Thorium Atomics · Tesseract TGR· High-temperature gas-cooled reactor
The Tesseract TGR is a helium-cooled pebble-bed reactor producing roughly 100 megawatts of firm electricity, or 750 °C process heat for the industries electrification cannot reach. It runs on TRISO fuel below 10 percent enrichment, so it does not depend on high-assay fuel supply, and a thorium blanket around the core breeds fresh fuel during operation, cutting lifetime mined uranium by roughly half against a comparable light-water reactor.
Roughly a quarter of global energy demand is industrial process heat, and most of it is burned rather than generated. Electrification handles the low end of that range well. Above about 400 °C the list of practical clean options gets very short, and above 700 °C it nearly disappears. That band is what this reactor was sized for.
| Band | Representative processes | Clean supply available today | |
|---|---|---|---|
| < 150 °C | Space heat, drying, food | Heat pumps and resistive electrification work well here. | |
| 150–400 °C | Paper, chemicals, refining | Clean options exist. Cost and scale are uneven. | |
| 400–700 °C | Hydrogen, ammonia, glass | Very few credible clean options at industrial scale. | |
| 700 °C + | Cement, steel, high-temperature process | Almost nothing clean serves this band today. |
Helium leaves the core at 750 °C and returns at 250 °C. That outlet temperature is high enough to drive chemical process heat directly, and high enough for an efficient power conversion cycle when the customer wants electricity instead. One reactor design serves both, and the choice is made at the site rather than in the reactor.
Helium is inert and single-phase. It does not change state, does not react with the fuel or the structure, and does not become strongly radioactive in the primary loop. That removes an entire category of accident from the design problem before analysis starts.
The arrangement, in schematic section. Concentric from the centre outward: the driver bed, the thorium blanket, the graphite reflector carrying the rod and riser channels, the core barrel, and the pressure vessel. Dimensions and channel counts are withheld.
| 01 | Reactor class | High-temperature gas-cooled, pebble bed |
| 02 | Thermal output | 250 MWth |
| 03 | Electrical output | ~100 MWe |
| 04 | Coolant outlet / inlet | 750 °C / 250 °C |
| 05 | Coolant | Helium. Inert, single-phase, chemically inactive |
| 06 | Fuel form | TRISO particles in graphite pebbles |
| 07 | Driver enrichment | Below 10 percent uranium-235 |
| 08 | Core arrangement | Uranium driver bed with a thorium blanket |
| 09 | Mined uranium over life | Roughly half a comparable light-water reactor |
| 10 | Design life | 60 years |
Instead of fuel assemblies fixed in place for years and replaced during an outage, the core is a slowly moving bed of graphite spheres. Pebbles are added at the top and withdrawn from the bottom while the reactor runs.
Each pebble holds many thousands of TRISO particles. Each particle is a fuel kernel a fraction of a millimetre across, wrapped in a porous carbon buffer, two layers of pyrolytic carbon, and a shell of silicon carbide. The silicon carbide layer is the engineered barrier. It holds fission products inside the particle at temperatures far above anything the reactor reaches in normal operation.
The consequence is that radioactive material is retained at the scale of a grain of sand rather than at the scale of a building. TRISO fuel has been manufactured and irradiated under United States Department of Energy programmes for decades, so this is a qualification path rather than an invention.
Pebbles leave the bottom of the bed through the defuelling chute, pass a measurement station that reads how much of their fuel has been consumed, and are then either lifted back to the top of the vessel or retired. Fuel that still has life in it keeps circulating. Fuel that does not is removed while the reactor stays at power.
Because fuel is replaced continuously rather than in large batches, the core does not need to be loaded with a large excess of reactivity at the start of a cycle to survive until the next outage. The core sits closer to its working condition for its whole life.
Every reactor loses neutrons out of the edge of its core. The Tesseract puts an annulus of thorium-bearing pebbles in that path. Thorium-232 is fertile rather than fissile: it does not sustain a chain reaction on its own, but when it captures a neutron it becomes uranium-233, which does. Fuel is therefore being made inside the reactor while power is being made.
This is a near-breeder, not a breeder. The reactor does not make all of its own fuel and we do not claim that it does. What it does is displace a substantial share of the uranium that would otherwise have to be mined, converted and enriched, which on the current reference design is roughly half the lifetime mined uranium of a comparable light-water reactor of the same output.
Thorium is also about three to four times more abundant in the Earth’s crust than uranium, and much of it already exists as a stockpiled by-product of rare-earth mining. The fuel resource is not the constraint.
Reactors that run thorium accumulate thorium-229 as a matter of physics. Thorium-229 is the parent of actinium-225, one of the scarcest and most sought-after isotopes in medicine. We did not design the reactor to make it. The reactor makes it anyway.
Targeted alpha therapy attaches an alpha-emitting isotope to a molecule that binds to a cancer cell, delivering a dose over a range of a few cell diameters. Actinium-225 is the leading candidate for it. Global supply comes largely from a small inventory of legacy thorium-229 held at a handful of national laboratories, and that supply has been the constraint on clinical programmes for years.
A reactor running a thorium cycle builds thorium-229 inventory as an ordinary consequence of operation. Recovering it is a fuel-cycle and separations problem rather than a reactor problem, and it is not on the critical path for the energy business.
We do not present this as the reason to build the reactor, and we do not put a number on it. Energy comes first. This is what sits underneath, at no additional design cost, because of the fuel choice already made.
The safety case rests on physical properties of the fuel, the moderator and the geometry rather than on systems that have to work. The formal analysis that supports a licence application is still ahead of us, and is listed as such on Sheet 09.
As the fuel heats up, the reaction slows down. This is a property of the fuel and the moderator, not of a control system, so it does not depend on power, signals, or an operator being present.
The core spreads 250 MWth across a large graphite mass. Graphite holds a great deal of heat and gives it up slowly, so the core takes hours to change temperature rather than seconds.
The vessel is slender enough that residual heat can reach the surrounding structure by conduction and radiation alone. Cooling the core after shutdown does not require pumps, power, or intervention.
Helium does not boil, does not burn, does not react with graphite or steel, and does not become strongly radioactive. Losing coolant pressure is a loss of cooling, not the start of a chemical event.
The twelve control rod channels sit in the reflector rather than in the fuel region, so reactivity control does not depend on penetrating the moving bed.
The silicon carbide shell around each fuel kernel retains fission products well above the temperatures the reactor reaches, including in the accidents the design is analysed against.
A compact, self-contained unit that can sit inside an industrial fence line, next to a data centre campus, or in a utility corridor. The reactor does not change between these cases. What changes is whether the output is taken as heat or as electricity.
750°C helium delivered directly to steel, cement, ammonia, hydrogen, and chemical production. Co-located with industrial facilities under long-term take-or-pay heat offtake agreements.
Behind-the-meter clean electricity for campuses where uptime, energy density, and carbon commitments are non-negotiable. No transmission constraints. No intermittency.
Firm, dispatchable electricity for municipal utilities, rural cooperatives, and growing communities. Compact footprint. Designed to be a neighbor, not a facility.
Resilient power and heat for critical facilities, defense installations, and strategic supply chains where fuel independence and long-duration reliability matter most.
Reactor programmes are slow largely because every design change is expensive to evaluate. We build and analyse the reactor in software first, so that trade-offs surface early, the reasoning is preserved, and the engineering record is assembled as the design is made rather than reconstructed afterwards.
Core geometry, fuel loading, enrichment and the blanket configuration were each decided by analysing the alternatives rather than by argument. The sensitivity of the design to each choice is on record, which is what lets us say plainly which parameters are locked and which are still open.
Each figure in the design basis traces to a specific analysis with recorded inputs, code version and date. A regulator and an engineering contractor both need that provenance, and it is far cheaper to keep from the beginning than to reconstruct at the end.
Drawings and the design basis are generated from the same source as the analysis, so the engineering record and the design cannot quietly drift apart as the work advances.
Our own tools inform the design. The structural, thermal-hydraulic, seismic and fuel-performance analyses that support a licence application are performed in established qualified codes, by organisations with the quality programmes to stand behind them.
A company is a revision history. This is ours, current to July 2026, with the parts that are not finished listed at the bottom rather than left out.
| Rev | Date | Description | Ref |
|---|---|---|---|
| A | 2026 · 03 | Reference concept independently reviewedReviewed by Dr. Ayman Hawari of Texas A&M University, whose work spans reactor physics and TRISO fuel systems, and found technically feasible. | External review |
| B | 2026 · 03 | United States provisional patent application filedCovering the reactor architecture. Patent pending. Further applications are in preparation with counsel. | Patent pending |
| C | 2026 · 05 | Reference design locked250 MWth, TRISO driver fuel below 10 percent enrichment with a thorium blanket. Confirmed by a multi-cycle depletion analysis run across the full fuel cycle. | 250 MWth |
| D | 2026 · 06 | Pre-application engagement opened with the U.S. Nuclear Regulatory CommissionNotice of intent submitted 23 June 2026 under 10 CFR Part 53, the NRC's technology-inclusive framework for advanced reactors. The Canadian pathway, a CNSC Vendor Design Review, is maintained in parallel. | 10 CFR Part 53 |
| E | 2026 · 07 | NRC project number assignedThe NRC assigned project number 99902174 and a project manager in the Office of Advanced Reactors. A regulatory engagement plan is in preparation. | No. 99902174 |
| F | 2026 · 07 | Design basis and preliminary drawing set issuedA documented design basis for the reactor island and module systems, issued together with an open-items register that states plainly what has not yet been analysed. Constructability and road transport of the major components were assessed against real delivery constraints rather than assumed. | TAI-DBD-001 |
| G | 2026 · 07 | Independent verification and demonstration siting in progressA Strategic Partnership Project agreement with Idaho National Laboratory is in preparation for Department of Energy review; the first task cross-checks our reactor physics results in MCNP and Serpent. A separate demonstration application was submitted to INL's Nuclear Energy Launch Pad programme. | INL |
Code-qualified structural, thermal-hydraulic and seismic analysis, the probabilistic risk assessment, and fuel qualification testing are all ahead of us. No prototype has been built. Nothing on this sheet should be read as a licensed or constructed design.
During my life I have witnessed extraordinary feats of human ingenuity. I believe that this struggling ingenuity will be equal to the task of creating the Second Nuclear Era. My only regret is that I will not be here to witness its success.
The second nuclear era Weinberg described is not abstract. It begins with decisions made now, by people with the physics and the industrial will to build.
Between them the team has led nuclear engineering at Ontario Power Generation, run the Tennessee Valley Authority, licensed new build for the Canadian regulator, and worked the thorium fuel cycle at Oak Ridge.
Four decades of nuclear and power industry experience. Former President and CEO of the Tennessee Valley Authority and of Ontario Power Generation, two of North America's largest nuclear-powered utilities. Prior senior leadership at Duke Energy, Progress Energy, and Chicago Bridge & Iron Power, and senior technical and management roles at the U.S. Nuclear Regulatory Commission. Currently on the Board of Directors for Dominion Energy and Aecon Group.
Founder of Algonquin Power and a veteran builder of public-market energy platforms, he helped scale the company from its initial listing into a multi-billion-dollar infrastructure enterprise. Decades of experience in board governance, capital formation, strategic transactions, and long-horizon company building across regulated power markets and energy infrastructure.
Over 30 years of leadership in nuclear safety and licensing from Ontario Power Generation (OPG) and Atomic Energy of Canada Ltd. (AECL), including as VP of Nuclear Regulatory Affairs at OPG. Spearheaded licensing efforts spanning licence renewals for operating plants and site preparation / construction licensing for new builds. Canada's representative at IAEA member state meetings on nuclear safety. Ph.D. Mechanical Engineering, University of Toronto.
Founder and President of Spectra Tech Inc., a nuclear services firm based in Oak Ridge, TN with ~350 employees across multiple DOE sites and national laboratories. Ph.D. Nuclear Engineering, University of Tennessee, under Dr. Bernard Spinrad (inventor of the Molten Salt Reactor Experiment). Direct thorium fuel cycle experience, including five years at ORNL Building 3019 (the U-233 repository), and Fort St. Vrain spent-fuel management for DOE since 2016.
Co-architected the Tesseract simulation platform and helped originate the company's strategic direction. Leads federal engagement and long-term platform positioning. Over two decades prior in institutional markets, including senior roles at RBC Capital Markets and CIBC World Markets. Co-founded the company around a simple conviction: the future of nuclear will belong to platforms that solve for industrial heat, fuel resilience, and energy sovereignty, not electricity alone.
Former Chief Nuclear Engineer and VP, Engineering and Modifications at Ontario Power Generation. Led nuclear engineering, major modification programs, and new-build technical strategy across one of North America's most demanding fleet environments. Deep experience in reactor design, core physics, refurbishment, life extension, and modernization of critical nuclear systems.
Helped design and build the company's internal simulation and development environment, spanning reactor engineering workflow, Monte Carlo simulation integration, data architecture, and regulatory compliance automation. Central role implementing and refining the computational platform that underpins Thorium Atomics' technical strategy.
15+ years of public-company finance experience across mining, precious metals, and energy. Director of Treasury & Planning at Americas Gold and Silver for nearly a decade; CFO and corporate controller roles at multiple TSX-listed uranium and resource companies including IsoEnergy, Premier American Uranium, and Western Metallica. Founder of Numbers First Inc.
Capital markets strategist with 30+ years driving growth across public and private companies. Previously Managing Director and Head of Institutional Equity Sales & Trading at Desjardins Securities. Central to the company's formation and early-stage capital strategy.
15 years in capital markets and corporate development. Founder of Thesis Capital, a Canadian merchant bank. Entrepreneur with a track record of founding, scaling, and exiting ventures in the technology sector. Advises and finances growth-stage companies and provides strategic transaction expertise and go-public strategies in the North American markets.
PhD Physics, Stanford. Thorium breeder modeling and Monte Carlo neutronics. Published in Nuclear Technology. Fusion neutron source development at Adelphi Technology.
PhD Nuclear Engineering, Texas A&M. Licensed CANDU reactor operator. Thorium breeder reactor design and fuel cycle modeling. PRA across 11 nuclear power plants.
Westinghouse safety advisor for AP1000. Senior Reactor Operator, Braidwood & AP1000. 19 years at Exelon; thermal-hydraulics and accident analysis.
We are a small team and we read everything. Tell us which part of this you care about and we will reply with something specific rather than a brochure.
Process heat, siting, and offtake structures.
Firm power, grid support, deployment pathways.
Verification, technical review, public-sector collaboration.
If the engineering interests you and you want to understand the company, write to us.
Collaboration, peer review, or an interest in working on this.