Behind the video
Why the AI Trade Is Crashing
Overall Thesis
The host attributes the AI-led market selloff to doubts about returns on enormous infrastructure spending and increasingly expensive financing. He expects scrutiny of AI spending to persist into 2027, emphasizing neocloud funding risks and potential pressure on supplier valuations without making an unambiguous asset-price forecast.
Narratives
The host argues that the index's substantial exposure to the AI trade makes it vulnerable to questions about infrastructure spending returns. He expects those questions to persist through the remainder of 2026 and potentially into 2027, but does not establish a definite index-price direction over that period.
Key Arguments
- The host estimates roughly half of the index is related to the AI trade.
- AI exposure extends across semiconductors, memory, energy and hyperscalers.
- Expensive financing and uncertain economic returns threaten the sustainability of spending.
- The host favors cash-generating businesses with long-term bargaining power.
Risks acknowledged
- Stocks need not fall every day.
- Established hyperscalers can pull back investment.
- AI can benefit companies outside the infrastructure trade.
NBISNebius GroupThe host identifies Nebius as a risky neocloud because convertible financing can require repayment if equity conversion fails. He interprets discounted convertible debt as evidence that investors demand greater compensation for AI financing risk.
Key Arguments
- Recently issued convertible debt is described as trading around 85 cents on the dollar.
- Failure to convert debt into equity creates a cash repayment obligation.
- The host frames the debt pricing as a warning about capital access.
Risks acknowledged
- Conversion following a higher share price could allow attractive equity financing.
METAMeta PlatformsMeta illustrates the host's concern that even major hyperscalers face funding pressure from AI infrastructure spending. He nevertheless sees established businesses as more resilient than neoclouds because they can reduce investment.
Key Arguments
- The host says Meta entered negative free cash flow in the third quarter of 2026.
- Rising credit insurance costs are cited as a broader financing warning.
Risks acknowledged
- Core business cash generation supports resilience.
- Management can pull back capital spending.
NVDANvidiaThe host raises the possibility that slower hyperscaler spending could weaken Nvidia's revenue growth, margins and valuation multiple. This is a conditional valuation-risk discussion rather than an explicit forecast of Nvidia's share-price direction.
Key Arguments
- Supplier demand depends on customers maintaining AI investment.
- Slower revenue growth could lead to margin contraction.
- Combined growth and margin pressure could compress valuation multiples.
Risks acknowledged
- Valuation pressure would not necessarily mean terrible operations.
- The business could remain profitable.
MUMicron TechnologyMicron is presented as exposed to a potential moderation in AI-related memory demand. The host outlines possible revenue, margin and valuation pressure rather than directly predicting a stock decline.
Key Arguments
- The host questions whether memory demand could weaken.
- Slower supplier growth could pressure margins and valuation multiples.
Risks acknowledged
- Valuation compression does not require operational failure or losses.
CRWVCoreWeaveCoreWeave is presented as especially exposed to rising financing costs because infrastructure investment is its core business. The host questions whether available cash, customer prepayments and future returns can support its spending commitments.
Key Arguments
- The host cites a roughly 14.6% yield on its 2031 debt, recently exceeding 15%, as a financial stress signal.
- Expected capital expenditures substantially exceed the cited cash balance.
- Existing debt and higher borrowing costs complicate additional fundraising.
- Customer prepayments may become less reliable if compute availability improves.
Risks acknowledged
- The company has raised capital since the cited quarter.
- Customer prepayments provide financing.
- Infrastructure assets could generate high returns if contracts and renewals perform as anticipated.
GOOGLAlphabetAlphabet is the host's only disclosed holding directly associated with the AI trade. He highlights financing pressure from AI investment while arguing that large hyperscalers retain flexibility through their established operations.
Key Arguments
- The host says Alphabet crossed into negative free cash flow in the second quarter of 2026.
- Rising credit default swaps are presented as evidence of increasing financing costs.
- The host discloses ownership without making a new directional investment call.
Risks acknowledged
- Core operations generate cash.
- Capital spending can be reduced.
AMZNAmazonAmazon is discussed as another hyperscaler whose infrastructure spending creates external financing needs. The host distinguishes that pressure from the greater vulnerability of companies whose entire business depends on continued AI buildout.
Key Arguments
- The host describes Amazon as having negative free cash flow for some time.
- Increasing credit default swap costs signal more expensive capital.
Risks acknowledged
- Established operations generate cash.
- Infrastructure spending can be reduced if necessary.
AMDAdvanced Micro DevicesAMD is included among suppliers exposed to a potential slowdown in AI infrastructure demand. The host discusses possible growth, margin and valuation compression without making an explicit absolute share-price forecast.
Key Arguments
- Reduced customer spending could slow revenue growth.
- Margin contraction could accompany weaker growth.
- Lower growth and margins could lead to multiple contraction.
- AMD is also included in the discussion of rising credit default swaps.
Risks acknowledged
- The company could retain sound operations and positive net income.
TSMC is included among semiconductor suppliers that could face slower growth if AI spending weakens. The host discusses a possible chain from slower growth to margin and multiple contraction, without an explicit asset-price prediction.
Key Arguments
- Supplier growth depends on sustained infrastructure demand.
- Revenue and margin pressure could change market valuations.
Risks acknowledged
- Operations could remain healthy and profitable despite valuation pressure.
Hedges & Caveats
- Large hyperscalers can reduce capital spending and rely on cash generated by their core businesses.
- Potential supplier valuation pressure does not necessarily imply poor operations or negative net income.
- Stocks are not expected to fall every day.
- Continued scrutiny through 2027 is expressed tentatively.
- The host owns Alphabet and says his other holdings can benefit from AI without being central to the infrastructure trade.
- CoreWeave customer prepayments and infrastructure returns could help fund investment, although their durability is questioned.