FAQ
Same questions also surface inline on the press page. Answers are written for a technically literate non-specialist; the underlying technical detail lives in the data room and the public whitepaper.
01.What is Laurelin in one sentence?
Laurelin is a containerized pulsed deuterium–deuterium fusion reactor program — a hardware company building a forty-foot transportable fusion module with electromagnetic direct conversion and an AI control model, founder-led by a team of six engineers in San Francisco.
02.What have you actually built?
A complete reactor design. The architecture is documented end-to-end: subsystem blueprints, the patent claim map, the procurement plan, and the white paper that ties them together. Bench articles for representative subsystems run upstream of integrated build, and an AI control model is in development to optimize reactor controls in real time. Public-tier program facts and the subsystem index live under the Program tab; the technical data room is available on request and shared on a tier basis.
03.Why a containerized 2H–2H reactor and not a campus-scale 2H–3H machine?
2H–3H cross-section is easier; 2H–3H engineering is not. A 2H–3H machine inherits a tritium plant, a breeding blanket, and a tritium-inventory accounting regime — none of which fit in a forty-foot envelope. Deuterium is in seawater and is not a special nuclear material. The architecture pays the cross-section penalty in exchange for an engineering integral that closes inside a transportable container.
04.Why now?
Three factors converge: (1) hyperscaler datacenter load is growing faster than transmission, forcing on-site firm power; (2) component cost curves for high-field magnets, fast switches, and pulsed-power capacitors have compressed enough to put a containerized 2H–2H machine inside a venture capex envelope; and (3) US fusion regulation has been formally separated from fission — the NRC is regulating fusion machines under the 10 CFR Part 30 byproduct-material framework, with a dedicated fusion-machine rulemaking proposed in February 2026. The combination makes a behind-the-meter commercial module viable for the first time.
05.How is this different from CFS, Helion, TAE, Zap, or Avalanche?
Most compact-fusion incumbents target campus-scale machines aimed at grid-scale power output, with siting that requires civil works, transmission interconnect, and utility-scale balance of plant. Laurelin starts from a forty-foot containerized envelope, runs on 2H–2H (no tritium plant or breeding blanket), treats electromagnetic direct conversion as a first-class architectural element rather than a downstream afterthought, and integrates an AI control model from the design stage. We are not racing to first plasma at scale; we are racing to a transportable, behind-the-meter module.
06.Why this team?
Joe Finberg authored the white paper, the subsystem blueprints, the patent claim map, and the procurement plan that the company is built on. Boruch Epstein (DPhil Oxford, materials science) leads engineering and runs the bench upstream of every measurement. Six engineers total, founder-led, headquartered in San Francisco. The team is organized around iteration speed and falsifiability, not around peer-review or institutional procurement.
07.What is the role of the AI control model?
Pulsed FRC operation is a real-time control problem with very tight timing budgets. We are building an AI control model whose explicit job is to optimize the reactor control loop — formation, merge, compression, and recovery timing — under live diagnostics. The model lives in the architecture from day one rather than being bolted on after first plasma. This is part of why a small team can credibly operate a machine that historically required a control-room of operators.
08.What is the business model?
Behind-the-meter firm power for large constant electrical loads — datacenters first, then microgrid power, remote and forward operations, and other off-grid sites where grid extension is uneconomic. Module sale or power-purchase agreement, not utility wholesale. The container form factor is the product, and the deployment model is unit economics, not power-plant economics. Operators sizing future load can register early interest.
09.What is the regulatory path?
Following SRM-SECY-23-0001 (April 2023), the NRC regulates fusion machines under the 10 CFR Part 30 byproduct-material framework rather than the Part 50/52/53 utilization-facility framework that governs fission. A dedicated fusion-machine rulemaking was published in the Federal Register in February 2026 implementing Section 205 of the 2024 ADVANCE Act. The compliance burden is materially lower than for a fission reactor of equivalent thermal output, and the licensing path is one a venture-backed company can carry. State-level radiation-control engagement is on the roadmap.
10.What is the safety and waste profile?
2H–2H fuel does not require an external tritium supply chain or a breeding blanket. The neutron spectrum from a 2H–2H machine is softer and lower-flux than a 2H–3H machine of equivalent fusion power, which materially reduces the activation budget and the shielding integral against the as-built container geometry. There is no fission inventory, no meltdown failure mode, and no long-lived high-level waste stream. We do not claim tritium-free — the 2H–2H reaction has a branch that produces tritium, which is managed in-situ.
11.What is the defensibility story — IP, integration, supply chain?
The patent claim map covers subsystem-level inventions and the integration that ties them into the container envelope. Beyond patents, the moat is engineering integration: the architecture is co-designed across coil families, pulsed-power switching, direct-conversion geometry, shielding, and the control model, in a way that cannot be replicated by buying any single subsystem off the shelf. Most components are commoditizable; the integration is not. The control model and the procurement-and-build playbook compound over time.
12.What partners and customers are you working with?
Founder-led partnership conversations are ongoing across hyperscale datacenter operators, defense and forward-operations buyers, and selected national-lab collaborators on diagnostics and shielding. We do not disclose specific counterparties on the public site. Inquiries via the contact form.
13.What does the next 12-24 months look like?
Continued bench characterization of representative subsystems, build-out of the diagnostic stack, training and validation of the control model on bench data, and a small number of pre-declared technical demonstrations that a non-specialist can verify. Specific measurement criteria and timelines are shared with counterparties under mutual NDA.
14.What does the technology data room contain?
Subsystem blueprints, the full technical roadmap, the patent claim map, the procurement plan, control-model architecture, and the measurement criteria for each program stage. The data room is available on request; specific reactor details inside it are shared under mutual NDA on a tier basis.