Jim Chanos & Val Zlatev: Long and Short Alpha in AI, Semiconductors, Data Centers, Neoclouds, and Data Centers | MacroMinds Symposium 2026 – Monetary Matters with Jack Farley
Jim Chanos & Val Zlatev: Long and Short Alpha in AI, Semiconductors, Data Centers, Neoclouds, and Data Centers | MacroMinds Symposium 2026
Bullish and bearish opinions expressed in this episode, paired with supporting transcript quotes. The quote confirms what was said—not whether the opinion is correct.
Bullish
$NVDA— Chanos is net long AI with NVIDIA being a key beneficiary of the AI infrastructure build-out, trading at reasonable valuations (15x 2027 earnings) compared to competitive CPU companies. The company benefits from oligopolistic market position versus competitive markets.
$AVGO— Broadcom trades at attractive valuation (12x 2028 earnings) despite being in an oligopolistic market position with strong competitive advantages, similar to NVIDIA's positioning.
$MEMORY— Memory companies (DRAM/NAND) are trading at extremely low valuations (6-7x forward multiples) while facing supply constraints that limit capacity growth to 30-35% annually. Demand from AI is driving 4-5x price increases with peak demand expected to last 2-4 years.
$SEMIS— Broad semiconductor exposure remains attractive as the sector benefits from AI infrastructure build-out, with valuations not at bubble levels like 1999-2000. Physical constraints on supply growth support pricing power.
$PHOTONICS— Val identified potential opportunities in the optical photonics space as part of the AI infrastructure build-out.
Bearish
$NEOCLOUDS— Neo-clouds (CoreWeave, Lambda, etc.) are equipment leasing companies with single-digit ROICs (5-8%) even in peak demand conditions. They lack technology differentiation and face competition from hyperscalers who can buy chips directly. Returns will compress as power and labor bottlenecks ease.
$BTC— Bitcoin miners are identified as inherently unprofitable business models attached to the AI ecosystem with de minimis returns on capital even under best-case assumptions.
$SPACEX— SpaceX's $200B valuation cannot be justified by current business fundamentals. Launch business still loses money, Starlink earns modest returns on $25-30B invested capital, and xAI is a cash sinkhole. Starship hasn't achieved Earth orbit in 12 flights. Space data centers face fundamental challenges including radiation, maintenance costs, and redundancy issues.
$META— Meta's AI strategy makes no sense according to Jack, who considers himself a bear on the stock despite acknowledging this view is becoming consensus.
$PCMAKERS— PC and smartphone component makers face unit declines of mid-teens as memory costs rise from 20% to 50% of bill-of-materials. These companies lack pricing power to pass costs to consumers in elastic markets, creating shorting opportunities.
$INTC— Intel trades at elevated valuations despite being in a competitive market and having lost its monopoly position, with the company not making money for a couple of years. Valuation disparity versus oligopolistic players like NVIDIA is extreme.
$SEMIEQPT— Semiconductor equipment companies (ASML, Lam Research, Applied Materials) trade at expensive valuations (35x forward multiples) despite growth being capped at 30% annually due to physical constraints. More expensive relative to chip makers with higher margins.