Semiconductor Decoupling Drives Energy Demand as US Onshoring Reshapes Power Markets

Semiconductor Decoupling Drives Energy Demand as US Onshoring Reshapes Power Markets America's push to bring chip fabrication home is accelerating power demand growth that benefits energy storage and nuclear baseload plays. The US government has started throwing billions of dollars at bringing chipmakers to American soil, moving beyond its traditional strategy of controlling semiconductor design while outsourcing fabrication. The shift, documented in Adam Tooze's Chartbook newsletter and the Economist, represents an attempt to hoover up lower-value production segments that Washington previously left to Asian manufacturers. The energy implications are direct and large.^5,6 Chip fabrication plants are among the most electricity-intensive industrial facilities in operation. Each new fab consumes power equivalent to a small city. When multiple facilities cluster in the same region, as Arizona and Texas are now experiencing, they create step-change increases in local grid demand that utilities must plan for years in advance. The semiconductor decoupling from Asia is, at its core, an energy demand story. Capital is already rotating into the companies positioned to supply that power. Fluence Energy's stock surged 98% in a single week as investors targeted energy companies that can serve AI data centre and industrial buildouts. Nuclear and renewable baseload generation are attracting the heaviest flows, reflecting a market view that intermittent sources alone cannot meet the round-the-clock demand profile of fabs and data centres.³ Fluence Energy sits at the intersection of grid-scale storage and industrial power demand. The company reported a record backlog and signed master supply agreements with two major hyperscalers, expanding into the data centre energy storage market. Management reaffirmed its 2026 revenue target of approximately $3.2 billion to $3.6 billion, with 85% of the midpoint already contracted. Supply chain disruptions that delayed around $80 million in shipments are resolving.² But the stock's trajectory reveals the market's ambivalence. Despite the 98% weekly surge, Fluence continues to report net losses. A secondary offering of 20 million Class A shares in mid-May, priced around $21.00, increased the public float and triggered immediate price volatility. The 52-week range of $4.40 to $33.51 reflects a company that the market cannot decide is a growth story or a cash-burning speculation.^2,1 The broader debate about industrial policy frames the energy trade. The Economist noted that the certainty with which policymakers once advocated free trade has given way to a consensus that globalisation needs guardrails. New industrial policies — chip subsidies, energy security mandates, domestic content requirements — are reshaping capital allocation across the energy sector. Each guardrail creates a demand signal for local power generation that would not exist under pure free trade.⁴ America's economic outperformance has been built partly on capturing the highest-value segments of the global semiconductor industry while letting others handle the energy-intensive manufacturing. Bringing those segments home reverses the energy arbitrage. The power that was previously consumed in Taiwan, South Korea, and Japan to make chips for American designers will now be consumed in Arizona, Ohio, and Texas.⁵ The scale matters for energy markets. A single advanced fab can draw 100 MW or more of continuous power. Multiply that by the dozens of facilities under construction or planned, and the incremental demand reaches gigawatt scale within a decade. Gas-fired generation, nuclear restarts, and grid-scale battery storage all benefit from the demand pull. Analysts expect Fluence's deferred Q2 shipments to bolster upcoming quarterly results as delivery schedules normalise. The company's enterprise value of $3.22 billion against a revenue target midpoint of $3.4 billion implies a price-to-sales ratio of roughly 1.0, modest by growth-stock standards. The question is whether the growth materialises fast enough to justify the cash burn.² The signal to watch is whether semiconductor onshoring announcements translate into utility interconnection requests and power purchase agreements in the regions where fabs are being built. The energy demand is real, but the timeline from announcement to first silicon — and first megawatt consumed — can stretch five years or more. Energy investors betting on the semiconductor reshoring theme need to match their positioning to that timeline, not to the headline cycle.^5,3