The timeline for commercial fusion power just got a lot tighter. On Thursday, Helion, the Everett-based startup backed by Sam Altman, announced a $465 million Series G funding round, pushing its total valuation to $15.5 billion. The goal is no longer just research; it is construction. The company is now racing to complete Orion, its first power plant, with a target to begin feeding electricity into the grid by 2028.

This is a high-stakes bet on a timeline that most fusion experts consider aggressive. While the broader fusion industry typically points to the mid-2030s for commercial viability, Helion’s deal with Microsoft requires it to deliver power much sooner. The capital, led by Thrive Capital, brings the company’s total funding to $1.5 billion, providing the runway necessary to move from prototype to production.

A Different Approach to the Grid

Most fusion startups are essentially building high-tech steam engines. Whether they use lasers to compress fuel or massive magnets to contain plasma, the end goal is usually to generate heat, boil water, and spin a turbine. Helion is taking a fundamentally different path.

Instead of relying on steam, Helion uses magnets to compress fuel until it reacts. When the fusion occurs, the resulting plasma expands, pushing against the magnetic fields. The company harvests electricity directly from these magnets, a process analogous to regenerative braking in an electric vehicle. By skipping the steam cycle, Helion aims to achieve significantly higher efficiency and a smaller physical footprint for its power plants.

However, this unconventional design is exactly why some physicists remain skeptical. Unlike many of its peers, Helion rarely publishes its findings in peer-reviewed journals, leaving the scientific community with limited data to verify its theoretical claims. CEO David Kirtley has consistently pushed back against this criticism, arguing that the company’s focus is on building functional hardware rather than debating theory in academic papers.

Why Microsoft is Betting on Fusion

For Microsoft and other hyperscalers, the hunger for energy is becoming an existential bottleneck. The massive compute requirements for training and running large language models have created a power demand that traditional grids are struggling to meet. Fusion offers a potential solution: a source of near-limitless, carbon-free energy that doesn't rely on the intermittency of wind or solar.

Investors are clearly buying into this narrative. The fusion sector has seen a flurry of activity in recent months, with companies like Focused Energy and Thea Energy securing significant capital. Even as the timelines remain long by venture capital standards, the potential to disrupt trillion-dollar energy markets is proving too tempting to ignore.

What This Means for the Energy Market

If Helion succeeds, it won't just be a win for Microsoft; it will be a proof-of-concept for a new way to power the world. The 2028 deadline is the critical pressure point. If the company can demonstrate a working, grid-connected plant by that date, it will force a massive re-evaluation of the global energy roadmap.

However, the engineering hurdles remain immense. Moving from a controlled laboratory environment to a reliable, commercial-scale power plant involves solving problems related to material durability, fuel supply, and grid integration that have stalled fusion efforts for decades.

Key Takeaways

  • Aggressive Timeline: Helion is aiming to deliver fusion power to the grid by 2028, significantly faster than most industry projections.
  • Direct Energy Capture: Unlike competitors that use steam turbines, Helion harvests electricity directly from magnetic fields, aiming for higher efficiency.
  • AI-Driven Demand: The funding reflects the urgent need for massive, reliable, and carbon-free energy sources to support the growing power demands of AI data centers.

Helion’s next few years will be defined by the construction of Orion. The company has moved past the phase of theoretical debate and into the phase of industrial execution. By 2028, we will know if the physics of their direct-capture model can hold up under the demands of a real-world power grid.