A 5,000-amp superconducting cable, cooled by liquid nitrogen, sits on a bench in a San Francisco lab. It is the heart of a bet that a fusion reactor can power a ship. It is also, according to its creators, a product ready for the AI datacenter market. For Maritime Fusion, the cable is both the wedge and the wallet.
The startup, founded in 2024, raised a $4.5 million seed round last November. The round was led by Trucks VC and included Paul Graham, Alumni Ventures, Aera VC, and Y Combinator [PR Newswire, Nov 2025]. The capital is split between two missions. The first is commercializing the patent-pending SHIELD superconducting cable. The second is developing the Yinsen, a low-power-density tokamak fusion reactor designed for maritime propulsion and off-grid power [PERPLEXITY SONAR PRO BRIEF].
The Dual-Use Wedge
The SHIELD cable is the company's core technical innovation. In a bench test, it carried 5,000 amps at 77 Kelvin, a temperature achievable with liquid nitrogen [PERPLEXITY SONAR PRO BRIEF]. The company claims its design can handle up to 8,000 amps and is engineered for the extreme conditions inside a fusion reactor [maritimefusion.com/blog, 2026].
This cable is intended to form the magnets for Maritime Fusion's Yinsen reactor. But the company does not plan to wait for fusion to find a market. It will sell the same SHIELD cable for high-power distribution in AI datacenters, a sector desperate for dense power transmission solutions [PERPLEXITY SONAR PRO BRIEF]. This creates a potential near-term revenue stream to fund the long-term reactor development, a classic deeptech hedge.
Why Ships First
The company's reactor strategy hinges on a specific market niche. Grid-scale fusion requires massive power output and near-perfect reliability. Maritime Fusion is targeting marine and off-grid applications, which it argues need 10 to 15 times less power and can tolerate lower uptime [PERPLEXITY SONAR PRO BRIEF].
The flagship Yinsen reactor is designed to generate around 30 megawatts of electricity, a scale suited for large commercial vessels or remote industrial sites [TechCrunch, Nov 2025]. The first unit is projected to cost approximately $1.1 billion to build and be operational by 2032 [LinkedIn, Felipe Amoroso Manzano, 2026]. The bet is that solving for a simpler use case provides a faster, cheaper path to a first commercial fusion device.
The Team Behind the Magnet
The founding team, Justin Cohen and Jason Kaufmann, bring backgrounds from SpaceX and Tesla [PR Newswire, Nov 2025]. Cohen is a nuclear engineer and plasma physicist with experience at Columbia University and the Princeton Plasma Physics Laboratory [BEAMSTART, 2026]. Kaufmann also hails from the SpaceX and Tesla engineering ranks [bestofshowhn.com, 2026]. As of their Winter 2025 Y Combinator batch, the team size was six [Y Combinator, Nov 2025].
Their collective resume points to experience in high-performance hardware and systems engineering, a relevant skillset for building complex superconducting magnets. It is less clear on direct experience in bringing a nuclear fusion device to regulatory approval and commercial deployment, a multi-decade challenge that has eluded well-funded giants.
The Capital and the Clock
The $4.5 million seed is a starting pistol. The roadmap to a 2032 reactor implies a capital journey measured in the hundreds of millions, if not billions, long before the first watt is sold. The table below outlines the disclosed funding and the projected next major capital event.
| Round | Amount | Lead Investor | Key Participants | Purpose / Context |
|---|---|---|---|---|
| Seed (2025) | $4.5M | Trucks VC | Paul Graham, Alumni Ventures, Aera VC, Y Combinator, angels | Lab build-out, initial cable development, reactor concept design [PR Newswire, Nov 2025] |
| Projected Next Round | ~$20-50M (estimated) | N/A | Likely deep-tech / climate funds | Scaling cable manufacturing, advancing reactor engineering toward a defined milestone |
The company's ability to attract subsequent funding will depend heavily on two signals. First, commercial traction for the SHIELD cable in the datacenter market. Second, technical milestones that de-risk the reactor physics, moving from bench tests to integrated subsystem demonstrations.
Where the Wheels Could Come Off
The fusion sector is a graveyard of ambitious timelines and technical dead ends. Maritime Fusion's plan faces several credible pressures.
- The Capital Chasm. A $1.1 billion reactor build is a venture-scale outcome, not a venture-scale check. Bridging from a $4.5 million seed to that figure requires a series of increasingly large, high-conviction rounds in a sector where investor patience is historically thin.
- The Dual-Market Distraction. Selling cables to datacenters and building a fusion reactor are fundamentally different businesses with different customers, sales cycles, and regulatory environments. Success in one does not guarantee success in the other, and resource allocation will be a constant tension.
- The Proven Competition. The company lists Commonwealth Fusion Systems and Type One Energy as competitors. Both are better capitalized and further along in developing HTS magnet technology for fusion, though focused on the grid. Maritime Fusion's maritime focus is its differentiator, but it does not escape the underlying physics and engineering hurdles.
The company's most plausible answer is that the cable business is not a distraction but a validator. Proving the SHIELD technology in a demanding commercial market de-risks the core component for investors. It also generates revenue that can extend the runway, buying more time for the reactor team.
The Next Twelve Months
Watch for three things. First, a named datacenter customer or partnership for the SHIELD cable. Second, a subsequent funding announcement, likely framed as a Series A, within the next 18 months. Third, a more detailed technical paper or demonstration moving beyond the 5,000-amp bench test to a prototype magnet coil.
The $4.5 million from Trucks VC and Paul Graham is a vote of confidence in a team with a compelling hardware pedigree. The question for 2026 is whether the cable can carry enough commercial current to keep the reactor dream alive. Can a product for AI power distribution fund a prototype for fusion-powered shipping? The first answer will come not from a plasma physics journal, but from a purchase order.
Sources
- [PR Newswire, Nov 2025] Y Combinator-backed Maritime Fusion Raises $4.5M From Trucks VC, Paul Graham | https://www.prnewswire.com/news-releases/y-combinator-backed-maritime-fusion-raises-4-5m-from-trucks-vc-paul-graham-302624213.html
- [maritimefusion.com/blog, 2026] SHIELD cable technical specifications | https://www.maritimefusion.com/blog/
- [TechCrunch, Nov 2025] Exclusive: This startup wants to build a fusion reactor, on a boat | https://techcrunch.com/2025/11/24/this-startup-wants-to-build-a-fusion-reactor-on-a-boat/
- [BEAMSTART, 2026] Justin Cohen background | https://www.beamstart.com/
- [bestofshowhn.com, 2026] Jason Kaufmann background | https://bestofshowhn.com/
- [Y Combinator, Nov 2025] Maritime Fusion company profile | https://www.ycombinator.com/companies/maritime-fusion