AI’s power needs are reshaping energy finance before the scale of that demand is clear. Energy startups raised $12.6 billion in initial public offerings during the first half of 2026, more than the $4.3 billion raised across all of 2025. The first-half total is the highest since 1999, when the dot-com boom drove a similar rush to market, according to the Financial Times. That comparison captures the speed and enthusiasm of the fundraising, but it doesn’t establish that the two periods have the same economic foundations. IPO proceeds show that investors are willing to fund companies; they don’t show that the companies can build projects, sell power profitably, or meet the timelines buyers expect.
Investors first poured money into companies tied directly to artificial intelligence, including chipmaker Nvidia. Now they’re looking further down the supply chain to the electricity required to run data centres. “Every chip needs energy to power it,” RBC clean energy analyst Chris Dendrinos told the Financial Times, describing the resulting demand as a “huge tailwind” for energy companies. The shift matters because electricity is a physical input to computing, not just another category of technology spending. More computing capacity requires power, but translating that broad relationship into reliable forecasts for particular energy companies is harder: demand depends on how quickly data centres are built, how intensively they’re used, and what sources can supply them.
Capital reaches technologies still seeking proof
The fundraising wave spans established energy businesses and early-stage technologies. Nuclear fusion, enhanced geothermal energy and space-based solar power are drawing attention despite unresolved technical and commercial hurdles. Meta has signed a deal with startup Overview Energy to develop as much as 1 gigawatt of space-based solar capacity. That agreement indicates interest in exploring a possible source of power, but a development deal isn’t the same as a completed facility delivering electricity. For investors, the distinction between a promising concept and an operating asset is where much of the risk sits.
Fusion has been particularly difficult to finance through public markets. Its high research costs and experimental status have historically left development to governments and large public projects. That’s changing as fusion startups pursue IPOs and attract investors looking for potential sources of abundant power. Public-market funding can give companies access to capital and put their plans under closer scrutiny, but it also brings pressure to show progress in a field where research and engineering challenges can take years to resolve. Investors are being asked to finance the possibility of future power, not simply the expansion of a technology already producing it at scale.
The case for backing multiple technologies rests partly on the uncertainty around AI’s eventual electricity consumption. Forecasts remain difficult as large language models change quickly and spread across more applications. A forecast made under one set of assumptions can shift as computing methods, customer use and data-centre plans change. Yet the expectation of rising data-centre demand is already influencing capital allocation. OpenAI co-founder and CEO Sam Altman said at the 2024 World Economic Forum in Davos that meeting future energy needs would require a breakthrough, adding that the prospect motivated investment in fusion. That’s a useful signal of how some technology leaders frame the challenge, though it doesn’t settle which energy source can meet it, or when.
Here’s the thing: when demand is uncertain, investors may spread money across several possible solutions rather than wait for a single winner to emerge. That can support research and project development, but it can also make market enthusiasm look broader than the near-term supply available. If you’re working in this space, the distinction between announced capacity, projects under development and electricity actually delivered is central to judging the claims. And this is the part most people overlook: the technologies competing for attention may have very different timelines, costs and technical risks, even when they’re grouped together as answers to the same power problem.
Renewables compete on cost and reliability
Funding for emerging technologies is one part of a broader contest to add power capacity. Solar photovoltaics was the largest source of new U.S. generating capacity for 28 consecutive months through the end of 2025, accounting for 72.6% of electricity additions, according to Federal Energy Regulatory Commission data. Those figures describe additions to generating capacity, rather than the share of electricity produced over time. That distinction matters: capacity can show what’s being built, while actual output depends on operating conditions and how generation fits into the wider power system. Still, solar’s position in new capacity points to an established technology already attracting a large share of investment and construction.
EDP chief executive Miguel Stilwell d’Andrade told Semafor that the United States was in “one of the best periods to invest in renewables” in two decades. His assessment points to an economic argument for clean power that can persist even amid a difficult policy climate: developers and customers need new supply as electricity demand grows. For energy buyers, the appeal of renewables will depend on whether projects can be built and connected when power is needed, alongside the cost and reliability of the supply. A positive investment climate can help projects advance, but it doesn’t remove practical constraints or make every proposal commercially sound.
That said, the funding race isn’t a simple contest between established renewables and speculative technologies. Electricity systems need supply that can meet demand across changing conditions, and the article’s figures don’t show how much each category will contribute in practice. Emerging options may offer potential, while solar is already adding capacity; the financing question is how those different paths fit together and how quickly they can produce usable power. The numbers here are striking. The proof will come in projects that reach operation.
What investors will need to watch
The current IPO surge reflects expectations, not proof that every emerging technology can deliver power at commercial scale. Still, the funding figures show investors are treating energy supply as a central part of the AI buildout. The market’s next test will be whether capital translates into dependable generation fast enough to meet demand. That test involves more than raising money: companies have to advance projects, overcome technical hurdles and demonstrate that customers can rely on the power they promise.
For readers weighing the significance of the boom, the comparison with 1999 is a reminder to distinguish a powerful theme from the performance of individual companies. The dot-com analogy describes a rush of investor interest, but it doesn’t by itself predict what happens next in energy. Demand for electricity may provide a strong reason to build, while uncertainty about the eventual scale of AI use leaves room for forecasts to change. Investors and customers alike will be looking for evidence in completed capacity and delivered power. Until then, the funding is a measure of confidence in the need, and a wager on which technologies can meet it.