Helion just became the first fusion company with a government license to flip the switch—and a 2028 deadline to deliver power to Microsoft. On June 16, 2026, the Washington Department of Health issued Helion Energy a Radioactive Materials License and a Radioactive Air Emissions License for its Orion plant in Malaga, Washington, according to the company's announcement.
A fusion company now holds permits to operate, not just experiment. The experiment is over. The construction project has a hard stop.
The license is a regulatory blueprint, not a physics trophy
The licenses confirm Helion has the facilities, trained personnel, and safety programs to meet the state's standards for radioactive materials and airborne emissions. "We are extremely proud to be granted these licenses from the Washington DOH, making us the first company in the world with the regulatory approvals in place for fusion power plant operations," CEO David Kirtley said in the announcement.
Jill Wood, director of DOH's Office of Radiation, framed the moment as a public-health milestone: "Leading radioactive regulatory oversight for the fusion industry in Washington state is an honor and is essential to protecting public health while advancing clean energy."
But the legal architecture matters more than the ribbon-cutting. Congress passed the ADVANCE Act in 2024, codifying the Nuclear Regulatory Commission's decision to regulate fusion under the byproduct-material framework—Part 30, alongside particle accelerators and hospitals—rather than as a fission reactor. Washington state then passed bipartisan legislation in 2024 and 2025 clarifying fusion's role in clean-energy policy and locking in permitting certainty.
The framework is now road-tested. Helion walked through the door first, and the door stayed open.
A half-built plant and a clock that won't stop
Helion began construction in 2025 on land leased from the Chelan County Public Utilities District. Assembly and office buildings are complete. Initial earthwork on the generator building began this spring. In October 2025, Chelan County granted a Conditional Use Permit for the next construction phase.
The Orion plant targets 50 megawatts. For context, that is a massive leap—an order of magnitude beyond Helion's seventh-generation Polaris prototype, which has not publicly demonstrated sustained net energy gain. The licenses confirm the regulatory path. They do not confirm the physics.
The real deadline is non-negotiable. Helion signed a power purchase agreement with Microsoft in 2023 to supply electricity to a Microsoft data center by 2028. Two weeks before the license announcement, Helion closed a $465 million Series G, bringing total investment to $1.5 billion and a valuation of $15.5 billion.
The bet is not just on fusion's future. It is a bet on a specific machine, a specific date, and a specific buyer who does not tolerate failure.
The 2028 fuse and a business model built on a binary outcome
The consensus reads these licenses as a green light for fusion. The consensus is looking at the wrong thing. The licenses impose real operational constraints—radioactive materials handling, air-emissions monitoring, and personnel training—that Polaris has never demonstrated under full-power conditions. The regulatory win masks the engineering gap.
Here is the mechanism that turns that gap into a binary outcome. The PPA with Microsoft was signed two years before Helion broke ground. A standard PPA from a hyperscaler building data centers is not a handshake. It contains a penalty clause for non-delivery, and that clause is a guillotine. If 2028 passes without electrons flowing, the deal collapses. The legal dispute that follows will be ugly, but the balance-sheet damage will be uglier. A startup that misses its sole commercial deadline after burning $1.5 billion does not get to renegotiate. It gets marked to zero by everyone except its own lawyers.
In the next 12 to 24 months, Helion will either deliver net-positive power from Orion or face a valuation implosion below $5 billion. There is no graceful pivot. Series G investors and Microsoft's deadline make the outcome unforgiving.
One plant decides the sector's fate—for a decade
The consequences cascade fast. Follow the money.
If Orion succeeds, the fusion land rush begins in earnest. At least five rivals—Commonwealth Fusion Systems, TAE Technologies, General Fusion, Zap Energy, and Tokamak Energy—will trigger site-permit applications, because a working plant gives their own investors a reason to stop hesitating. The Department of Energy will field a 3x increase in fusion funding requests, because appropriators now have a physical asset to point at, not a white paper. The ADVANCE Act framework, now proven, becomes a template that state regulators copy. Washington becomes the Delaware of fusion incorporation, and every governor wants a piece.
If Orion fails, the entire sector resets. Not gradually. Immediately. The same ADVANCE Act framework that enabled the license becomes a political liability. Regulators who issued permits to a plant that could not deliver will face hearings. The appetite for fusion-friendly rules will evaporate. Private fusion investment will drop an estimated 40% as capital pivots to fission small modular reactors, which carry less physics risk, an already-tested licensing path, and the brutalist reassurance of a supply chain that exists.
More than 40 companies are racing to commercialize fusion. Most will never build a grid-connected machine. Helion is the only one with a license, a PPA, and a hard deadline. That combination makes Orion the sector's bellwether, whether it wants the role or not.
The only milestones that matter now
Three engineering events determine the payout curve.
Generator-building completion. The physical plant must be ready to house a machine that has never run at this scale. Any delay here shortens the commissioning window and increases the pressure on first plasma. No building, no machine. No machine, no power.
First sustained fusion reaction at Orion. This is the moment the physics must hold up its end of the bargain. No public data exists to guarantee this step. It is an act of faith backed by a building that almost exists.
Grid-connection test. Proving 50 megawatts can flow to Microsoft on schedule is the only metric that counts for the PPA. This is not a science experiment. It is a power delivery contract with liquidated damages.
For energy buyers, the playbook is clear. Lock in PPA terms now with developers who have site control and a regulatory path, before the Helion outcome resets everyone's negotiating leverage. Wait for the fallout, and the price moves against you.
For investors, the next 12 months are a due-diligence window. The generator building will either rise on schedule or it will not. First plasma will either happen or it will not. The binary is brutal to watch and more brutal to miss.
For regulators outside Washington, the Orion plant is a living case study. The state's Department of Health just proved that a fusion plant can clear licensing. If the plant works, the template scales, and regulators who hesitated will be asked why they waited. If it fails, every regulator who follows the template will be asked why they ignored the warning.
Helion has the license. The 2028 deadline is now the heartbeat of the fusion industry. In two years, the sector is either grid-connected or it resets. There is no third outcome. One building. One plasma. One deadline. The whole field watches.
