GrowthTurtle
Data Science PhD & IEEE Fellow, Ex-Two Sigma
GE Vernova jumped to the top of the Bottleneck Index after the Defense Production Act's Section 303 was invoked for power transformers. Policy is demand, and a DPA invocation is an unlimited checkbook walking into a market that is already short, so stacked on a bottleneck the upside becomes arithmetic rather than forecast. x.comÂ
GE Vernova is the cleanest way to own the power transformer and switchgear bottleneck choking the AI buildout, which retail is still not pricing in. Half of 2026 US data centers are delayed by equipment shortages, prices are up 77-95% since 2019 and US peak power demand is forecast to rise 166 GW in five years. It guided $41-42B for 2026, and a PEG of 1.9 shows it is earning its multiple. Risks: the stock nearly tripled, wind still loses money, and Eaton, ABB and Hubbell compete for orders. x.com x.com x.com x.com x.com x.comÂ
GE Vernova's turbines and grid equipment make it a clear winner from the AI data-center power crunch. Q3 equipment sales rose 42%, total revenue 27% and operating income 281% as margins expanded. Grade: A. x.comÂ
The AI power trade is starting to separate promises from electrons, favoring GE Vernova over NuScale Power because the market wants power now, not in 2033. GE Vernova has a $176bn backlog with gas turbines effectively sold out for years. NuScale reported just $75,000 of revenue last quarter, thousands not millions, then filed to sell another $750m of stock. x.com x.comÂ
GE Vernova is a cyclical gas-turbine manufacturer priced as a secular compounder, the basis for a sell rating. Per @StockWorthyApp, the stock fell 8.62% in a day and 9.43% over a week on concerns about an AI development slowdown. x.comÂ
Demand for AI power is accelerating, and power is one of the AI bottlenecks where GE Vernova is positioned. Its Q1 2026 new orders rose 71% year over year, led by equipment orders up 106%, and the quarter alone exceeded the first three quarters of 2025 combined. x.comÂ
Grid suppliers are winning the next AI bottleneck, because power must reach data centers through transformers, substations, switchgear and transmission the grid lacks. Transformer lead times stretched from 24 months in 2019 to five years, per HSBC, and PJM got 220 GW of generation applications by April 2026 versus 2.1 GW connected in 2025. The White House designated switchgear and substations defense-essential. Forgent, Eaton, Powell, Hubbell, nVent, Quanta, MYR, AMSC, GE Vernova and Itron fit. … Show more
GE Vernova is highly expensive, and infrastructure booms always look unstoppable until they stop; a cyclical boom-bust stock should not need a three-years-out view to reach a 20 P/E. Gordon Johnson's $470 price target is right in theory. Barring a crash, it is probably among the safer expensive stocks through year end as fund managers pad top holdings, but the cycle will pass, as similar ones did before the financial crisis. x.comÂ
Gas turbines, not nuclear or renewables, are the real AI infrastructure bottleneck, and GE Vernova owns that chokepoint. Its order book is full through 2029 and now booking into 2031, turbine prices are up 300% in three years to over $250 million each, and AI and data centers make up 20% of its gas power orders. Microsoft bought seven turbines for 2.7GW in Texas, and xAI's Colossus and OpenAI's Stargate already run GE Vernova turbines. Nuclear headlines are noise; gas powers the build now. x.comÂ
GE Vernova and the other big turbine makers now behave like memory makers, restricting output and raising prices rather than repeating their 1990s-2000s boom-bust cycle. That discipline has been great for their stocks, and GE Vernova is sold out until 2030. Unlike memory buyers, though, power customers have alternatives, so the congested turbine backlog leaves the field ripe for disruption by behind-the-meter generation. x.comÂ
GE Vernova's tanking earnings are a dealbreaker, even after the stock's 60% gain in six months. It had been an earlier pick for 2026 as the year of energy following AI's year in 2025. x.comÂ
GE Vernova's AI-driven power demand is compounding faster than its Wind segment can drag on results. Data center orders already exceed $5B, more than double 2025's full-year total, even as headline EPS came in below estimates. The market is fixating on the wrong line. x.comÂ
