PodcastThe Monetary Matters NetworkWarren Pies

Earnings Estimates Are Going Vertical on AI Demand | Warren Pies

Discussion balanced a disruptive geopolitical shock to oil flows through the Strait of Hormuz—which could lift prices into the $150–$200+ range and drain significant inventories—against a powerful AI‑driven demand cycle for compute that has sparked GPU shortages, hyperscaler capex and a semiconductor rally. Hosts pointed to firm tools (Caliban), rapid ARR growth at vendors like Anthropic, and near‑vertical analyst EPS revisions as evidence that revenue traction, not capex accounting, should guide the bull narrative. Markets are top‑heavy in their gains, prompting tactical recommendations to overweight equities and commodities while hedging oil risk, with a base case that the oil shock is temporary over a ~6–9 month horizon.

Artificial IntelligenceCorporate EarningsSemiconductorsAI InfrastructureEnergy MarketsGeopolitical Risk

Preview · 16 more points in the app

Read the Full Summary

Key Takeaways

A Strait of Hormuz closure has already collapsed oil shipping and could lift prices into $150–$200+ scenarios, mechanically widen Brent–WTI spreads via policy responses, and is expected to persist until diplomatic engagement eases — creating acute market risk for months.

Speakers' base case is that the oil shock is likely temporary (reopening and fading within ~6–9 months) and equities can remain bid until real‑economy evidence forces a rerating, but sustained oil above the flagged $140/barrel macro pain threshold would meaningfully depress U.S. activity.

The physical oil strain (a discussed multi‑month supply hole on the order of ~10 mbd and potential ~1 billion barrel inventory drain from an ~8 billion stock base) increases regional shortages (jet fuel in Asia first) and limits the effectiveness of SPR draws unless prices spike or governments impose managed restrictions.

Portfolio positioning has shifted to overweight equities and commodities (funded by bond underweights) with modest energy‑equity allocations (~4–5%) used as a risk‑dampening diversifier; practitioners recommend pairing equity overweights with explicit oil exposure (example: long August‑Brent) and scaling out into near‑term spikes (example sell target: Brent ≈ $130).

A true compute squeeze — Blackwell GPUs effectively unavailable and a collapse in GPU availability from mid‑March after a December ramp — plus Mythos frontier evidence validated the causal chain (model improvement → compute demand → hyperscaler capex), triggering April upgrades and powering the semiconductor rally.

GEO-POLITICS

Hormuz Disruption — Speakers contrast a prolonged Strait of Hormuz closure — with oil shipping collapsing to only a few shipments per day and an unprecedented barrels‑per‑day loss — against an AI narrative that lets equities look through near‑term oil shocks; this disruption could lift prices toward $150–$200+ and, if export bans occur, mechanically widen Brent–WTI spreads. They expect the disruption to persist until diplomatic engagement eases shipping constraints, creating meaningful near‑term market risk.

“I've spoken to oil analysts who I respect and they tell me that the shortage caused by the blockage of the Strait of Hormuz is the biggest ever in terms…”—

TECHNOLOGY

GPU Shortage — Proprietary data showed GPU availability collapsed in mid‑March and the newest Blackwell GPUs were effectively unavailable, signaling a scramble for high‑end compute that began ramping in December and helped lead the semiconductor rally.

“so availability collapsed. We were wondering you know what's going on? seems like there's a real scramble for compute…”—

Mythos Leak — Participants said the Mythos leak indicated frontier models keep improving with more compute (the leak was withheld for cybersecurity reasons) and argued the chain—model improvement → compute demand → hyperscaler capex—underpins the current bull market.

“That mythos leak I think what it represented was that frontier models continue to get better you feed more compute better models that whole train of events is intact And…”—

Caliban Adoption — Caliban was productized and rapidly adopted by clients this year, now producing the majority of the firm's charts and automating research tasks like PDF ingestion, presentation regeneration, charting and reporting, making it an operational backbone.

“I'd like that. And so I guess it was the beginning of this year we started allowing our clients to have to have access to it…”—

ECONOMY

Earnings Revisions — Year‑to‑date analyst EPS estimates have risen roughly 20–25% for the next 12 months and over 252 days, a near‑vertical pace described as almost unprecedented and not driven by a prior earnings drawdown.

“Highly anomalous. We only see this coming out of really recoveries where earnings are starting to hockey stick higher.”—

Structural Upside — Expected year‑end downgrades did not materialize and analysts historically 'sandbag' forecasts; with corporate beat rates near ~70%, speakers argue current revisions may understate true earnings strength.

“The reason we had that confidence is that usually in when you're talking about Q4, you focus in on Q4 estimates,...”—

Capex Circularity — Blockbuster hyperscaler capex was initially seen as a capex 'top' that would depress free cash flow, but speakers argue capex reduces FCF today while feeding an AI‑driven earnings boom, and that high oil and housing dynamics make a 'peak Goldilocks' less likely.

“You've got capex which is you not being realized right now it does it knocks free cash flow but not earnings obviously ly and it's feeding into the earnings boom…”—

MARKETS

Top‑Heavy Gains — A disproportionate share of recent EPS upgrades comes from a handful of names and sectors—notably high‑bandwidth memory (e.g., Micron), Nvidia and large oil companies—producing a top‑heavy headline gain that raises breadth concerns.

“One of the big push backs is that this is a topheavy earnings advance…”—

Breadth Improvement — Despite headline concentration, median estimates and many individual stocks have improved; speakers stress practical confirmation rules (for example, equal‑weight confirming cap‑weighted new highs within ~90 days) because leadership shifts can signal risk.

“first like it's it's true like so what you're saying is true there is a huge spread between the median industry earnings estimate growth versus the index level and that's…”—

Valuation Signal — The valuation spread between semiconductors and software proxies is treated as a market proxy for investor views on 'how much compute is needed' versus 'how disruptive software will be,' with movements in the spread signaling confidence in the AI trade.

“I think that is a until we sort the software, how software actually fits into this world of AI…”—

FORECASTS

Track Record — The host invokes Warren's prior S&P‑7,000 by 2026 call to contextualize and lend credibility to the firm's present forward views.

“The last time I spoke to him, Warren was calling for S&P to 7,000 by 2026, which obviously happened.”—

S&P Targets — The firm published a base‑case S&P 500 target of ~7,000 in 2026 and suggested 8,000 in 2027, explicitly tying further upside to how far AI drives earnings and valuation.

“We wrote a report earlier in the year where we were talking about that 7,000 call for 2026…”—

Oil Hedge — One speaker says they will only hold an equity overweight if paired with long oil exposure (example: August Brent) to hedge oil‑related tail risk.

“The only way I can get okay and comfortable with holding an overweight equity position and being a bull is by pairing that with a long oil commodity exposure over…”—

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