Teardown

Energy · Deep dive

Aalo Atomics

The microreactor company betting hyperscaler AI needs factory-built nuclear.

emerging

The question that decides it: Does Aalo convert the Crusoe partnership and Idaho Falls MOU into a second commercial PPA at an LCOE that proves out the hyperscaler thesis before Series B cash depletes?

My take

Founded
2022
Revenue
Pre-revenue
Headcount
~120+ (targeted post-Series B)
Screen
Published
2026-11-17
Web
www.aalo.com
Elsewhere
LinkedIn · Crunchbase

Founders and leadership

  • Matt Loszak

    Co-founded Humi (Toronto HR SaaS) in 2016; Queen's University engineering physics.

  • Yasir Arafat

    Prior at Westinghouse (eVinci microreactor) and Idaho National Lab (MARVEL project lead).

Snapshot

Aalo Atomics is the Austin-based advanced-reactor company trying to compress the 15-year gap between an engineering drawing and an operating nuclear plant to something closer to 4 years — and to sell behind-the-meter power to hyperscale AI data centres while it does it. Founded in late 2022 by Matt Loszak and later joined by MARVEL microreactor lead Yasir Arafat, Aalo has raised more than $136M across a $6.26M seed, a $27M Series A led by Fifty Years and Valor Equity Partners in July 2024, and a $100M Series B led by Valor in August 2025. The company's Critical Test Reactor achieved first criticality on July 4, 2026, and a Crusoe partnership announced July 30, 2026 targets a first commercial Aalo-X at an INL data centre in 2027 and 50-MWe Aalo Pods at Crusoe sites by end of 2029. The AI-power narrative is the single reason Aalo looks fundable today, and it is also the single risk that could reshape the thesis if hyperscaler demand curves cool.

Founding story

Matt Loszak co-founded the Toronto HR-and-payroll SaaS company Humi in 2016 with Kevin Kliman, grew it to roughly 150 employees over six years, and then moved to Austin to start Aalo in late 2022. Loszak is a Queen's University engineering physics graduate with a short-form video of himself doing arithmetic on a whiteboard that became a Twitter moment during the 2024 round close. Yasir Arafat — the only person in the industry who can credibly claim to have shipped two microreactor programmes — spent about a decade at Westinghouse starting the eVinci programme in roughly 2014, then went to Idaho National Laboratory as the project lead on MARVEL (Microreactor Applications Research Validation & Evaluation), which he carried to 90% design-complete status under DOE approval. Arafat formally joined Aalo as CTO in October 2023. The founders' pitch to Valor and the other investors was that the MARVEL design was ready, non-TRISO fuel sidesteps the supply bottleneck most SMRs face, and factory-built modular manufacture is the only path to the kind of deployment cadence hyperscalers need.

How it works

The Aalo-1 is a sodium-cooled, pool-type liquid-metal reactor sized at about 30 MW thermal and 10 MW electric, based directly on INL's MARVEL. The original design used uranium-zirconium hydride (UZrH) fuel, with hydrogen in the fuel providing the neutron moderation. Aalo moved to a UO₂ fuel enriched to ~5% LEU after signing a supply contract with Urenco in September 2025 (Data Center Frontier, Sept 2025). This is below the 5–20% "HALEU" band and matters because Aalo sidesteps both the HALEU supply crunch and the proliferation-concern debate. The Aalo Pod — marketed as the "XMR" or extra-modular reactor — is a 50-MWe plant built from five Aalo-1 units sharing one turbine, with a 100-MW footprint under five acres. Factory build happens at Aalo's 40,000-square-foot Austin facility; the plant ships out and is assembled on site. External water is not required for cooling, which is the practical reason the company can site at data-centre campuses where water rights are a constraint.

Product and business overview

Three product lines so far. Aalo-X is the full-scale 10-MWe demonstration unit at INL and the vehicle Aalo is using to prove the manufacturing playbook. Aalo Pod / XMR is the commercial 50-MWe plant sold to data-centre operators. Aalo Pod extensions push the architecture to 100 MWe and beyond in modular blocks. The business is build-own-operate, not reactor-sale — Aalo retains asset ownership and sells electricity on long-dated PPAs, which is how the hyperscaler customers want to transact. The Critical Test Reactor achieved first criticality on July 4, 2026, which is a meaningful operational milestone (not full operating life, but a proof that the lattice and controls work as designed).

Business model and pricing

Pricing is not public. Loszak has said in interviews and investor commentary that the non-TRISO fuel route saves 10–20 cents per kWh relative to TRISO competitors (Power Magazine, Aug 19, 2025), and the Pod's LCOE is pitched below most competing SMRs. The target customer is a hyperscaler needing 50–500 MW of behind-the-meter or on-site power for AI training and inference clusters, with secondary customers in remote industrial, mining, and process heat. The Idaho Falls Power MOU is for a 75-MW fleet (seven Aalo-1 units) at the INL Energy Research Park; the Crusoe partnership covers an Aalo-X at an INL site plus commercial Aalo Pods at Crusoe sites by 2029.

Traction over time

Capital, people, and regulatory milestones have moved on roughly a 12–18 month cadence. 2023 seed $6.26M; October 2023 Arafat joins as CTO; May 2024 DOE site-use MOU; July 2024 $27M Series A co-led by Fifty Years and Valor Equity Partners, with Alumni Ventures, Harpoon Ventures, Earthshot Ventures, Preston Werner and Jeff Dean participating; December 10, 2024 DOE identifies tentative INL site for Aalo-X; February 2025 selected for the Texas A&M RELLIS "proving ground" (up to 1 GW); July 2025 Urenco fuel contract; August 20, 2025 $100M Series B led by Valor and bringing cumulative financing to more than $136M, plus selection into the DOE Reactor Pilot Program (which allows some non-NRC pathways for the pilot build); July 4, 2026 Critical Test Reactor criticality; July 30, 2026 Crusoe partnership. Headcount was approximately 60 at Series B close, with the Series B earmarked to double to 120+.

Market analysis

The driver is data-centre electricity. Grid Strategies' 2025 utility filings imply roughly 90 GW of data-centre-driven summer-peak growth in the US between 2025 and 2030, about 55% of total load growth of 166 GW (Grid Strategies, 2025). Lawrence Berkeley/DOE put US data-centre electricity use at between 6.7% and 12% of national consumption by 2028. Average new data-centre capacity rose from 150 MW in 2023 to 300 MW in 2024, and about half of 2030-era projects are planned at over 1,000 MW. Behind-the-meter generation is already a feature — roughly 9% of new builds include on-site power, representing a third of planned capacity. For advanced reactors, there is no clean public TAM, but Oklo alone claims a ~14 GW customer pipeline across LOIs, implying a $50B-plus addressable opportunity at $4–5M per MWe of CapEx. The 10x30 executive order and DOE Pathways Report together target 200+ GW of US advanced-reactor fleet by 2050.

Competitive intel

Oklo (NYSE: OKLO, listed May 10, 2024 via AltC SPAC) is the most direct analog — a sodium-cooled fast reactor family (15–75 MWe), co-founded by Jacob and Caroline DeWitte, with Sam Altman as SPAC chair; Oklo broke ground at INL on September 29, 2025 and is targeting late-2027 commercial. Oklo has more public-market capital and a bigger pipeline but a similar design-stage risk profile. NuScale (NYSE: SMR) is the cautionary tale: its 77-MWe LWR Carbon Free Power Project was cancelled in November 2023 at an $89-per-MWh price. X-energy has an 80-MWe HTGR using TRISO fuel, got a $500M-plus Amazon investment in October 2024, and is slower to deploy. Kairos Power has a FLiBe molten-salt pebble-bed design with a 500-MW Google offtake signed October 14, 2024 and a 2030 first-unit target. TerraPower Natrium (Bill Gates) broke ground at Kemmerer, WY in June 2024 for a 345-MWe sodium-plus-molten-salt-storage unit. Last Energy ships a 20-MWe PWR micro, UK-focused, with reported $19B of European MOUs. Westinghouse eVinci (Arafat's prior programme), BWXT (DOE Project Pele transportable), and Rolls-Royce SMR are slower to US data-centre markets. Aalo's positioning is explicitly the "sweet spot" between Oklo-scale micros and NuScale-scale SMRs: a 50-MWe pod, non-TRISO fuel, no external water, factory-built, and sited behind the meter.

History and evolution

Late 2022 founded; mid-2023 $6.26M seed; October 2023 Arafat joins as CTO; May 2024 DOE site-use MOU; July 2024 $27M Series A; December 2024 DOE INL tentative site decision; February 2025 Texas A&M selection; July 2025 Urenco contract; August 20, 2025 $100M Series B and DOE Pilot Program selection; July 4, 2026 Critical Test Reactor first criticality; July 30, 2026 Crusoe partnership, targeting an Aalo-X-powered Crusoe Spark modular data centre at INL in 2027 and commercial Aalo Pods at Crusoe sites by end of 2029. The regulatory story is deliberately heterodox: the DOE Pilot Program lets Aalo build the pilot under DOE authority rather than NRC licensing, with NRC still required for commercial sales afterward.

What people say

The case for. Bloomberg, Power Magazine, the American Nuclear Society, Data Center Frontier, NEI Magazine and Reuters pickups have all run broadly favourable coverage emphasising the manufacturing playbook, the explicit AI-data-centre thesis, and the non-TRISO fuel decision. The Crusoe partnership is read as the first concrete commercial validation outside DOE funding. Valor's willingness to lead back-to-back rounds, and participation from Hitachi Ventures, NRG Energy, Tishman Speyer and Fine Structure Ventures, suggests credibility across strategic and venture investors.

The complaints. Three serious skepticisms recur. First, 2027 and 2029 timelines are "very optimistic" given NRC licensing history — the industry's track record is consistent slippage. Second, arms-control analysts flagged the broader HALEU-microreactor cohort as a potential proliferation stepping-stone (Kemp et al. in Science, June 2024; Bulletin of the Atomic Scientists, November 2025); Aalo's switch to 5% LEU UO₂ largely sidesteps this, but it is a residual concern for the US advanced-reactor fleet overall. Third, the NuScale/UAMPS cancellation in November 2023 is the industry's standing cost-overrun warning: a project priced below $90/MWh fell apart at bid; a project trying to beat that price without visible cost discipline carries the same risk.

Outlook: the open question

The open question is whether Aalo can convert a convincing manufacturing narrative into a second signed commercial PPA before cash from the $100M Series B is consumed — and whether that PPA clears at an LCOE low enough to prove the hyperscaler thesis against renewables plus storage and against competing microreactor plays. The bullish answer: the Crusoe partnership, the Idaho Falls Power MOU, the Urenco fuel contract and the DOE Pilot Program selection together produce a defensible path to first revenue in 2027 and portfolio contracting in 2029. The bearish answer: three competing microreactor programmes (Oklo, Kairos, X-energy) have more capital and comparable technology; hyperscaler demand for behind-the-meter nuclear is real but still priced aggressively against renewables-plus-batteries; a cost overrun on the first commercial unit is the industry-standard outcome and would require a bridge round at a difficult moment.

How to attack it

The specific wedge is a brownfield microreactor repowering play — approaching retiring coal and gas plants with a modular-swap design that reuses the existing interconnection, cooling water (where available), and turbine hall, dramatically shortening the siting and permitting cycle. Aalo's exposed weaknesses point to attack vectors a new entrant could exploit. First, Aalo is single-design: everything hangs on the Aalo-1 lattice, so any design iteration ripples through the Pod architecture. Second, factory manufacture is a bet on repeat volumes that are not yet contracted — the Austin 40,000-square-foot facility is a fixed-cost liability if the order book slips. Third, licensing is split awkwardly between DOE (for the pilot) and NRC (for commercial sales), which creates regulatory arbitrage complexity for customers who want long-dated PPAs. Fourth, the hyperscaler customer concentration (Crusoe, future Oracles and AWSes) means a renewable-plus-storage break-through could suddenly deprioritise nuclear in that segment. Fifth, the fuel supply (Urenco UO₂ at 5% LEU) is less of a bottleneck than TRISO, but it is a single-vendor dependency that an attacker could flip into a multi-source story.

Adjacent-segment play

The most attractive adjacent segment is district heat and industrial process heat — the Aalo-1's 550°C outlet temperature is well-matched to process steam at chemical plants, refineries and large greenhouses. Last Energy has pitched exactly this in Europe with its MOU book, and there is no US incumbent — a repackage of the Aalo-1 for industrial heat (rather than electricity) could sell to Dow, Nutrien, and the North American chemical complex without touching data-centre demand. Other credible adjacencies: Caribbean and remote-grid replacement for diesel baseload at island utilities (which pay $200–300/MWh today); defence and critical-infrastructure microgrids (Westinghouse eVinci and the DOE Project Pele already occupy this slot, but demand exceeds supply); and desalination co-siting in water-stressed markets where no-external-water cooling is doubly attractive. The adjacency to avoid is residential or small commercial retail — the fixed cost of nuclear siting and the regulatory overhead do not scale down.

Sources and further reading

Capital history

DateRoundAmountValuationLead(s)
2023 $6.26M — —
2024-07 $27M — —
2025-08 $100M — —

Investors / owners: Valor Equity Partners, Fifty Years, Hitachi Ventures, NRG Energy, Tishman Speyer, Fine Structure Ventures, Harpoon Ventures, Alumni Ventures, Crosscut, Kindred Ventures, Nucleation Capital

Competitive set

  • Oklo —
  • X-energy —
  • Kairos Power —
  • NuScale —
  • TerraPower Natrium —
  • Last Energy —
  • Westinghouse eVinci —