Quaise Energy

Phase A — Understand the business

Lens 1 · Company Overview

Quaise Energy is a deep-geothermal drilling company founded in 2018, headquartered in Cambridge, Massachusetts, spun out of the MIT Energy Initiative (MITEI) and MIT's Plasma Science and Fusion Center (PSFC). It launched publicly in June 2020 with a $6M seed. Roughly 71–72 employees across four continents as of April 2026.

The one thing it does: replace the mechanical drill bit with directed energy. A surface-mounted gyrotron — a high-power vacuum tube borrowed from nuclear-fusion research — emits a millimetre-wave beam (~30–300 GHz) that is piped down standard oil-and-gas tubing acting as a waveguide, and ablates (melts/vaporizes) rock at the bottom of the hole with no downhole hardware. A purge gas carries the vaporized rock to surface. Conventional rotary drilling gets through sediment to crystalline "basement" rock; the beam then takes over for the hard, hot granite/basalt that destroys steel bits.

Why it matters (the thesis): if it works, millimetre-wave drilling reaches superhot / supercritical geothermal (~400–500°C) at 5–20 km depth anywhere on Earth, turning geothermal from a geographically-lucky niche into a globally-sited, 24/7, zero-carbon baseload source. A superhot well produces up to ~10x a conventional geothermal well. That is precisely the "firm clean power anywhere" profile the AI-data-center buildout is starving for — the same demand thesis that just carried Fervo Energy to a >$10B IPO (Lens 7).

Business model — note it is not yet chosen, and that is a risk. Quaise straddles two very different companies:

1. Drilling-technology / services provider — sell or license the mm-wave rig to oil-and-gas drillers and geothermal developers (asset-light, Nabors-shaped).
2. Independent power producer (IPP) — own and operate the geothermal plants and sell electrons (asset-heavy, capital-intensive, utility-shaped). The Series B narrative ("build the world's first superhot geothermal power plant," Project Obsidian) tilts hard toward model 2 — Quaise becoming a power developer, not just a drill vendor. That is the higher-capital, lower-multiple path, and it changes what the company is.

Customers / offtake (nascent): no commercial revenue. Announced development relationships: Nevada Gold Mines (Barrick + Newmont JV) — evaluating a retrofit of Newmont's TS Power Plant (coal) to deep geothermal heat; and the Project Obsidian plant in central Oregon feeding grid stability across the Pacific Northwest. The AI-data-center framing (a 250 MW Phase to power an AI campus) appears in secondary coverage but is not confirmed in Quaise's own Series B/Obsidian materials, which describe grid power, not a named data-center offtaker — flagged as an unverified secondary claim.

Lens 2 · Supply Chain (name the actual stakeholders)

Upstream inputs → Quaise → end customer, with the real names:

• Gyrotron sources (the critical input, single-ish source): high-power CW gyrotrons are made by a tiny global set — Communications & Power Industries (CPI) (US), Thales (France), Bridge12 Technologies (US, MIT-adjacent), GYCOM/IAP (Russia — now geopolitically off-limits), and Canon/Toshiba (Japan). Fusion-grade megawatt gyrotrons are effectively artisanal, long-lead, and export-controlled. Quaise's jump from 100 kW → 1 MW class hardware depends on this thin, specialized supplier base. This is the sharpest single-source chokepoint in the story.
• Rig / drilling integration: Nabors Industries — one of the world's largest land-drilling contractors — is both a $12M strategic investor and the integration partner; Quaise mounted its 100 kW gyrotron onto a Nabors oil-and-gas rig and ran the Texas field tests on it. Nabors supplies the conventional-drilling front-end, rig crews, and a ready supply-chain/human-capital base to scale.
• Tubing / waveguide: standard OCTG (oil-country tubular goods) repurposed as a dielectric waveguide — commodity steel from the existing oilfield supply chain, but the thermal/mechanical survivability of the waveguide over multi-km at high power is unproven (Lens 13).
• Purge gas / process: industrial gases to flush ablated rock — commodity.
• Downstream / offtake: utilities and grid operators (Pacific Northwest via Obsidian), industrial heat users (Nevada Gold Mines), and — prospectively — hyperscale data-center buyers (the Fervo/Sage template).
• Strategic energy partners: JERA (Japan's largest power generator) and Idemitsu Kosan (Japanese refiner/energy major) entered on the Series B — potential Asian deployment/offtake channels and balance-sheet muscle.

Chokepoints: (1) megawatt gyrotron supply; (2) the waveguide-at-depth engineering; (3) permitting on federal geothermal leases (Obsidian sits in the Deschutes National Forest, i.e. BLM/USFS + geothermal-lease process).

Lens 3 · Competitive Advantages (moats)

• IP / process moat (the real one): Quaise holds the exclusive commercialization path from Paul Woskov's decade-plus of MIT PSFC mm-wave rock-ablation research, plus its own field-integration patents. If mm-wave drilling becomes the way to reach superhot rock, Quaise owns a genuine head start and a patent thicket. This is a deep-tech / know-how moat, not a brand or network moat.
• Depth/temperature reach = the differentiated wedge: every other next-gen geothermal player (Fervo, Sage, Eavor) uses mechanical drilling and is therefore bounded by the same bit-wear/temperature wall that stopped the Kola borehole at 180°C (Lens 8/13). Quaise's only reason to exist is that directed energy sidesteps that wall. If it does, no incumbent can follow without licensing the technique.
• Founder/operator credibility: CEO Carlos Araque spent ~15 years at Schlumberger — Quaise speaks the oilfield's language, which is why Nabors and Mitsubishi backed it.
• Bargaining power — currently weak. Pre-revenue, single-critical-supplier (gyrotrons), and dependent on partners (Nabors) for the rig. Quaise needs its suppliers and partners more than they need Quaise today. That inverts only if/when the technology is field-proven at depth.
• Moat durability caveat: the moat is entirely contingent on the physics working at commercial depth. Until then it is an option on a moat, not a moat.

Lens 4 · Segments / revenue architecture

n/a — pre-revenue, private. No segment or geographic revenue exists to break out (segments.csv empty). The forward architecture is two-legged (drilling-tech vs IPP, Lens 1). The instructive "segment" trend is capital intensity migrating from R&D toward first-plant construction: the Series B explicitly funds Project Obsidian construction + project-level equity/debt, i.e. the spend mix is shifting from lab to balance-sheet-heavy plant development. That shift — from a $200M cumulative-funded science program to a plant that needs hundreds of millions to gigawatt-scale billions in project finance — is the single biggest change in what this company will need to raise.

---

Phase B — Measure performance (+private: funding, traction, cap table, milestone events)

Lens 5 · Funding & valuation trajectory (swaps "Earnings Result")

All amounts ``, unaudited. Cumulative total is the well-sourced anchor; intermediate round sizes vary across outlets and are reconciled below.

• Valuation: n/a — not disclosed. No priced valuation is public for any round (PitchBook maintains a profile but the figure was not retrieved/confirmed here). Do not infer a mark.
• Burn signal: ~$230M raised over 8 years against a still-pre-commercial technology, now pivoting into plant construction. The Series B's explicit need for additional project-level equity + debt "imminently" signals that $134M does not fund Obsidian to completion — more capital is required, and the CEO framed the round as taking Quaise "from field-proven technology to first commercial revenues," i.e. revenue is still ahead, not here.
• Read: the round is a genuine up-round-scale event and the JERA/Idemitsu entries are real validation, but the capital structure is turning from venture-equity into infrastructure project finance, which is a different, harder, lower-multiple game.

Lens 6 · Founder communications / sentiment trend (swaps "Earnings Calls")

No earnings calls exist. Proxy = founder interviews, MIT features, and press-release tone over time:

• 2022 (MIT News): framing was scientific & visionary — "million-year energy source below our feet," deepest holes in the world, 20 km / 500°C ambition.
• 2025 (MIT Tech Review / Canary / IEEE): tone shifted to proof-point / show-me — "from lab to field," first 100 m, careful about what remains (directional drilling, well longevity, plasma control).
• 2026 (Series B): tone shifted again to commercial & grid — "field-proven technology to first commercial revenues," "powering the grid," Project Obsidian, gigawatt-scale.
• Trend: a deliberate, healthy migration vision → proof → commercialization. The thing they stopped leading with is the headline "20 km / 500°C anywhere" moonshot; the thing they started leading with is a ≥5 km / 315–365°C buildable plant. That de-scoping toward reachable depth is a maturity signal — but also a quiet admission that the full moonshot is further out than the early pitch implied.

Lens 7 · Cap table & syndicate quality (swaps "Comps")

Syndicate:

• Climate/deep-tech VCs (repeat leads): Prelude Ventures (lead across A, A1, B), Safar Partners (repeat lead), The Engine / Engine Ventures (MIT tough-tech, original backer), Collaborative Fund.
• Strategic corporates: Nabors Industries (drilling), Mitsubishi Corporation, Standard Investments (Standard Industries' deep-tech arm), JERA, Idemitsu Kosan.
• Board: includes Jeremy Wertheimer (founder of ITA Software, acquired by Google; a serious tough-tech operator/investor).

The tell that matters — crossover funds are ABSENT. The +private IPO-proximity signal is a Fidelity / T. Rowe / Coatue / Wellington entry marking a company toward public markets. Quaise's syndicate is climate-VC + strategic-corporate, with no visible public-markets crossover. Contrast the public comp:

Read: Fervo proved the market (AI-data-center demand → $10B geothermal IPO), but it did so with the de-risked mechanical approach and actual contracted revenue. Quaise is 2–3 category-risk notches behind Fervo on the tech-maturity curve, and its cap table reflects that — strategics who want optionality on a breakthrough, not crossovers pricing an exit. Peer multiples are not applicable to a pre-revenue private; EV/Sales, P/E = n/a / not meaningful.

Lens 8 · Milestone / catalyst events (what moves the private mark)

Events that would move a secondary mark (analog to the ">5% stock-move" lens):

• 2020 launch ($6M) — existence.
• 2022 Series A + Nabors — oilfield validation.
• 2024 first gyrotron-on-rig integration; 4-inch hole, ~30–40 ft deep on the Nabors rig.
• 2025-07 the de-risking event: 100 m drilled with mm-wave in the field (central Texas quarry; basalt columns per MIT Tech Review), first field ablation without downhole hardware — up from a few cm in the MIT lab.
• 2024-12 Nevada Gold Mines partnership — first industrial-offtake signal.
• 2026-07-07 $134M Series B + Project Obsidian unveiled — the commercialization pivot and the biggest positive mark event to date.
• Company claim (unverified independently): "drilled over 100 m through granite and approaching one kilometre" at the Texas site as of the Series B — note the rock-type inconsistency (2025 independent coverage said basalt; 2026 PR says granite) and treat "approaching 1 km" as a company figure pending third-party confirmation.
• Pattern: the mark reacts to field-depth records and strategic-capital entries, not to lab results or vision decks. The next mark-moving events are the 1 MW gyrotron field test (targeted 2026) and multi-km field depth.

Traction & unit economics (+private add-on)

• Revenue run-rate / ARR: ~$0 commercial. n/a — pre-revenue.
• Physical traction: 100 m verified field depth (2025); "approaching 1 km" claimed (2026). Target near-term commercial depth ≥5 km (Obsidian); ultimate 10–20 km. Drilling-rate goal 3–5 m/hr.
• Economics (target, company + literature): conventional geothermal LCOE $64–106/MWh, beating US nuclear ($142–222/MWh) at ~90% capacity factor; superhot EGS modeled feasible at ~5¢/kWh with ~$657M capex for a 162 MW superhot development. Quaise's own thesis: one superhot well ≈ 10x a conventional well, driving LCOE toward fossil parity at terawatt scale. These are targets, not results.

---

Phase C — Judge people & books

Lens 9 · Management

• Carlos Araque — CEO & co-founder. MIT engineering degree; ~15 years at Schlumberger in drilling technology. The credibility spine — an oilfield-drilling insider running a drilling-tech startup. Archetype: technical founder-operator, well-matched to this stage.
• Matt Houde — co-founder & Chief of Staff. Geologist; previously at AltaRock Energy (EGS pioneer). Geothermal-resource domain depth.
• Paul Woskov — scientific co-founder. MIT PSFC research engineer; originator of the mm-wave rock-ablation technique (answered a 2008 MITEI RFP). The IP wellspring.
• Kevin Bonebrake — CFO. Present ahead of the capital-intensive plant phase — appropriate for a company about to raise project debt.
• Dr. Geoffrey Garrison — VP Operations; Dr. Trenton Cladouhos — VP Geothermal Resource Development (superhot EGS); Diane Hughes — VP Marketing/Comms. A credible technical bench; Cladouhos in particular is a recognized geothermal-resource name.
• Board: Jeremy Wertheimer (ex-Google/ITA Software) named — a tough-tech-savvy director. Board appears thin (one named director in profiles) for a company entering project-finance scale — worth watching for institutional governance build-out.
• Skin in the game: founders retain meaningful equity (typical for an 8-year-old venture-backed private); exact ownership n/a — not disclosed.
• Capital-allocation history: disciplined-ish — ~$230M over 8 years, staged against milestones, heavy reliance on strategic co-investors to de-risk. The forward test is the plant-construction capital-allocation decision (own-and-operate vs license), which will make or break returns.
• Red flags (management): none egregious. Mild watch-items: (1) the two-model strategic ambiguity (Lens 1); (2) a thin public board for the capital scale ahead; (3) PR-vs-independent inconsistency on field results (granite/basalt, "approaching 1 km").

Lens 10 · Forensic red flags (private lens: governance, burn, dilution, disclosure quality)

No audited financials exist to forensically analyze — that itself is the headline caveat: all figures are self-reported and unaudited. Private-company risk items:

• Disclosure quality: results are communicated via press release, not audited filings; note the granite vs basalt and "approaching 1 km" claims that lack independent confirmation. Not fraud — but the standard "trust-but-verify" posture for pre-revenue deep-tech PR applies.
• Dilution / burn: 5 priced/announced rounds over 8 years; the pivot to project finance implies substantial further dilution and/or leverage ahead. Founders' and early backers' stakes will compress as infrastructure capital enters.
• Concentration: dependence on Nabors (rig) and a thin gyrotron supplier base are operational single-points-of-failure.
• Going-concern framing: typical venture risk — the company lives round-to-round; the Series B explicitly needs follow-on project capital to build Obsidian.

Regulatory findings (required sub-section).

• SEC (EDGAR EFTS — LR + AAER): regulatory/regulatory-findings.md (generated 2026-07-10) reports 0 SEC findings — Quaise has no CIK, is private, and is not required to file; no EDGAR enforcement search is possible.
• Non-SEC (web search — FTC/DOJ/FDA/EPA/consent-decree/settlement/fine): no material enforcement actions, litigation, or penalties surfaced against Quaise Energy in web search as of 2026-07-10.
• Item 3 Legal Proceedings: n/a — no 10-K exists (private, no filings on shelf).
• Sector-regulatory watch (not enforcement): the real regulatory exposure is prospective, not historical — federal geothermal leasing/permitting (BLM/USFS) for Obsidian in the Deschutes National Forest, and induced-seismicity/environmental review for any deep-drilling program. No induced-seismicity incidents attributed to Quaise to date (its no-fracking approach is a seismicity advantage vs EGS peers).
• Conclusion: No material regulatory or legal findings — verified via SEC EDGAR EFTS (LR, AAER), web search, and the absence of any required filing, as of 2026-07-10.

---

Phase D — Project & stress-test

Lens 11 · IPO-readiness & path-to-tradeable (swaps "Forward Projection")

No private-watch.json entry exists for Quaise as of this run (the overlay ledger has no stage/ipo_readiness/catalyst for the slug) — IPO-readiness is assessed web-only here. (Per --watchlist wave rules this dossier does not write back to private-watch.json; flag to the master session that an entry could be seeded: stage = "field-proof / pre-commercial", readiness = LOW, catalyst = "1 MW gyrotron field test 2026 + multi-km depth".)

Readiness: LOW / early. Path-to-tradeable: ~2030+, contingent on Project Obsidian.

• What Fervo's IPO proved: the public market will pay ~$10B for next-gen geothermal — but only with contracted revenue ($7.2B backlog) and a plant under construction (500 MW Cape Station). Fervo had product-market-and-capital-market fit; Quaise has neither yet.
• Milestones that would unlock an S-1, in order: (1) 1 MW-class gyrotron drilling in the field (targeted 2026); (2) maintained open borehole at multi-km depth and >300°C — the never-before-done step; (3) Project Obsidian first electrons to grid (company target 2030, phase I ~50 MW; note this slipped from the ~2028 / ~20 MW pilot described in mid-2025 coverage — a right-shift of ~2 years in ~12 months); (4) signed offtake (utility or hyperscaler PPA) converting the story into backlog.
• Estimated window: a public listing realistically requires the full drill-to-electrons loop demonstrated — earliest ~2030, more plausibly early-2030s. This is a venture/deep-tech option, not a near-term tradeable.
• Tracked binary (Brier-loggable, not logged per --watchlist rules): "Quaise drills and maintains a producing borehole to ≥3 km at ≥300°C in the field before end-2028" — p ≈ 0.30 ``. That single binary is the hinge of the entire thesis.
• No EPS forecast — pre-revenue; forecast.ts create intentionally skipped (watchlist mode).

Lens 12 · Bull vs Bear

Bull case. Quaise owns the only geothermal approach that structurally escapes the temperature/bit-wear wall that has capped drilling since Kola. If mm-wave ablation works at depth, it unlocks superhot baseload power anywhere on Earth — the single most valuable energy profile in an AI-electrified grid, and a market Fervo just proved is worth ~$10B for the inferior mechanical version. The IP is a decade of MIT science with an oilfield-fluent CEO; the syndicate now includes two Japanese energy majors (JERA, Idemitsu) plus Nabors' rig muscle; and Project Obsidian gives a concrete, financeable first asset. The physics is not disproven — Quaise's own line is "we're not working against the laws of physics, only cost and engineering." The surprise: if the 1 MW gyrotron hits multi-km fast, the re-rate from "science project" to "category-definer" is violent, because there is no mechanical follower who can match superhot depth.

Bear case (2–3 permanent-impairment risks). (1) Borehole survivability at depth — vaporizing heterogeneous rock yields asymmetric, non-uniform boreholes that may collapse under high pressure/temperature; the proposed fixes (vitrified glass lining, intermittent redrilling) are unproven beyond the lab. No one has ever maintained an open hole at 5–20 km / 400°C. (2) Waveguide power delivery — piping megawatt-class mm-waves down km of tubing without catastrophic thermal/mechanical loss or plasma formation (which damages equipment) is unsolved at scale. (3) Time-and-cost patience — Cornell's Jefferson Tester ("very aspirational… it's all about cost… investors may lose patience"); the pilot already slipped 2028→2030. Deep-tech that needs 10+ years and infra-scale capital can die of funding fatigue before physics vindication.

Pre-mortem (18 months out, thesis broke): the 1 MW gyrotron field campaign stalls — boreholes collapse or the waveguide can't carry the power past ~1–2 km — Obsidian's depth target quietly retreats again, the "imminent" project debt doesn't close on acceptable terms, and Quaise raises a flat/down bridge while Fervo/Sage soak up every AI-data-center PPA with shovels-in-the-ground. Quaise becomes a licensing/IP story, not the IPP it now pitches.

Are the (implied) multiples too high? No priced mark is public, so there's nothing to call rich — but the narrative premium (gigawatt-scale, "first commercial revenues") is running well ahead of the demonstrated ~100 m–1 km reality. The gap between story and proof is the risk.

Contrarian view (what the market refuses to see): the market is treating Fervo's IPO as validation for the whole category — but Fervo's win actually raises the bar for Quaise, because it proves you can win AI-power contracts with boring, de-risked mechanical EGS today. Quaise's exotic physics has to beat not just fossil/nuclear but a now-public, well-capitalized mechanical incumbent that is already signing the offtakes. The bull "no follower can match superhot depth" only matters if superhot depth is required — and Fervo is proving a lot of the demand can be met shallower and sooner.

Lens 13 · Devil's Advocate (short-seller)

Dismantling the bull case:

• The core product does not yet exist. After 8 years and ~$230M, the verifiable field result is ~100 m (and even the rock type — basalt vs granite — is inconsistent across the company's own timeline). Commercial requires 5,000–20,000 m. That is not an incremental gap; it is 50–200x, across a regime (open hole at 400°C+) no human project has ever sustained — Kola stopped at 180°C/12.2 km, KTB at 260°C/9.1 km, and both were plagued by hole instability and malleable rock. Quaise's differentiator (energy vs bit) removes the bit-wear problem but does nothing about the rock going plastic and the hole closing at depth.
• Physics they under-sell: millimetre-wave attenuation and mode purity over kilometres of imperfect waveguide; plasma breakdown at the beam front; keeping the hole clear of melt re-solidification. Each is a potential hard stop, not a cost line.
• Strategy incoherence: is this a drill vendor (asset-light, high-multiple) or an IPP (asset-heavy, utility-multiple)? The Series B pivots to owning a power plant — the worse business — precisely because selling a not-yet-proven rig is hard. That's a tell.
• Capital-market reality: the absence of any crossover fund after a $134M round says the smart late-stage money isn't pricing an exit. The follow-on need ("project equity + debt imminently") means the $134M headline overstates runway.
• Most dangerous competitor bulls underestimate: Fervo — now public, ~$10B, $7.2B backlog, taking the AI-data-center PPAs now with technology that works today. Also Sage (Meta 150 MW) and Eavor (no-frack closed loop). Quaise could win the physics and still lose the market to whoever electrifies data centers first.
• What breaks it permanently: a definitive field finding that boreholes cannot be kept open/producing at target depth/temperature with mm-wave ablation. Plausibility: material — this is the single unproven step and it's the whole company. If growth (i.e., field depth) disappoints even 20–30% versus the Obsidian plan, the plant is un-financeable and the equity story reverts to a patent-licensing shell.

Lens 14 · Fifteen Questions for the CEO (ordered by information value)

1. At what field depth and rock temperature have you maintained an open, stable, producing borehole — not just ablated rock — and what is the longest such hole to date? (The whole thesis.)
2. What is the demonstrated power-delivery efficiency of the waveguide at 1 MW over 1 km, and how does attenuation/plasma-formation scale to 5 km?
3. Independent third-party verification of the "approaching 1 km" and rock-type (granite vs basalt) claims — who has audited the Texas results?
4. Project Obsidian slipped from ~2028/20 MW to 2030/50 MW in one year — what specifically caused the slip, and what confidence interval is on 2030?
5. How much total capital (equity + project debt) does Obsidian require to first electrons, and how much beyond the $134M is committed vs. hoped-for?
6. Drill vendor or IPP — which company are you building, and why not license the rig to Nabors/others and stay asset-light?
7. What is your contingency if borehole collapse at depth proves unsolvable with vitrified lining / redrilling?
8. Gyrotron supply: who builds your 1 MW+ units, what is the lead time, and how export-control-exposed is that supply chain?
9. What offtake (PPA) is signed or in exclusivity for Obsidian, and at what price/tenor?
10. How do your delivered LCOE and capex/MW compare to Fervo's demonstrated $7,000/kW → $3,000/kW learning curve, on a like-for-like basis?
11. What is the latest priced valuation, and why has no public-markets crossover investor participated?
12. What is current monthly burn and runway on the $134M before project financing?
13. What directional-drilling capability exists — Obsidian's fractured-reservoir design implies deviated wells you've said you can't yet do?
14. What induced-seismicity / environmental review does the Deschutes federal-lease site require, and what's the permitting timeline risk?
15. What single field result in the next 18 months would you treat as thesis-invalidating, and are you pre-committed to acting on it?

---