Jesse Felder argues the stock market is in a prolonged topping process, with the AI/semiconductor boom now nearing an inflection point as data-center delays and cancellations collide with extreme valuation expectations. He also sees stagflation risks rising from war-driven energy shocks, higher inflation, and weaker hiring, which could force the Fed to tighten into a slowing economy. His preferred positioning is defensive diversification, with emphasis on commodities/real assets and especially energy, while he thinks precious metals may need to cool before offering a better entry.
Watch on YouTube ›Get the market thesis, key claims, assets, contradictions, and follow-up questions from any financial video — then unlock a version personalized to your portfolio, watchlist, and favorite speakers.
Jesse Felder’s core thesis is that the current equity bull market is in its late-stage topping process and that the AI trade is beginning to transition from boom to bust. He argues the market is pricing in a future of near-certain, very high AI-related revenue growth, but the physical buildout required to support that path is running into hard limits: data centers have been delayed for much of the year, and cancellations are now starting to appear. In his view, that disconnect between analyst forecasts and construction reality is the key fault line in the AI story. He builds that case by pointing to several linked mechanisms. First, semiconductor and hyperscaler earnings look artificially strong because AI capex creates an accounting timing mismatch: revenues are recognized now while depreciation and related expenses show up later. …
Tactically, the AI/semiconductor trade looks vulnerable now: watch for project cancellations, analyst cuts, and sharper tech volatility. Energy and broader real assets are the cleaner near-term hedge if inflation and geopolitics keep pressure on rates.
Over the next few months, the base case is lower AI growth expectations, weaker hyperscaler margin sentiment, and more rotation out of crowded tech leadership. Confirmation would come from further buildout delays, softer demand signals, and continued firmness in commodities.
Structurally, Felder sees a transition from tech-led concentration toward a more inflation- and scarcity-driven market regime. If he is right, real assets and diversifiers matter more while monopoly-like growth stories face a much higher bar to sustain valuations.
The stock market is in a prolonged topping process and the AI trade may already be transitioning into a bust.
He repeatedly says the bull market is in its final innings and that AI is metastasizing into a bust.
Data-center delays and cancellations mean the AI revenue outlook is running into physical constraints.
He argues there is no way the announced buildout pace can be met, so sales estimates must fall.
AI earnings strength is partly a timing and accounting effect from capex, not pure underlying profitability.
He says revenues are recognized immediately while depreciation and expenses show up later.
What is your current macro and market assessment given stocks near all-time highs, the Iran war, and massive AI-related capital spending?
Jesse believes we're in a prolonged topping process in the stock market, with semiconductors in a final blowoff phase. He sees earnings growth as largely a timing artifact — revenues from the AI buildout are recognized immediately while expenses (depreciation) are deferred. Meanwhile, data centers that chips are destined for are being delayed or cancelled, with two-thirds of centers supposed to be completed next year not even broken ground yet. He also sees stagflationary dynamics building, with small businesses raising prices and reducing hiring, partly driven by the Iran war.
What would it take for you to change your mind and believe this market move is sustainable?
Jesse would need to see the physics of the data center buildout change miraculously — that the ability to build announced data centers rapidly was actually possible. He cites Satya Nadella saying last year they could buy all the chips they want but lack the 'warm shells' (data centers) to plug them into, and that constraint has only worsened. Without this change, semiconductor revenue growth can't be maintained and could reverse.
On top of the physics constraint of building data centers, what about the political backlash against data centers, such as the moratorium in Reno?
Unlock the full claims, asset map, scores, related transcripts, follow-up questions, and AI chat — shaped around your portfolio, watchlist, favorite speakers, and risks.
