Intel Just Broke a 26-Year Curse. Here's What Falls Next.

In its most recent fiscal year (ended January 2026), Nvidia invested $17.5 billion in private AI companies and infrastructure funds. It also provided $3.5 billion in land, power, and shell guarantees to early-stage companies — essentially backstopping the physical infrastructure its own customers need to buy its chips.

That's not a chip company. That's a quasi-sovereign infrastructure bank. Nvidia is financing the demand side of its own income statement: fund the startups, guarantee their data center buildouts, then sell them GPUs. The flywheel is self-reinforcing, and it's backed by over $60 billion in cash and marketable securities.

When one company controls both the supply of critical hardware and the financing of the ecosystem that buys it, the usual competitive dynamics don't apply. Nvidia isn't just riding the AI wave — it's engineering the tide.

Intel's foundry push puts it in a weird spot. It competes with TSMC while also pitching itself as a manufacturing partner to AMD, Qualcomm, and other chip designers that rely on Taiwanese factories today. If this earnings beat speeds up talks with chip designers looking for a non-TSMC option to reduce geopolitical risk, that's the real domino.

AMD and Arm both rallied on April 24 alongside Intel, as the broader semiconductor sector advanced. But the rotation story goes deeper than sympathy trades. Every fabless company watching the U.S.-China tension over Taiwan has a strategic reason to want Intel's foundry to succeed — it's the only credible Western alternative to TSMC for leading-edge manufacturing.

A 26-year ceiling turning into a floor matters less for Intel's stock chart than for the strategic calculus of every chip designer weighing supply chain risk.

Nvidia deployed $17.5 billion into private AI companies last fiscal year — the same companies that turn around and buy Nvidia's chips. That creates a self-reinforcing loop. Nvidia's chip revenue justifies its $5 trillion market cap. That validates the private valuations of companies Nvidia has invested in. And those valuations support the story that AI demand is broad-based and here to stay.

The conflict is subtle but real. Nvidia has a financial incentive to keep private AI valuations high, because those marks reinforce the growth story that supports its own public multiple. It's simultaneously the arms dealer and the venture capitalist — and both roles benefit from the same narrative.

None of this means the AI spending cycle is fake. But it means the feedback loop between public and private valuations has a single point of failure: Nvidia's continued dominance. If that cracks, the entire valuation chain unwinds in both directions.

Nvidia's capital expenditures nearly doubled to $6.1 billion from $3.4 billion the year before — spending that flows directly to foundry partners. Meanwhile, Intel is mounting its own foundry revival and competing for some of the same leading-edge process nodes. When several big customers all rush to book 3nm and 2nm slots at once, TSMC's choices about who gets capacity become the real bottleneck.

Who gets first dibs on the next generation of chip nodes isn't just a factory question. It's where competitive advantage is quietly handed out. A chip designer locked out of leading-edge capacity for even one product cycle falls behind in ways that take years to recover from.

Historically, foundries have shown more stable returns than fabless chip designers during cycles of disruption in AI architecture. TSMC's wide range of customers shields it from risks tied to any single chip design. But the real edge isn't diversification — it's the power to decide who ships first.

With $102.7 billion in operating cash flow, Nvidia had room to fund buybacks, invest in the ecosystem, and maintain a fortress balance sheet. The result: net income per diluted share hit $4.90 for fiscal year 2026, up 67% year-over-year — two percentage points faster than revenue growth. That gap is the buyback at work.

But here's the tension most investors miss. At $40.4 billion in buybacks against $102.7 billion in operating cash flow, Nvidia is returning nearly 40% of its cash to shareholders rather than reinvesting it. That's a telling choice about where to spend cash. Management is saying that buying back shares beats reinvesting every dollar into the business on a risk-adjusted basis.

For a company supposedly in the early innings of an AI supercycle, that's a subtle bearish signal hiding inside a bullish headline. If the growth opportunity were truly limitless, you'd expect the buyback to be smaller and the reinvestment to be larger.