After years of compressed or nonexistent exit windows and limited price discovery, the IPO and M&A market is opening up. Benjamin Graham called this the difference between a voting machine and a weighing machine: popularity in the short run, fundamentals in the long run. In 2026, clean energy is being weighed, and the scale of what is getting done signals that capital markets are open for business in this sector.

The deals speak for themselves. X-Energy raised over $1 billion in an upsized IPO in April. Fervo Energy became clean energy's biggest-ever IPO in May, popping 33% on debut on the strength of next-generation geothermal technology built to power AI data centers. Oklo, a small nuclear reactor company, carries a $10 billion market cap. Whether these companies deliver strong long-term returns for their early investors remains to be seen. What is clear is that public markets are assigning real, meaningful valuations to clean energy companies for the first time in years.

Not every company is getting through. Sunnova, a residential solar installer with 500,000 customers, filed for bankruptcy in 2025. When interest rates rose and California changed its solar incentive rules, bondholders recovered roughly $118 million on billions in debt. The failure was not a clean tech story. It was a business model story. Sunnova had no path to profitability once the conditions it was built for shifted.

The emerging picture — what goes public, what gets acquired, and what fails — is giving the sector something it has long needed: real data on what the market will ultimately reward and what it will not. For anyone building or backing early-stage companies in this space, that proof point accumulation matters enormously. It sharpens how investors underwrite risk, how founders think about their path to exit, and how everyone calibrates what a durable business in this sector actually looks like.

AI data centers consumed roughly 500 TWh of electricity in 2025, a figure comparable to the total electricity consumption of a large country, and that total is on track to double by 2026. A single large AI campus can require as much power as a mid-sized city. Across public, private, and infrastructure sectors, the realization is setting in: AI is fundamentally an energy story, and access to affordable electrons is becoming a hard constraint. The U.S. hosts nearly half of all global AI data center capacity, and the electrical grid was not built for this.

The result is a sudden, urgent market for reliable clean electricity, for smarter coordination of the grid capacity that already exists, and for technology that can make AI itself less energy-intensive. What was a long-horizon climate story has become a near-term infrastructure challenge. Clean energy is not just good for the planet. It is now strategically critical to the economy being built on top of AI, and that urgency is a large part of what is drawing capital markets back to the sector.

While the first leg of the AI and data center race was about massive centralized compute that trains models with racks of GPUs over months, the next leg will almost certainly be about distributed inference: bringing AI into the physical world, sector by sector. Token utilization is going parabolic, and the bottleneck isn’t chips. It’s access to a grid connection and the distributed space to put the chips. Everyone is hunting for capacity, and majors are acting on it. For operators who can load-shift peak demand to support the grid during high-stress periods, the financial incentives are significant, reducing the need to throttle compute precisely when it matters most.

At VoLo Earth, we back early-stage companies with step-change technological differentiation and a credible path to winning economics, with or without policy support. We have always believed the most durable businesses in this sector are built around genuine cost advantages, not around a favorable policy cycle. AI's energy challenge sits squarely in that frame, and it has two sides: too much demand, and a grid that cannot efficiently deliver the supply it already has. Our first two Fund II investments this year address one each.

Texture: The software layer the grid is missing. The U.S. grid has a surprising amount of capacity sitting dormant in home batteries, rooftop solar, and other distributed assets. The DOE estimates 80 to 160 GW of this exists but goes uncoordinated, in large part because the thousands of utilities, installers, and equipment manufacturers involved are still working off emails and spreadsheets. Texture is building the platform that brings all of them onto the same system, enabling utilities to orchestrate existing clean energy assets in real time rather than defaulting to building new power plants every time demand spikes. We co-led their Series A.

Refiant AI: Improving AI inference quality while also reducing energy demand. Today's AI models have fundamental efficiency and performance problems: the more information (context) you (or agents) reason over, the more energy models consume and the more they “summarize” completed work. Refiant's architecture, inspired by how the human brain focuses selectively rather than processing everything at once, achieves orders of magnitude improvements to both context window and energy efficiency compared to leading AI models. In independent benchmarking, Refiant required only a laptop's worth of memory rather than a data center rack. At the scale AI is now operating, that difference is not incremental. It is structural. We led their Seed round.

While we track many themes across the energy transition, one area that ties together everything above particularly stands out: physical AI entering the built environment. 

Buildings and construction account for 37% of global energy-related emissions yet remain largely untouched by modern software and smart hardware. As AI moves from centralized data centers into commercial and industrial buildings, the demand for systems that can deploy compute efficiently, reduce energy loads, and support grid flexibility is growing fast. We already have a front-row seat through our portfolio company Blue Frontier, which delivers significant advantages to both the grid and buildings while decarbonizing air conditioning. We think the most important companies in this space are being built right now, and we are actively looking for more. 

The weighing machine has arrived. Exit data is beginning to confirm which business models hold their value under real market conditions. Our job is to find those companies before the verdict is in, and the evidence so far tells us the window is open.

Pulse 2.0 sat down with co-founder and managing director Kareem Dabbagh for an in-depth profile of VoLo Earth covering the firm's operator-first investment approach, its focus on backbone sectors where better economics and decarbonization move together, and key milestones including closing Fund II at $135M and top-decile Fund I performance on Carta's benchmarks.

Partner and COO Elaine Hsieh joined the Startup Day Mainstage to discuss how decarbonization solutions are being "rewrapped" as language adapts to today's policy and market environment (featuring Skyven's Arun Gupta as a Fund I portco example) and served as a mentor during Speed Mentoring. Senior Associate Erin Kirton spoke on "Storage Goldilocks: Finding the Right Battery Solution for Your Organization," helping commercial and industrial decision-makers navigate the battery landscape.

ImpactAlpha's Shaping the Algorithm series exploring how African startups are building AI-native infrastructure named VoLo Earth as the lead investor in Refiant AI, highlighting the firm's thesis that ultra-efficient, edge-deployable AI is critical for markets where cloud access, compute power, and reliable energy are constrained.

During SF Climate Week, Elaine delivered the Closing Keynote at the BERC Energy Summit at UC Berkeley, speaking on pragmatic approaches to moving decarbonization markets. Kareem spoke at a PowerLines panel on market design and regulatory pathways, and joined Heatmap News' Heatmap House for a conversation with Katie Brigham on grid reliability, making the case that data centers requiring 99.999% uptime cannot function without a firm grid connection.

VoLo Earth was recognized as #23 on Climate50's list of leading climate tech investors, joining a cohort of firms driving capital into the technologies reshaping the energy system.