AI Hyperscaler Deals Drive Fluence 98% Higher as Natural Gas Extends Weekly Rally

AI Hyperscaler Capex Is Now Moving Energy Markets Directly The feedback loop between artificial intelligence infrastructure spending and power demand is no longer theoretical — it is printing in gas prices and energy equity valuations simultaneously. June Nymex natural gas settled at $2.96 per million British thermal units last Friday, a gain of 2.3% on the day and 7.4% for the week, as power-sector demand expectations hardened alongside hotter weather forecasts and resilient LNG export flows.² The LNG side of the equation reinforced the constructive read. Weekly vessel departures reached 141 billion cubic feet, up 26 Bcf from the prior week despite maintenance activity at several export facilities — a signal that global demand is absorbing supply without difficulty as the summer demand season approaches.² The equity market is pricing the longer-duration version of this story with far less patience. Fluence Energy, a battery storage and grid software company, closed at $24.16 on May 8 after a 98.2% single-week surge, triggered by the disclosure of master supply agreements with two unnamed hyperscalers and a record $5.6 billion backlog. Capital is rotating into companies positioned to supply power and storage infrastructure for AI data centre buildouts, with nuclear and renewable baseload generation regarded as the cleanest load solutions for hyperscalers facing emissions scrutiny.¹ The operational case at Fluence has improved materially. The company delivered positive adjusted EBITDA of $2.0 million in Q1 2026, the fourth consecutive quarter in the black, with non-GAAP gross margin expanding to 52%. Its PowerTrack software platform manages 37.5 gigawatts of solar assets under management, with annualised recurring revenue guided to $65–70 million by year-end.¹ The balance sheet is the countervailing risk that traders should not ignore. Stockholders' equity stands at negative $265.88 million and cash at just $36.59 million — a combination that leaves the company dependent on continued backlog conversion and capital market access at precisely the moment when rate expectations remain fluid. A hyperscaler capex slowdown, or any schedule slippage on those supply agreements, would test the stock's new valuation floor hard.¹ The broader IPO pipeline amplifies the demand case but adds its own uncertainty. OpenAI, currently valued at around $500 billion in private markets, and Anthropic, valued at roughly $180 billion in a recent funding round, are both said to be exploring public listings. Any of these transactions could rank among the largest US IPOs since Alibaba raised $25 billion on the New York Stock Exchange in 2014.⁴ Fresh public-market capital at that scale would directly fund the next wave of datacenter construction — meaning the power and gas demand implied by current AI capex plans has not yet been fully built, let alone energised. In European markets, the demand-and-supply tightness is already registering. German Q2 spot power prices are forecast to surge 17% year on year, while gas prices face an even steeper move, with analysts projecting a 40% increase over the same period, even as expanded solar generation and softer industrial demand work against the magnitude of those gains.³ The contrarian read is that the energy market is front-running a demand story that may not materialise at the expected scale or timeline. Hyperscaler commitments to suppliers like Fluence are contractual, but the ramp schedule for hundreds of gigawatts of new AI compute has repeatedly shifted to the right. If IPO windows close before the largest AI companies go public, the capital unlock that underpins the next datacenter wave gets deferred. Gas at $3 already prices in a meaningful demand tailwind; positioning here is essentially a bet that the AI buildout stays on schedule. What to Watch Nvidia's earnings report is the immediate catalyst — guidance on data centre revenue and inference chip demand will either validate or complicate the power demand thesis priced into both gas futures and energy infrastructure equities. The formal progress of SpaceX and OpenAI toward public listings deserves tracking as a leading indicator of how much new capital will be committed to compute infrastructure in the next 12–18 months. On the gas side, the weekly storage report will clarify whether the demand draw from power generation is running ahead of or behind seasonal expectations at the $3 threshold. Any announcement of additional hyperscaler supply agreements in the storage and grid software space — following the Fluence template — would confirm that the capital rotation into AI power infrastructure is broadening beyond a single-name trade.