced. That is an inference, not a certainty, and positioning can shift quickly in a week. But the evidence so far suggests the market is not repeating May’s mistake in the same form.

What Washington and Beijing Are Actually Discussing

Xi Jinping will make a state visit to the United States from 23 to 25 September – his first state visit to Washington since 2015 – with the principal meeting with Trump expected on 24 September, followed by a state dinner that several leading technology executives are expected to attend.

The timing is not incidental. The one-year trade truce agreed in Busan in October 2025 is winding down, and negotiators have been working on a framework for reciprocal tariff reductions covering roughly $30 billion of trade on each side. The agenda also extends well beyond tariffs: artificial intelligence, technology restrictions, critical minerals, Taiwan and broader geopolitical stability have become central components of the relationship.

AI deserves particular attention. The two leaders discussed possible AI guardrails in May, and preliminary talks ahead of this visit – which China’s Xinhua described as candid, in-depth and constructive – reportedly included AI. That represents a potentially important area of cooperation between two countries otherwise competing aggressively for technological leadership.

But investors should distinguish between AI dialogue and semiconductor liberalisation. The two are not the same. Washington may be willing to discuss AI safety while maintaining restrictions on advanced computing technology. And Beijing, as May demonstrated, is pursuing technological self-sufficiency with growing determination: just this week, DeepSeek chief executive Liang Wenfeng called using more domestic chips for AI training a “major priority,” with Huawei deliveries expected as early as the fourth quarter.

For markets, diplomatic language may therefore matter less than the details – and the most important detail may now sit in Beijing rather than Washington.

The Semiconductor Question

For Nvidia and AMD, one of the most consequential issues is whether the summit changes expectations around access to the Chinese AI accelerator market. The size of the prize is real: KeyBanc has estimated that Chinese companies would like to purchase around 1.5 million H200 processors, equivalent to roughly $30 billion of revenue. But since the licences already exist, the binding constraint is now whether Beijing permits its companies to buy.

For Broadcom, the exposure is different, spanning networking, custom silicon and the wider AI infrastructure build-out. Micron introduces another strategic layer through memory and high-bandwidth memory. Further upstream, semiconductor equipment and manufacturing companies remain exposed to export-control policy and restrictions on advanced manufacturing technology.

The market often trades these companies collectively under the label “AI.” Policy does not. That means the reaction to the summit could be highly uneven across the semiconductor supply chain.

Three Possible Market Paths

1. Stabilisation without a major technology deal

A continuation or extension of the trade truce, constructive diplomatic language and no significant change to semiconductor access would reduce near-term geopolitical uncertainty. This may be close to what markets already expect. If so, the fundamental result could be positive while the equity reaction remains muted – though, given the evidence above, perhaps less sharply negative than in May.

2. A technology surprise

A genuine surprise would now require more than new US licences. It would require a signal that Beijing will allow its companies to buy – or another concrete concession on AI-chip market access or export controls. For companies with direct China exposure, such an outcome could alter revenue assumptions rather than merely improve sentiment.

A diplomatic headline changes sentiment. A policy change can change earnings.

3. Renewed friction

Semiconductor export controls, Taiwan, tariffs, critical minerals, supply-chain restrictions and US sanctions targeting China’s energy trade with Russia remain areas where strategic interests diverge sharply. A harder line from either side, or a failure to extend the truce, could rapidly restore the geopolitical risk premium that markets have recently been willing to compress.

Why May Should Not Be Treated as a Prediction

There is a temptation to look at May’s rally and subsequent semiconductor sell-off and assume that September must follow the same pattern. That would be a mistake. Historical analogues are useful because they identify market behaviour – not because they guarantee repetition.

The macro environment is different: the Federal Reserve has since begun raising rates. Valuations are different. Investor positioning is different, and, as shown above, the leadership of the rally itself is different. May therefore provides a framework, not a forecast. The relevant lesson is that markets can price diplomatic progress before the diplomacy actually produces it.

The Second Trade Hiding Behind the Trump-Xi Summit

There is another variable capable of overwhelming the geopolitical story entirely: US Treasury yields. May itself illustrates the point – the post-summit sell-off came alongside a jump in yields, making it difficult to separate the diplomatic disappointment from the duration effect.

Long-duration technology valuations remain highly sensitive to the discount rate, and that sensitivity has only increased since the Fed’s 16 September hike. Even a constructive Trump-Xi summit outcome could therefore coincide with falling semiconductor stocks if yields rise materially. The opposite is equally possible: an inconclusive summit combined with falling yields and stronger AI capital-expenditure expectations could allow technology stocks to keep climbing – much as falling yields helped power Monday’s record.

For investors, this creates two simultaneous trades: the geopolitical trade and the duration trade. Ignoring either one risks misreading the market’s reaction.

What Investors Should Watch on 24 September

Rather than focusing exclusively on the first headline from Washington, several details deserve attention:

Any signal that Beijing will approve Chinese purchases of licensed US AI chips – now the key variable for Nvidia
Language surrounding semiconductor export controls and any change to licensing policy
Extension or modification of the trade truce, and progress on the ~$30 billion reciprocal tariff framework
Commitments involving critical minerals and supply chains
Taiwan-related language
The reaction of US Treasury yields
Whether semiconductor stocks confirm or reject the diplomatic narrative – and whether China-exposed names behave differently from the CPU names that led Monday’s rally

Price action itself will provide information. If chip stocks absorb an inconclusive summit without surrendering their recent gains, it would suggest that the AI rally remains primarily anchored in earnings and capital-expenditure expectations. If apparently constructive headlines are followed by a sharp semiconductor reversal, the market may instead be signalling that optimism had moved too far ahead of policy – and that would look considerably more like May.

May vs September: The Question Behind the Chart

The most useful comparison may be a simple one: normalise Nvidia, AMD, Broadcom and the Nasdaq to 100 on the trading day before each summit, then compare their trajectories into and immediately after the event. The objective is not to claim that two charts must produce the same outcome. It is to ask whether investor behaviour is beginning to rhyme:

Rally into the event → Expectations rise → Policy arrives → Prices reveal what was already discounted

That is the experiment worth watching this week – and one we will revisit in our post-summit update.

The Orion View

The Trump-Xi summit is not inherently bullish or bearish for AI equities. Its significance depends on the gap between expectation and outcome.

May demonstrated how quickly semiconductor prices can incorporate expectations of diplomatic progress – and how quickly they adjust when the concrete outcome falls short. It also revealed that the decisive constraint on Nvidia’s China business sits increasingly in Beijing rather than Washington.

September presents another test, but on different terms. The structural AI investment story remains powerful, and this time the market’s enthusiasm appears anchored more in agentic-AI compute demand than in a China deal. Geopolitical optimism has entered the valuation equation again – but perhaps not as heavily as it did in May.

For investors, the key question after 24 September will therefore not simply be “Was the summit successful?” It will be: “What changed relative to what the market had already priced in?”

Because markets rarely trade the headline itself. They trade the distance between expectation and reality.

Post-Summit Update

Investments with a Soul.

Disclaimer: This material is provided for informational and research purposes only and does not constitute investment advice or a recommendation to buy or sell any security.

Sources

About the Author

Ieva Bružaitė is the Founder of Orion Capital, an independent research platform covering AI, semiconductors and long-term investing.

← Back to Orion Capital