GE Vernova: The Electrification Supercycle Is Bigger Than Data Centers, and the Services Flywheel Is Just Getting Started
Bernstein 42nd Annual Strategic Decisions Conference, May 27, 2026 — CEO Scott Strazik lays out a case for a decade-long earnings inflection that the market is still underpricing
The single most important message Scott Strazik delivered at the Bernstein Strategic Decisions Conference on Tuesday was one of deliberate reframing: GE Vernova is not a data center play. It is something considerably larger and more durable. With a $163 billion backlog, $10 billion of cash on hand, and an electrification business growing at a pace that is already ahead of its own aggressive financial targets, Strazik used his opening remarks not to sell a story but to correct one.
Electrification: The $300 Billion Addressable Market That $20 Billion in Revenue Has Barely Touched
The electrification segment is the fulcrum of the GE Vernova investment case right now, and the numbers Strazik presented make clear why. The business ended 2022 with a $9 billion backlog. By the end of the first quarter of 2026, that figure had reached $42 billion, already well past the midpoint of the company's stated target of doubling its year-end 2026 backlog of $30 billion to $60 billion by 2027. Strazik noted that electrification orders in Q1 2026 alone exceeded the total for all of 2025, and he flagged that Q2 is shaping up to be another strong quarter.
The financial target for the segment — $20 billion of revenue at 22% EBITDA margins by 2028 — has been described by management as a floor, not a ceiling. When set against a serviceable addressable market that Strazik pegged at $300 billion today, even hitting that floor implies a market share that leaves an enormous runway. "I can't give that same proportional market share in gas power, in wind power in North America primary markets, where our share of the global market is just substantially larger," he said. "So by default, this is a business that we see an incredible amount of opportunity for us to serve and meet this moment."
The four businesses within electrification — power transmission, grid systems integration, power conversion, and grid software — are each at different stages of growth. Power transmission, the largest, benefited immediately from the February 2026 close of the full ProLec acquisition: GE Vernova has already booked $500 million of incremental year-to-date orders by routing non-U.S. factory capacity from India and Turkey to North American customers, something the previous joint venture structure explicitly prohibited. Strazik was direct about the strategic error the JV represented: "With a 50-50 JV, that was clunky with the customers when we couldn't fully integrate those transformers to solutions they wanted without interacting with another partner." The fix, in his telling, is already generating returns well ahead of deal-day expectations.
The Stability Block: A New Product That Signals Where Integrated Power Revenue Is Heading
The most actionable new product disclosure at the conference came from the power conversion segment. GE Vernova is developing what it is calling a "stability block" — a medium-voltage uninterruptible power supply solution designed to sit inside or adjacent to hyperscaler data centers and absorb the extreme frequency volatility that AI workloads impose on power infrastructure. Strazik described it as part of a "string of pearls" integrated solution that stretches from power generation through electrical equipment and into software, and he positioned it as a direct response to what hyperscalers are co-designing with the company right now. The energy management system software disclosed at Q1 earnings is the connective tissue; the stability block is the next physical component in that stack.
The strategic logic behind this product is significant. GE Vernova's competitive moat in integrated solutions is that almost no other company can credibly offer power generation equipment, electrical infrastructure, and the software layer simultaneously. As Strazik put it, "we're one of the few companies that can provide the power generation equipment, the electrical equipment and the software to allow the end application to work the way it's supposed to." Each layer of integration added to a customer relationship deepens switching costs and extends the aftermarket opportunity — which is precisely how the services flywheel accelerates into the next decade.
Gas Power: 100 Gigawatts Contracted, But the EPC Bottleneck Is Real and Is Already Shaping the Order Book
On gas, the headline figure is 100 gigawatts on contract, primarily delivering through the end of the decade with some units now trickling into 2031. Strazik contextualized the scale: GE Vernova's total installed base is 720 gigawatts, of which only roughly 200 gigawatts runs as full firm baseload today. The company's projection is that baseload capacity at least doubles to 400 gigawatts by the mid-2030s, which is the mechanical driver of the services backlog expansion.
The EPC and construction bottleneck, however, is the most honest constraint Strazik acknowledged. The inability of engineering and construction firms to commit to project timelines is already distorting the order pattern in a visible way: coming into 2026, GE Vernova still had available 2029 turbine slots, yet customers were locking in 2030 capacity because no EPC would guarantee a 2029 installation timeline. "There won't be a patio in this country that's ready for a gas turbine that won't have a gas turbine on it," Strazik said. "But there's also not a need to build them ahead of when that moment is there." The discipline is deliberate, but it underscores that the constraint on near-term growth is not demand — it is installation capacity.
Nuclear: The Game This Decade Is Booking Contracts, Not Booking Revenue
The BWRX-300 SMR program remains sub-$1 billion in revenue in 2026, but the strategic posture is shifting. Construction is underway at the Darlington site in Ontario, which is a four-unit site even though only unit one is being built now. Up to 10 U.S. contracts could be signed before year-end. Sweden represents up to five units. A Polish engineering study is in progress. Strazik's framework was explicit: the purpose of building a contracted book of roughly 20 units in the near term is not to generate meaningful near-term earnings — it is to give the company leverage to go back to long-lead suppliers and drive down unit costs by reindustrializing a supply chain that has been starved of investment for fifteen years.
The $40 billion SMR funding commitment embedded in the U.S.-Japan trade deal, which Strazik said covers approximately 10 SMRs for GE Vernova, is the policy catalyst that underpins this logic. He was unsentimental about the timeline: "Will it really replace gas? No. But over time, will more customers carbon dollar cost average in some zero-carbon baseload power that probably will still be at a pricing premium relative to unabated gas? Yes, in a healthy mix." Nuclear is a next-decade income statement story. This decade, success is measured in contracts signed and supply chain investment committed.
AI and Automation: Dilutive Now, Accretive Starting 2027
One of the more precise disclosures Strazik made was the phased financial contribution of AI and factory automation. Through 2026, AI is a net modest negative to GE Vernova's financials given investment spend relative to returns. That inflects in 2027. The company's AI foundry has grown to more than 50 people and is scaling to 80 by fall, focused on engineering applications that are already enabling the company to absorb rapid growth without a proportional increase in headcount. Strazik described meaningful progress in matching installed base service needs with parts and resource scheduling using decades of proprietary operating data.
Factory automation follows with roughly a one-year lag. Eight "lighthouse" automation programs are running as proof-of-concept through the first half of 2027. Once validated, the capital expenditure to roll those applications across GE Vernova's global factory network is already embedded in the financial plan, with the resulting productivity gains expected to become visible in 2028 margins. The small acquisition of Robotech, a 35-engineer Montreal robotics firm, fits this architecture: Strazik is pulling it out of its other commercial relationships to serve as a dedicated automation center of excellence, with access to talent from McGill University.
On supply chain, Strazik noted that the rare earth situation — which was described as urgent as recently as summer 2025 — has materially improved. The company has secured multi-year inventory and is investing in alternative sourcing simultaneously. Castings and forgings suppliers are performing ahead of schedule, in part because GE Vernova funded their capacity expansion directly through CapEx in 2024. The company has installed 305 new machines across its gas power factories in the past five-plus quarters and has another 100 to install by year-end.
Capital Allocation: The Buyback Is the Default, M&A Must Clear a High Bar
With $10 billion of cash on hand and free cash flow generation running well ahead of the $11 billion four-year R&D and CapEx envelope, GE Vernova's capital return program is substantial. The company has repurchased approximately 10 million shares for $4.6 billion over the past five to six quarters. Strazik's characterization of M&A discipline was notably frank: the test for any transaction is whether it makes life easier for his management teams to execute on existing backlog or adds another ball for them to juggle at a moment when "few in our industry have experienced as much core growth as we're experiencing right now." Vertical supply chain integration in gas, SMR long-lead items, and electrification adjacencies are in scope. New product categories are not, at least not now.
The Macro: This Is 1945, Not a Data Center Cycle
Strazik pushed back explicitly on the framing that GE Vernova is riding a three-to-five-year hyperscaler capex wave. Data centers represent only 20% to 25% of the current backlog. The analogy he reached for was the post-World War II electrification of the American economy — a structural, multi-decade build-out driven by economic growth, national security, decarbonization, and globalizing demand across Asia and the Middle East simultaneously. Vietnam alone, with a 90 gigawatt grid, has signed contracts for 8 gigawatts of LNG-to-power projects in the first five months of 2026. Japanese utilities are contracting for 2030 and 2031 shipments against 2033 commission dates, and Strazik noted their business cases are entirely insensitive to current LNG spot prices given the delivery timeline.
The biggest misconception Strazik wanted to correct, when asked directly, was precisely this conflation of GE Vernova with the data center trade. The $87 billion services backlog — set to generate approximately $20 billion of services revenue in 2027 — is the foundation. The equipment backlog growth, impressive as it is, is the mechanism that makes the next decade's services business larger and more profitable than the current one. Emerging demand vectors including defense, drone infrastructure, and what Strazik called "golden dome" military applications for electrical equipment add further optionality that is not in any current consensus model.