Dan Ives Says the AI Trade Is Only in the "Third Inning." Should Investors Stay Long AI Stocks?

Dan Ives’ “third inning” framing argues the AI build-out is shifting from GPU capex to monetization across cloud, networking, software, and security. He highlights CoreWeave and Cisco as beneficiaries as spending spreads beyond chips into data-center “plumbing,” and cites examples like MSFT Copilot, Meta AI agents, and PLTR AIP deployments showing AI moving up the stack toward recurring revenue. He estimates Nvidia spend could translate to ~$8–$10 of total AI spend across infrastructure layers, while noting risks from tighter macro conditions and regulatory scrutiny that could slow AI capex.

Analysis

The more important signal is not “AI is early,” but that the profit pool is migrating from one-time accelerator purchases to recurring spend on network, storage, orchestration, and application layers. That favors companies with usage-based revenue or installed base monetization power: MSFT and AMZN can turn infrastructure into sticky consumption, while CSCO benefits from data-center interconnect upgrade cycles that often lag GPU orders by 2-4 quarters. CRWV is the highest beta expression, but also the most financing-sensitive; if rates stay elevated, its growth multiple is more fragile than the operating leverage story suggests.

For the next 1-3 months, this reads more like a relative-value rotation than a fresh alpha signal. NVDA likely stays the center of gravity, but the next leg of multiple expansion may compress as the market asks who captures durable software margin rather than who ships the picks and shovels. PLTR and META are better positioned than generic “AI” names because both can point to monetizable workflows, but execution risk matters: the market will punish any evidence that AI features raise cost before they lift ARPU.

Contrarian view: consensus may be underestimating how much of the current AI basket is already priced for perfect capex continuation. If enterprise budgets pause, the weakest links are CRWV and other financing-dependent infrastructure names, then NVDA on sentiment, not fundamentals. The structural bull case remains intact over 6-18 months, but the trade likely shifts from chasing the original compute leader to owning monetizers and avoiding low-quality AI wrappers.

Key Decisions for Investors

• Pair trade: long MSFT / AMZN, short NVDA for 1-3 months if you want AI exposure with less multiple risk; thesis is monetization visibility versus crowded compute ownership. Falsify if NVDA re-accelerates backlog commentary or MSFT/AMZN cloud growth slows.
• Add CSCO on pullbacks for a 2-4 quarter trade; networking usually benefits after GPU deployment and can re-rate if AI-related orders show up in guidance. Risk: if hyperscalers delay networking refresh, the trade stalls.
• Treat CRWV as a tactical momentum name only, not a core long. Use smaller size or call spreads rather than outright equity given higher sensitivity to rates, financing, and utilization. Falsify on any evidence of slower customer ramp or tightening capital markets.
• Maintain a watchlist long in PLTR or META only on confirmed commercial traction; do not chase on AI branding alone. The upside is operating leverage if AI lifts seat/usage monetization, but the stock will de-rate quickly if revenue conversion lags.
• No actionable read-through for NFLX or GETY from this setup; stay out unless AI spend begins to affect ad-tech or content economics directly.