The Terrifying Truth About Next 5 Years in the Stock Market | Erik YWR on “Reverse Crash” Risk
The conversation centers on a split market driven by hyperscaler AI capex and geopolitical supply risk. A cited capex surge (roughly $200B to $600–800B) has fueled semiconductor and memory rallies while private credit and exchanges are underwriting data‑center builds amid market volatility. Speakers warn of both upside scenarios — rerating to ~25x S&P or extreme Nvidia outcomes — and tangible downside risks, from a possible ~20% drawdown to an ~80% left‑tail collapse for individual names.
Preview · 13 more points in the app
Read the Full SummaryKey Takeaways
Speakers flag a multi‑year Persian Gulf disruption—citing “if 20% of the energy comes out of the Persian Gulf”—as a structural upside for US refiners, shipping and chemical/refining names (stronger margins, higher jet‑fuel/prices), even though consensus models forecast negative YoY energy earnings in 2027 if oil cools; sustained geopolitical risk could instead push oil materially higher for years.
Hyperscaler cloud revenues have accelerated (from ~15% to mid‑20%+), and capex surging from roughly $200B to $600–800B is driving a concentrated hardware/peripherals rally (Taiwan, South Korea, China) and large estimate revisions across the supply chain, with institutional private‑credit inflows (record Blue Owl inflows) materially financing the data‑center buildout despite public market volatility.
The semiconductor/memory boom—fueled by that hyperscaler capex—has produced extreme price and earnings action but carries structural cyclical risks (a potential “depreciation bomb” and capacity‑pause if customers defer builds); at the extreme tails Nvidia could see a peak PE near ~100 or an ~80% crash, so investors should size exposure carefully and many bullish views favor a NASDAQ/semiconductor tilt.
Many fund managers warn of malinvestment: corporate AI spending is being treated as contingent on near‑term demonstrable ROI, so the current capex surge (especially memory and data‑center builds) could be scaled back if ROI disappoints—threatening the very hardware rally driven by hyperscaler leasing/monetization models.
GEO-POLITICS
“This is one of the biggest trends potentially in the world, Jack…”—
TECHNOLOGY
“I think the most investors are sensing the business model may need to change in two ways at least one they may go like this all this seatbased pricing feels…”—
“One thing I'm seeing on the other side though is cloud revenues are ramping quite a lot and I know we're kind of dinging the hyperscalers right now because they're…”—
ECONOMY
“And this has been something I call project Zimbabwe that I'm like we're in a higher inflationary era and you're going to be surprised how high the market goes.”—
MARKETS
“A lot of this memory stuff is like six times or eight times and I'm like may I'm like I don't know maybe it's an earnings bubble but it's definitely…”—
“So they invested in OpenAI and Anthropic and they recognized gains now that the valuation has gone up so much.”—
“I was I like to track Blue Owl, right?…”—
FORECASTS
“And I look at my numbers on the S&P 10,000. That could be quite conservative actually. You know, I'm now I'm like lowering the PE to get there…”—
Intro
Intro: Eric frames a contrarian bull thesis arguing higher inflation and accelerating earnings could drive nominal markets far higher despite widespread bearish sentiment.
"Project Zimbabwe" Bull Case For Stocks
Project Zimbabwe Bull Case For Stocks: He recounts Zimbabwe experience to argue inflation can lift nominal asset prices and that investors anticipating rising nominal earnings can push markets up.
Why Shouldn't This Bull Market Be Bigger Than Dot Com Bubble?
Why Shouldn't This Bull Market Be Bigger Than Dot Com Bubble?: Eric questions 1999 as a cap on multiples, arguing revolutionary technologies like AI could justify even larger reratings.
Earnings Growth is PHENOMENAL
Earnings Growth is PHENOMENAL: The S&P’s nominal earnings acceleration (now ~12–15% versus 8% historically) can mathematically support higher P/E ratios and much stronger index returns.
The Bear Case (Counterargument)
The Bear Case (Counterargument): The hosts and guest acknowledge concentrated AI/semiconductor exposure and risks from capex misallocation, potential depreciation bombs, and episodic drawdowns.
The True Long-Term Risk: Reverse Crash Higher In All Assets
The True Long-Term Risk: Reverse Crash Higher In All Assets: A less-discussed risk is missing an upward nominal repricing where assets surge during inflation, leaving non-participants permanently worse off.
Exchanges Idea: CME ICE / Perpetual Futures
Exchanges Idea: CME ICE / Perpetual Futures: They debate perpetual futures and prediction-style contracts, CME/ICE’s franchise defense via deliverability, and potential retail-driven disruption to incumbents.
Software
Software: Software stocks face structural uncertainty from AI-driven business-model change, potential usage-based pricing, and derating despite mixed-to-rising analyst estimates.
The Sectors Where Earnings (and Earnings Estimates) Are
The Sectors Where Earnings (and Earnings Estimates) Are: Semiconductors, memory, and energy/shipping/refining show the strongest earnings momentum and estimate upgrades, despite low multiples for many beneficiaries.
Banks Are Out of the Doghouse, Being Unleashed
Banks Are Out of the Doghouse, Being Unleashed: Global banks are recapitalized, showing loan growth and regulatory willingness to lend again, making European and Japanese banks attractive value plays.
Chinese Tech Stocks
Chinese Tech Stocks: Hong Kong software-heavy names disappointed, while mainland hardware and semiconductors look cheaper and more aligned with the AI capex cycle, but property and consumer weakness remain constraints.
Related Episodes

Important Disclaimers
Not Investment Advice: For education only — not financial, tax or legal advice.
AI-Generated: Made by AI from source transcripts and may contain errors.
No Guarantees: Past performance does not predict future results; investing carries risk.
Our Commentary · Request Removal