Helion Energy, Inc.
Pre-revenue deep-tech IP + capital-intensive first-of-a-kind power plant; future revenue via merchant power-purchase agreements (PPA) with hyperscalers/utilities. Vertically integrated manufacturing (in-house capacitors, magnets, pulsed-power).
Only Series E/F/G carry disclosed post-money valuations; Series A-D and early 2013-14 grants/seed were never priced-and-disclosed, so omitted. No publicly-disclosed secondary/tender mark.
Earnings, margins, COGS & capex
Revenue trend
Margins
Bull & bear
The single most de-risked path to first commercial fusion: a named hyperscaler offtake (Microsoft), the first-ever fusion operating licenses in hand, a uniquely capital-efficient direct-electricity architecture, and the deepest-pocketed sponsor in tech (Altman) backing a valuation that tripled in 16 months as AI-data-center power demand goes vertical.
- De-risked demand: world's first fusion PPA (Microsoft, May 2023) for ≥50MW by 2028, with Constellation Energy (CEG) as power-marketing partner — a real, penalty-bearing offtake, not a science grant. AI data-center load is the structural tailwind (data-center demand ~1,050 TWh by 2026, hyperscaler capex >$600B in 2026).
- Regulatory first-mover: on Jun 16, 2026 Helion became the FIRST company worldwide to secure fusion operating licenses (Radioactive Materials + Air Emissions Licenses, WA Dept of Health), under the lighter byproduct-material framework codified by the bipartisan ADVANCE Act of 2024 — fusion is NOT regulated like a fission reactor, collapsing a major permitting risk competitors haven't cleared.
- Capital-efficient architecture: the FRC + direct-electricity-capture design is essentially a giant RLC circuit — no steam turbine, no thermal cycle — so a plant is smaller/cheaper than a tokamak. Helion claims a faster, lower-capex path per MW than CFS's SPARC/ARC.
- Real technical progress: Polaris (7th-gen) hit 150 million °C D-T fusion (Feb 13 2026), beating its own prior 100M°C record — first privately-developed machine to demonstrate measurable D-T fusion at that temperature.
- Sponsor + balance sheet: Altman (chairman) + SoftBank Vision Fund 2, Lightspeed, Thrive Capital (Series G lead), Mithril, Nucor, Bill Ford; ~$1.5B raised, $15.5B valuation (Jun 2026) vs $5.4B (Jan 2025) — capital is not the binding constraint in the near term.
- Vertical integration as a moat: in-house capacitors, ultrahigh-pressure ceramics, and now magnets reduce dependence on a China-concentrated pulsed-power supply chain — a strategic + cost edge if it can scale manufacturing.
A pre-revenue science bet priced at $15.5B that has never produced net electricity, on the single hardest fuel cycle in fusion (D-³He), against a 2028 deadline almost no independent expert believes — the classic fusion pattern of slipped milestones now wrapped in a hyperscaler-hype valuation.
- No net energy, ever: as of Dec 2025 Polaris had NOT demonstrated net electricity; the original 2021 plan targeted net electricity in 2024 — missed. No peer-reviewed/independently-verified net-power result exists. The 2028 commercial date is, per multiple experts, not credible.
- Hardest physics in the field: Helion fuses deuterium + helium-3, whose ignition temperature is ~4x (some estimates ~10x) higher than mainstream D-T and whose reaction cross-section is far lower. Critics (incl. retired PPPL physicist Daniel Jassby) have publicly labeled the approach 'voodoo fusion'; even co-inventor John Slough has voiced doubts.
- Helium-3 fuel circularity is unproven at scale: ³He is rare/expensive; Helion's plan is to breed and recycle it from D-D side reactions — an unproven closed loop that must net-produce fuel AND energy simultaneously.
- Schedule pressure is showing: Helion is building a 'Tiny Merge' testbed to de-risk the aggressive Microsoft timeline — a tell that the path to 2028 is not locked. A missed PPA triggers (undisclosed) financial penalties to Microsoft/Constellation.
- Valuation is a venture lottery ticket: $15.5B (~3x in 16 months) on zero revenue and binary technical risk; comps (CFS ~$6.85B raised but pre-electricity; TAE; Pacific Fusion) are ALL pre-net-energy — the entire sub-sector is priced on narrative, not cash flow. A failed Polaris net-energy demo could re-rate the equity down severely.
- Manufacturing + supply-chain risk: highest-cost items (capacitors, pulse-power semiconductors) are supply-constrained and China-concentrated; scaling from one prototype to a fleet of plants is an unproven industrialization, not just a physics, problem.
What it is worth
Venture/strategic last-round mark + scenario-weighted optionality (no DCF possible — pre-revenue, binary technical outcome). Anchored to the Jun-2026 Series G ($465M at $15.5B post-money), cross-checked against private fusion comps (CFS ~$6.85B raised / reportedly $8B+ implied; TAE; Pacific Fusion ~$1B), all pre-net-energy.
<$3-5B (or a down-round / distressed recap) — Polaris fails to show net energy, the D-³He fuel cycle stalls, the 2028 Microsoft date slips with penalties, and a sector funding-winter compresses fusion marks broadly. Equity is a venture option that can lose most of its value on a single failed demo.
~$15.5B (Series G post-money
Jun 4 2026) — the market's current mark, pricing a credible-but-unproven path to first commercial fusion + the Microsoft optionality.
$30B+
Polaris demonstrates verified net electricity and Orion stays on track for 2028; Helion becomes the clear fusion leader and repeats the Microsoft template with other hyperscalers. The valuation re-rates on the first-ever commercial-fusion proof point.
Not financial advice. This is a private, pre-revenue, binary-outcome science bet — the $15.5B mark is a venture price on narrative + AI-power demand, NOT a cash-flow valuation. Conviction is LOW by construction; sizing should reflect total-loss risk.
SWOT
Strengths
- World's first fusion PPA (Microsoft) — a named, penalty-bearing commercial offtake
- First-ever fusion operating licenses (WA DOH, Jun 2026) under the lighter ADVANCE-Act byproduct framework
- Capital-efficient direct-electricity (RLC-circuit) architecture — no steam turbine
- ~$1.5B raised; Altman + tier-1 VC sponsorship; capital not the near-term constraint
- Vertical integration: in-house capacitors, ceramics, magnets reduce China-supply-chain exposure
- Polaris demonstrated 150M°C D-T fusion (Feb 2026) — measurable progress
Weaknesses
- No net electricity demonstrated — the core unproven milestone
- Hardest fuel cycle in fusion (D-³He): ~4x+ higher ignition temperature, lower cross-section
- Helium-3 breed-and-recycle loop unproven at scale
- Long history of missed deadlines (net-power-by-2024 missed)
- Pre-revenue; deep, sustained cash burn with extreme capex
- Industrialization risk — scaling from one prototype to a plant fleet
Opportunities
- AI data-center power demand boom — hyperscalers underwriting nuclear/fusion offtake at scale
- Behind-the-meter / dedicated-power model for data centers (Microsoft template repeatable to Google/Amazon/Meta)
- First-mover regulatory clarity (ADVANCE Act) accelerates US fusion deployment generally
- Direct-electricity architecture could undercut tokamak $/MW if physics works
- Helium-3 / pulsed-power IP licensing and manufacturing as a standalone business
Threats
- A competitor (Commonwealth Fusion's SPARC burning-plasma late-2026/2027) reaching a credible net-energy milestone first
- Polaris failing to show net energy → valuation re-rate + financing freeze
- Microsoft/Constellation PPA penalty + reputational hit on a 2028 miss
- Fusion 'winter' if the sector over-promises and capital retreats
- Supply-chain (capacitors, pulse-power semis) concentration and cost
- Physics simply not closing on the D-³He fuel cycle at energy gain
Moats, dependencies & bottlenecks
Moats
World's first fusion PPA (Microsoft) + first-ever fusion operating licenses Real first-mover assets — a named offtake and regulatory clearance competitors lack. But the PPA is contingent on delivering electricity, and licenses are replicable now that the ADVANCE-Act path is established.
Direct-electricity-capture / FRC architecture IP (no steam turbine) Genuinely differentiated and capital-efficient IF the physics closes; patent-protected. Worthless if D-³He net energy never works — the moat is unproven.
ceramics, magnets) + pulsed-power manufacturing know-how Hard-won manufacturing capability that de-risks a China-concentrated supply chain; durable industrial asset regardless of which fusion fuel cycle ultimately wins.
~$1.5B raised buys runway, but capital access is sentiment-driven and would evaporate on a failed net-energy demo or a fusion-funding winter.
Dependencies
The existential dependency. No fusion company has produced grid net electricity; Helion's D-³He fuel cycle is the hardest. Everything downstream (PPA, valuation) is contingent on this.
customer/commercial The sole anchor demand. A 2028 miss triggers undisclosed penalties and removes the commercial proof point; over-concentration in one customer.
³He is rare/expensive; the plan to self-breed it from D-D side reactions and recycle it is unproven at energy-positive scale.
Highest-cost items, China-concentrated; mitigated by in-house capacitor/magnet manufacturing but switches still bought externally (Polaris needs ~50,000).
Pre-revenue with extreme capex; depends on sentiment-driven follow-on rounds. Funded for now (~$1.5B) but burn is large and binary-event-sensitive.
Largely de-risked — licenses granted Jun 2026; the byproduct-material framework is codified and bipartisan.
Advantages
- Only fusion company with a named, penalty-bearing hyperscaler PPA (Microsoft)
- First and only fusion operating licenses in hand (WA DOH, Jun 2026)
- Capital-efficient direct-electricity architecture (no thermal cycle / turbine)
- Vertical manufacturing reduces China-supply-chain exposure
- Deepest sponsor/capital network in private fusion (Altman, SoftBank, Thrive, Nucor)
- Physical construction underway (Orion, Malaga WA)
Weaknesses
- Has never produced net electricity — the core claim is unproven
- Chose the hardest fuel cycle in fusion (D-³He)
- History of missed deadlines (net-power-by-2024 missed)
- Single-customer concentration (Microsoft)
- Valuation (~$15.5B) wholly narrative-driven on zero revenue
- Helium-3 fuel loop unproven at scale
Bottlenecks
- Net-energy physics on the D-³He fuel cycle — the single gating unknown
- Helium-3 fuel breeding/recycling at energy-positive scale
- Industrializing from one prototype to a fleet of plants (manufacturing, not just physics)
- Pulse-power semiconductor + capacitor throughput (Polaris alone = ~50,000 semis, ~150 containers of capacitors)
- Meeting the 2028 Microsoft delivery date — schedule risk flagged by the 'Tiny Merge' de-risking testbed
Top signals & trends
Top signals
The whole thesis hinges on a verified net-energy result. Until then it is a science bet at a $15.5B price.
Capital availability + tier-1 validation; also a sign the market is pricing fusion on narrative + AI-power demand.
Removes a regulatory unknown competitors still face; fusion not licensed like a fission reactor.
An extra de-risking step is a tell the 2028 date is under schedule strain.
Moves from PowerPoint to poured concrete; real capital and permitting commitment.
Best-funded rival; a credible CFS net-energy result would re-order the leadership narrative.
Asymmetric downside if the schedule slips; magnitude not public.
Structural demand that makes a fusion MW commercially valuable if/when it arrives.
Trends
Hyperscaler capex >$600B in 2026, data-center demand ~1,050 TWh; makes a future fusion MW commercially valuable and bankable via PPA — the reason Helion's valuation tripled.
Collapses permitting risk vs fission; Helion is the first to monetize this with granted licenses.
Abundant capital today, but a single high-profile failure could trigger a fusion-funding winter that hits the most aggressive timelines (Helion's) hardest.
Fission SMRs are nearer-term and compete for the same offtake dollars; fusion must prove it before the data-center power gap is filled by other sources.
The whole sector remains pre-commercial; narrative valuations are exposed to the first real net-energy data point — in either direction.
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
Power semiconductors / pulse-power switches — Helion buys semiconductor switches externally (its highest-cost item alongside capacitors); IGBT/MOSFET-class device makers are the relevant supply base.
Power-semiconductor supplier class for pulsed-power switching (Polaris needs ~50,000 pulse-power semiconductors).
Steel producer and Helion investor — high-quality metals for plant + magnet structures; strategic supplier-investor.
Vertically integrated: makes its own capacitors, ultrahigh-pressure ceramics, and increasingly its own high-power magnets to reduce China-concentrated pulsed-power supply-chain risk.
World's first fusion PPA (May 2023): ≥50MW from Orion by 2028 for data centers. The anchor customer and the entire near-term revenue thesis.
Investor + intended industrial offtake partner — steel mills as a future fusion-power customer beyond data centers.
Best-funded private fusion (~$6.85B raised through May-2026 $3.85B round). HTS-magnet tokamak SPARC ~75% built, targeting burning plasma late-2026/2027; commercial ARC early-2030s. The benchmark rival; a credible SPARC result would challenge Helion's first-mover narrative.
~$1.79B raised; beam-driven field-reversed-configuration (closest architectural cousin to Helion's FRC) also pursuing aneutronic p-B11/advanced fuels. Direct technical competitor on the FRC path.
~$1B raised; pulsed-magnetic-inertial approach — architecturally adjacent to Helion's pulsed-power concept. Newer but well-capitalized.
$327M raised; sheared-flow Z-pinch — no magnets, low-cost pulsed approach competing for the 'cheap, fast' fusion narrative Helion also claims.
$269M raised; stellarator path, partnering with TVA on a US deployment site — competes for the 'first US fusion plant' headline.
~$1B raised; phased fusion (isotopes/neutron sources first, energy later) — revenue today from non-energy products, a different commercialization model.
Public SMR/fission (not fusion) but competes for the same hyperscaler data-center offtake dollars; a public comp for 'pre-revenue advanced-nuclear at a large valuation.'