Iran-war risk is escalating. Oil markets are unstable. Bond yields are volatile. And yet — tech stocks rose this week, Nvidia jumped, and indexes closed near all-time highs. The reason is singular: AI optimism is currently carrying global market sentiment.

This is no longer a sector story. It’s a market structure story. And it has implications that extend far beyond whether Nvidia hits $6 trillion.

The Numbers Are Extraordinary

Nvidia reported $81.6 billion in Q1 revenue — up 85% year-over-year — and guided Q2 to $91 billion. The stock barely moved. Not because the numbers were disappointing, but because the market had already priced in this level of growth. When 85% revenue growth at $80 billion+ scale is “expected,” something structural has changed in how markets value AI.

Nvidia’s market cap sits at $5.23 trillion — the most valuable company on Earth. Alphabet is at $4.63 trillion. Apple at $4.53 trillion. Microsoft at $3.11 trillion. The top four companies by market cap are all AI infrastructure — as explored in the economics of AI compute infrastructure — or AI platform companies. The combined market cap of the top 5 AI-exposed companies exceeds $20 trillion.

AI Is the Market’s Shock Absorber

What’s remarkable is not that AI stocks are rising. It’s that AI is absorbing macro shocks that would normally trigger broad selloffs. Geopolitical risk in the Middle East, oil price spikes, interest rate uncertainty — each of these has historically caused 5-10% market corrections. In 2026, they’re being offset by AI capital expenditure commitments.

The mechanism: hyperscalers (Microsoft, Google, Amazon, Meta) have committed $300+ billion in combined AI capex for 2026. These are contractual obligations — data center leases, GPU purchase agreements, power contracts — that cannot be unwound quarter-to-quarter. The spending continues regardless of whether Iran escalates or oil hits $120. This creates a floor under the tech sector that macro volatility can’t easily break.

The result: AI capex has become a structural support for the entire equity market. When Nvidia, TSMC, and Broadcom report strong earnings driven by AI infrastructure demand, it lifts the S&P 500 because these companies represent an outsized share of the index. The market is, in a very literal sense, riding on AI spending commitments.

The Three IPOs Will Test This

SpaceX-xAI pricing on June 12 at $1.75 trillion. Anthropic filing at $965 billion. OpenAI targeting $852 billion. If public markets absorb $3.5+ trillion in AI listings while maintaining current index levels, it confirms that AI is not a bubble — it’s a structural reallocation of global capital toward the technology layer.

If the IPOs struggle — if SpaceX-xAI prices below target, if Anthropic delays, if OpenAI’s filing reveals unit economics that don’t justify the valuation — the entire market narrative shifts. The $300 billion in AI capex commitments don’t disappear, but the premium the market places on AI exposure would compress. And since AI exposure is what’s holding the indexes up, that compression would ripple far beyond the tech sector.

The Bigger Picture

We are in a moment where a single technology cycle — AI infrastructure buildout — is simultaneously the largest driver of corporate capital expenditure, the largest driver of equity market performance, and the largest driver of new company creation since the internet. The last time a single technology carried this much weight across all three dimensions was 1999.

The difference: in 1999, the spending was speculative — companies were building for demand that hadn’t materialized. In 2026, the spending is contractual — Nvidia’s $91 billion Q2 guidance is backed by purchase orders, not projections. Microsoft’s $80 billion capex is committed to data center leases already signed. Meta’s $145 billion is financing AI infrastructure that’s already generating ad revenue.

That doesn’t make AI stocks cheap. Nvidia at 25x forward revenue, Anthropic at 80-100x revenue, OpenAI at 50-60x revenue — these multiples require sustained hypergrowth to justify. But it means the AI trade is grounded in real spending in a way that the dot-com boom was not. The question is not whether the spending is real. It’s whether the spending generates returns that justify the valuations built on top of it.

For now, public markets are answering that question with a resounding yes. The next six months — three mega-IPOs, a potential geopolitical crisis, and the first full year of post-Blackwell earnings — will determine whether that answer holds.

For the full structural map of the AI economy, read The Map of AI Redrawn on Business Engineer.