GEV turbine pricing doubled to $2,500/MW with a record $150B backlog while NVDA operating income surged 147%, yet consensus models still underestimate both.
Crawford argues sell-side linear models can’t capture an exponential capex cycle, leaving Street EPS estimates systematically too low across the entire AI supply chain.
Don’t wait: the analyst who called NVIDIA in 2010 just revealed his top 10 AI stocks. See the full list FREE now.
On a recent episode of the Animal Spirits podcast titled Talk Your Book: AI Is Not a Bubble, Alger portfolio manager Dr. Ankur Crawford made a deceptively simple argument that cuts against most of the current debate over AI valuations: investors are arguing about the price-to-earnings ratio without first agreeing on what the earnings actually are.
TippaPatt / Shutterstock.com
“The first thing you need to get right when you think about valuation is the E. Only then can you come up with a PE,” Crawford said. She runs a concentrated 30-name portfolio and targets companies she believes can double or triple over a roughly three-year horizon. She contends that sell-side models are linear extrapolations bolted onto an exponential capex cycle, leaving Street estimates “just too low” across the AI supply chain.
The Case Study: GE Vernova
Crawford pointed to GE Vernova (NYSE: GEV) as the cleanest illustration. Only three companies make combined-cycle gas turbines globally, GE Vernova holds roughly a third of that market, and pricing has doubled from about $1,250 to $2,500 per megawatt in under a year as hyperscalers scramble for firm power.
The financials support the framing. Q1 2026 revenue came in at $9.34 billion, beating the $9.3 billion consensus, while organic orders jumped 71%. The Electrification segment booked $2.4 billion in data center equipment orders in a single quarter, more than all of 2025. Backlog hit a record $150 billion at year-end 2025 and grew $13 billion sequentially in Q1. CEO Scott Strazik told investors the company expects to reach at least 110 GW of combined gas turbine backlog and slot reservations by year-end 2026.
Don’t wait: the analyst who called NVIDIA in 2010 just revealed his top 10 AI stocks. See the full list FREE now.
The stock has run hard, up 72.84% year-to-date and 132.18% over the past year, trading at a forward P/E of 40x. Crawford’s point is that this multiple is built on consensus EPS, and consensus has missed Vernova’s revenue every quarter shown in the data, with beats ranging from 0.36% to 9.17%.
The Picks-and-Shovels Backlog
The same pattern shows up across the AI power buildout. Vertiv Holdings (NYSE: VRT) posted Q4 2025 organic orders growth of 252% year over year, lifting backlog to $15.0 billion, and has beaten EPS estimates by between 4.62% and 25.01% across three quarters. Eaton (NYSE: ETN) saw Electrical Americas’ trailing 12-month organic order growth accelerate from 7% in Q3 2025 to 42% in Q1 2026, with electrical backlog up 48%.