Nuclear Energy

Sequoia partners Sean Maguire (physics PhD) and Liam (Harvard physics, Olympic gold, nuclear industry) convinced the partnership to invest $300M in a nuclear venture (Isaiah's company) despite multiple technical risk layers, betting on domain expertise and execution.

South Park Commons portfolio includes companies building nuclear-powered ships with 18-month milestone targets for Series A; reflects broader hard tech ambition shift toward energy infrastructure.

Berkeley physics spinout Deep Vision ($150M raised) is drilling test holes for a 'gravity reactor' a mile underground in Parsons, KS, aiming to generate electricity by 2027-28 — an astonishingly fast timeline if successful, but faces local NIMBY opposition despite remote location.

NASA's SR1 Freedom spacecraft repurposes Gateway power element and INL reactor components for a near-term nuclear electric propulsion demo, enabling faster outer solar system missions and Mars return without cryogenic refueling miracles.

Coastal Ventures' incubation of Commonwealth Fusion demonstrates fusion's transition from science project to investable hard tech. Technical edge (high-temp superconductors) + regulatory tailwinds + energy demand from AI data centers create convergent opportunity.

Tiny nuclear reactors for resilient military base power are moving from concept to funded prototypes via DOE acceleration programs and venture capital, representing a new dual-use nuclear investment category.

Valar Atomics aims to become the world's largest energy company by vertically integrating nuclear reactor manufacturing and deploying at unprecedented hardware speed — from one reactor per year to monthly — to power AI infrastructure directly, having already demonstrated a reactor powering an Nvidia Blackwell chip.

The administration has set a 2028 deadline for a space-qualified nuclear reactor (for Mars transit propulsion) and notes unprecedented private capital flowing into fusion (37 venture-backed firms), making fusion a credible near-term energy source rather than a perennial 50-year promise.

Bipartisan regulatory reform enables new nuclear ventures; Valor's philosophy of extreme simplicity (off-the-shelf parts, in-house manufacturing) and rapid iteration mirrors SpaceX's path from Falcon 1 to Starship, targeting gigawatt-scale sites to deliver cheapest energy for AI factories and heavy industry.

China's Institute of Plasma Physics has installed a 582-ton superconducting magnet for a fusion reactor that achieved 30-minute sustained plasma at 100M°C, demonstrating progress toward gigawatt-scale fusion power from seawater; this industrial-scale advancement is underestimated by Western observers focused on solar.

First nuclear-powered AI demo (Blackwell chip), vertically integrated reactor manufacturing aims to scale from 1/year to 1/month, backed by Sequoia and JPM credit facility, nuclear can be co-located with large data centers.

Administration targeting nuclear reactor in space by 2028 with enough propulsion power for Mars transit; part of bold national missions to inspire STEM talent and demonstrate American technological leadership.

Bipartisan regulatory reform and founder intensity are unlocking nuclear startup scaling; Valor's simple commodity design physics-based safety approach mirrors SpaceX Falcon 1 to Starship progression; gigasite model targets cheapest energy for AI and industrial use.

Fusion physics works but economic viability unknown; current VC-funded fusion startups face existential funding risk from finicky investors. Fuse's three-step plan uses cash flow from integrated energy business (steps 1-2) to fund fusion bet (step 3) without company-existential risk.

The nuclear deterrent exemplifies why European defense integration stalls: France will not share nuclear command authority with Germany, making true industrial rationalization politically impossible regardless of economic logic.

Advanced nuclear approaches promise energy abundance but are blocked by vetocratic regulatory regimes; overcoming this is a prerequisite for AI-era energy demands.

The Wendelstein 7-X stellarator achieved full-power operation in 2022, retiring the last major physics risk (high-energy particle confinement). Proxima now pursues an engineering-first path: SMC full-scale HTS magnet by end-2027, then Alpha net-energy device, then Stellaris commercial plant. Steady-state operation avoids tokamak disruptions, enabling dispatchable baseload.

Marvel Fusion investment reflects need for European energy independence after German energy policy decisions created vulnerabilities exposed by geopolitical conflicts.

1 GW of aggregated distributed flexibility (batteries, EVs, heat pumps) equals the capacity of a modern nuclear power plant. The 20 GW target (20 nuclear equivalents) can be reached via software optimization of existing assets, avoiding decade-long construction and capex.

Microsoft's Three Mile Island restart and 20-year fixed-price PPA signals a shift: electricity is becoming a denominated asset class where long-term supply contracts are more valuable than bonds, creating durable demand for nuclear and independent power producers.

Founders with SpaceX-style aggressive execution are now tackling small modular nuclear reactors in Europe, a category that was unimaginable a decade ago, driven by urgent power constraints from AI data centers.

Rising electricity prices from data center demand could accelerate fusion timelines; Altman predicts first profitable fusion reactor within 5 years, earlier than consensus.

Nuclear power is undergoing a renaissance largely because AI data centers require reliable, carbon-free baseload power that intermittent renewables cannot provide, making nuclear a structural complement to AI infrastructure build-out.

100,000 tons of commercial spent fuel retains 97% energy content; policy shift turned waste liability into asset with 27 governors competing for fuel cycle campuses, unlocking domestic energy for AI data centers.

Proxima Fusion's stellarator approach using high-temperature superconducting magnets can demonstrate net energy with ~500M equity by sharing demonstration reactor costs with governments, compressing timelines from decades to years through AI simulation and European supply chain leverage.

Isaiah Taylor explains Valar's high-temperature gas reactor generates electricity without water cooling, partnering with Nvidia to enable scalable AI data centers independent of local water resources.

SMRs remain unproven at scale; only AP1000 large reactors with known supply chains (Westinghouse, BWXT) can deliver tens of GWs by 2033-2034, requiring government-led first-mover projects to derisk supply chain and revive US nuclear construction capability.

No small modular reactor has been built economically in competitive markets; last US nuclear project (Vogtle) costs 30-40¢/kWh; solar+storage is proven and cost-effective. Host requires demonstrated economic viability before investing, not stories or government subsidies.

Birol expects a significant nuclear renaissance including both traditional large reactors and small modular reactors (SMRs), similar to the post-1970s oil crisis wave that built 40% of today's nuclear capacity, as countries seek energy security and firm low-carbon power.

Europe's lack of fossil energy sources and exposure to Middle East shipping disruptions will drive faster nuclear development as part of green transition, reducing dependency on volatile oil/gas transit routes.

Nuclear thermal propulsion and surface fission power are unique government missions with no commercial use case today; NASA's investment de-risks the technology for future cislunar and Mars economies.

Operating at 0.97 criticality with particle accelerator neutron source makes runaway reaction physically impossible — eliminating the core regulatory hurdle that has blocked nuclear deployment since 1980s, enabling 300MW plants at ~$1B with existing supply chains.

Senra identifies nuclear micro-reactors as 'really difficult wire harnesses to build' with few qualified manufacturers; as Westinghouse and new incumbents scale, Senra's automated manufacturing and engineering partnership model captures this inflection.

Next-gen nuclear startups like Deep Vision and Oklo use mile-deep boreholes where earth provides containment, enabling 70-80% cost reduction, 6-month deployment, and 2-3 acre footprint for 1.5 GW data centers; DOE fast-tracking 11 companies including Deep Vision for July reactor test.

Hartz identifies fusion as the most impactful potential technology for the next decade, calling it 'almost critical for humanity,' and has invested in Fuse to capture the upside of inexpensive, plentiful energy.

The US lost all commercial enrichment capacity by 2013 and now imports HALEU from Russia; domestic centrifuge enrichment (General Matter) with DOE milestone funding and NRC fast-track licensing (12-month target) is the keystone to unlock SMR deployment at scale by 2030.

Nuclear power and propulsion in space is structurally required for Mars missions, outer solar system exploration, and sustained lunar presence; private industry won't fund it due to opportunity cost on terrestrial nuclear, making it a unique NASA mission that also attracts top talent.

Switzerland lifts ban, upgrades 4 reactors (40% power); France's mass-produced single design proves cost discipline; Europe's energy scarcity culture + lost Russian gas + heat dome deaths forces pivot; AI demand makes energy a commercial capacity constraint not environmental issue; SMRs/Gen III 2030s, fusion 2030s, grid time-shifting now.

Activists shut down nuclear in US while China scales aggressively. Data centers require massive baseload power; nuclear is the only viable clean solution at scale. US regulatory capture on nuclear is a strategic vulnerability for AI competitiveness.

Data center energy demand exceeds grid capacity; SMRs provide baseload power with 24/7 availability, driving 13x valuation for Valor Atomics and attracting hyperscaler offtake agreements as the only scalable zero-carbon solution.

The AI-driven electricity gap cannot be closed with renewables alone; nuclear provides the only scalable baseload, but US permitting and NRC processes are the binding constraint, making regulatory reform a prerequisite for continued data center expansion.

Hyperscalers signing power purchase agreements at >2x prevailing rates for guaranteed supply; this cost structure forces leverage and industrial-style balance sheets; nuclear restarts becoming strategic compute enablers.

Radiant Nuclear's 1.5MW container microreactor (5-year no refuel) and Form Energy's iron-air battery (100-day duration, rust-based chemistry) address data center needs for firm, long-duration power. Both leverage physics advantages for stationary grid applications.

Valar Atomics' 27x valuation jump and $450M raise for shipping-container-sized reactors that can co-locate with data centers validates nuclear as the scalable clean energy backbone for AI infrastructure demand.

The failure to build nuclear capacity 50 years ago is described as 'one of the greatest mistakes humanity has made,' creating today's energy scarcity that fuels data center opposition; this historical deficit now structurally supports nuclear renaissance narratives.

Liquid metal heat pipes enable passive cooling at atmospheric pressure, deployable in water-scarce environments. DOE/Army regulatory pathway operational (Project Pele → Advance Act). TRISO fuel supply chain mature (BWXT, DOE AGR program). Military bases need resilient power for cyber/satcom/nuclear command — willing to pay premium and invest equity in supply chain. Civilian data center demand follows military proof points.

Successful continuous operation of Valor Atomic's nuclear reactor demonstrates viable clean baseload power for energy-intensive AI infrastructure, with host expressing high conviction.

Gary Tan showcases nuclear reactor at YC demo day (no uranium); Har Tagar notes team building nuclear reactor inspired by Elon/Palmer/Anduril; investor appetite for hard tech now includes nuclear; structural shift from software-only funding.

Advanced reactors segment by size/application: 1 GW AP1000 for grid baseload; 100-300 MW SMRs for behind-the-meter data centers (5-10 yr horizon); 1 MW microreactors for remote diesel displacement. All need HALEU/LEU fuel, creating pull for enrichment capacity.

Uranium pellet equals ~1 ton coal; first-principles cost should beat fossils. Historical cost escalation came from stopping construction (lost iteration). Advanced reactor factories + domestic enrichment (cost per SWU) can restore learning curve and make nuclear cheapest baseload.

US has zero commercial enrichment capacity; Russia supplies ~25% of enriched uranium. 2028 import ban creates hard supply cliff for both HALEU (advanced reactors) and LEU (existing fleet). Building domestic enrichment (General Matter) captures the highest-value, most constrained step in the fuel cycle.

AP1000 proven but too expensive and labor-intensive; SMR/fusion economics unknown and require public-private partnership; industry risks funding cliff before free cash flow if too many players pursue parallel paths.

One truckload of uranium powers a reactor for a year vs. gas pipeline constraints; Valor generated electricity in 13 months from bare dirt (Falcon 1 moment); proactive community engagement (barbecues, city councils before construction) turns nuclear from NIMBY target into local economic asset; nuclear + data center co-location solves power siting bottleneck for AI.

Rickover's nuclear submarine culture — 100x safety standards, zero radiation deaths, founder-led engineering — created an enduring asymmetric advantage vs China. The model of heretic founders building dual-use industrial capacity (Chrysler, Kodak, GM) is the historical blueprint for modern deterrence.

Hyperscalers are locking in long-term nuclear capacity (X Energy's XC100 reactors) to meet AI compute power demands; private nuclear startups with utility-scale contracts represent a new investable category before public markets.