Introduces CoreWeave as a direct play on AI infrastructure demand, up 150% since IPO.
CNBC Anchor
Mike Intrator
Accepts the five-word take on AI: 'violent change in violent demand' and says it's not circular.
Asks for elaboration on the AI demand thesis.
CNBC Anchor
Mike Intrator
Describes sustained demand across the entire AI ecosystem (data centers, silicon, applications) from enterprise, cloud, and retail clients trying to access infrastructure at a scale that doesn't yet exist.
Notes that clients are large, well-heeled companies, which differentiates this cycle, but raises concern about staggering capital requirements and potential circularity of deals.
CNBC Anchor
Mike Intrator
Explains CoreWeave serves the entire AI ecosystem, from Microsoft/Meta/NVIDIA to AI labs and student startups, providing the software and physical infrastructure they need.
Asks about the long-term payoff, comparing the AI build-out to railroads, and whether the ROI will be 2-3x.
CNBC Anchor
Mike Intrator
States that in 5-10 years, AI will be embedded in everything and will pay dividends for the next 100 years.
Outlines three key risks: 1) Some companies (independents, not giants) run out of capital. 2) A tech breakthrough (like DeepSeek on steroids) reduces compute needs. 3) Chip depreciation schedules are shorter than assumed, requiring constant updates.
CNBC Anchor
Mike Intrator
On Risk 1 (capital): Acknowledges some companies will fail (Pets.com analogy), but this is natural for a new industry and won't take down the overall market. CoreWeave manages risk by selling across startups, clouds, and enterprises.
Mike Intrator
On Risk 2 (tech breakthrough): Says step-function efficiency gains are expected in this business (DeepSeek wasn't a shock).
Mike Intrator
On Risk 3 (depreciation): Argues that long-term (5-6 year) client contracts define the real depreciation/obsolescence curve, not shorts' opinions.