Space
The only credible bet on FULL (both-stages) reuse besides SpaceX — a metallurgy/physics moat the others ducked — but it is a single-vehicle, zero-revenue, zero-flights company whose entire value is gated on one un-flown second stage surviving reentry; WATCHING, not investable, until Nova reaches orbit and the upper stage comes home intact.
Research
The verdict
The only credible bet on FULL (both-stages) reuse besides SpaceX — a metallurgy/physics moat the others ducked — but it is a single-vehicle, zero-revenue, zero-flights company whose entire value is gated on one un-flown second stage surviving reentry; WATCHING, not investable, until Nova reaches orbit and the upper stage comes home intact.
Stoke Space Technologies is a Kent, Washington launch startup founded in 2019–2020 by two ex-Blue Origin propulsion engineers building Nova, a fully and rapidly reusable two-stage medium-lift rocket. The entire company is one product and one thesis: that the unsolved half of rocket reusability — bringing the second stage home intact and flying it again with little-to-no refurbishment — can be cracked with a novel actively-cooled metallic heat shield, and that doing so collapses cost-to-orbit by ~20x.
Business model. Launch-as-a-service: sell payload-to-orbit on Nova to commercial satellite operators, constellation builders, and the US government, plus the differentiated "to, through, and from space" services full reuse enables — in-space mobility, satellite servicing/repositioning, and downmass (returning cargo from orbit), which an expendable upper stage physically cannot offer.
Products. A single vehicle, Nova:
Customers / contracts. Government anchor + a "substantial" commercial manifest (Stoke's own characterization; unverified line items):
n/a — private, not disclosed.Contract structure / payment terms. Government IDIQ-style competitive pools (bid per task order, no guaranteed volume); commercial launch service agreements are typically milestone deposits + balance-on-launch. Concentration, deferred-revenue balance, and take-or-pay terms are all undisclosed → n/a — private.
Suppliers. Vertically integrated on the hard parts (engines, heat shield, structures manufactured in-house in Kent); buys raw metals, avionics components, and ground-systems hardware on the merchant market. No disclosed single-source supplier dependency.
Map: upstream inputs → Stoke → end customer.
Chokepoints / single-source dependencies. (1) SLC-14 is the only launch site — single pad, single coast, no second range; any pad anomaly grounds the entire company. (2) The additive-manufacturing line for the 24-chamber Andromeda ring is a concentrated internal chokepoint — yield/throughput on 3D-printed thrust chambers gates cadence. (3) Exotic-alloy supply for the metallic heat shield is the input most exposed to a sole-source squeeze. Names or it didn't happen: the public record does not disclose the specific alloy vendor or the printer OEM → flagged as an open item, not asserted.
The moat is a physics-and-metallurgy bet nobody else made. Every other Western launcher pursuing reuse — SpaceX (Falcon 9), Blue Origin (New Glenn), Rocket Lab (Neutron), Relativity (Terran R) — reuses only the first stage and throws the second stage away (Starship aside, which reuses both but is super-heavy, a different market). Stoke is the only credible small/medium-lift program targeting full, both-stages reuse from day one.
The differentiator is the actively-cooled metallic heat shield: liquid hydrogen flows through channels in a metal shield during reentry, absorbing heat and feeding it back into the turbomachinery — no fragile ceramic tiles, no ablative layer, "ready to fly again immediately without inspection or refurbishment". The claimed insight: regenerative cooling already survives a rocket combustion chamber (~20× more thermally intense than reentry), so applying it to a heat shield is a de-risked extrapolation, not a moonshot.
Durable moats (if it works):
Bargaining power. Today: weak. Pre-revenue, pre-flight, zero pricing power; customers hold all the leverage and can wait for SpaceX. Post-Nova-success: power flips toward Stoke specifically on the downmass + in-space mobility services that are uncontested — there, who-needs-whom favors Stoke because no one else can offer it at price.
n/a — private, pre-revenue, not disclosed. There are no revenue, EBITDA, or earnings segments — segments.csv is an empty scaffold and the company reports nothing. The asset table is the company. The only meaningful "segmentation" is the addressable use-case wedge Nova is built to attack:
| Use-case wedge | Why Nova is positioned for it | Contestable? |
|---|---|---|
| Medium-lift to LEO (constellations) | Fills the Electron→Falcon-9 gap (~300 kg → 22,800 kg) | Highly — Neutron, Terran R, Falcon 9 rideshare all here |
| National-security small-sat (NSSL Lane 1, OSP-4) | On-ramped, US-sovereign, responsive | Moderately — vs Rocket Lab, Firefly |
| In-space mobility / satellite servicing | Restartable upper stage that returns | Lightly contested |
| Downmass (cargo from orbit) | Only full-reuse upper stage can do it | Uncontested |
Trend/cause: not measurable without revenue. The strategic read is that Stoke is deliberately not competing head-on with Falcon 9 on price-per-kg of bulk LEO (a loser's game vs. Starlink-subsidized cadence) but staking the differentiated edges full reuse uniquely enables.
There are no earnings. The scoreboard is the cap raise and the valuation step-ups:
| Round | Date | Amount | Cumulative raised | Post-money valuation | Lead / notable |
|---|---|---|---|---|---|
| Seed → Series B | 2020–2023 | ~$220M cumulative | ~$220M | n/a | Bond, Point72 Ventures, Y Combinator, Breakthrough Energy |
| Series C | 2025-01-15 | $260M | ~$480M | ~$944M (PitchBook) | Breakthrough Energy, Glade Brook, Point72, Seven Seven Six, Univ. of Michigan, Woven (Toyota), YC |
| Series D (initial) | 2025-10-08 | $510M (+$100M SVB debt) | ~$990M | ~$1.95B | USIT (Thomas Tull's US Innovative Technology Fund) |
| Series D (extension) | 2026-02-10 | +$350M (round → $860M total) | ~$1.34B to date | ~$1.95B+ (not re-marked publicly) | extension to existing round |
Conflict surfaced (per provenance discipline): an early scrape claimed "$1.34B raised under Series D" as if Series D alone were $1.34B. That is wrong: $1.34B is total cumulative capital raised across all rounds; the Series D round itself is $860M (initial $510M + $350M extension). The dossier uses the cumulative figure.
Burn signal. A two-stage, full-reuse rocket with two factories and a refurbished Cape pad is enormously capital-intensive. The $510M Series D was explicitly sized to "complete development and demonstrate Nova through its first flights" — i.e. it funds to first flight, not through commercial operations. The Feb-2026 $350M extension "to accelerate other elements of the product roadmap" signals the program needed more than the October raise contemplated — a soft tell that costs/timeline ran ahead. ``: implied burn is plausibly $300–450M/yr given the cumulative raise pace ($480M→$1.34B in ~13 months) — unverified, directional only.
No earnings calls exist. Proxy: founder communications and milestone messaging. The arc is consistently confident-but-engineering-grounded — Andy Lapsa's public register is technical and cadence-focused, not promotional-hype (a favorable tell vs. SPAC-era space promoters). Recurring themes:
Sentiment trend: steadily more operational, appropriately hedged on timeline. No detectable over-promise spiral. This is a credibility positive.
Syndicate quality — the IPO-proximity tell. The cap table has migrated from classic VC (Bond, Point72, YC, Seven Seven Six) toward strategic / national-security capital: the Series D lead is USIT (Thomas Tull), a fund explicitly oriented to "critical technologies relevant to the national interest"; plus Toyota's Woven Capital, Breakthrough Energy (Gates), and the University of Michigan endowment. What's notably absent: a marquee crossover round — no disclosed Fidelity / T. Rowe / Coatue entry. Under the +private framework, the absence of a crossover fund is itself the signal: this is not yet an IPO-proximate cap table. USIT is strategic, not a public-market-on-ramp investor.
Secondary marks (the live tell). Stoke trades on secondary venues:
Read: the secondary print at/near $40 sitting below implied step-ups from the $1.95B primary mark is a mild softening / illiquidity-discount signal — consistent with a hard-tech name where the value is gated on an un-flown vehicle and secondary buyers demand a haircut for binary flight risk. No mutual-fund markup/markdown disclosures are public → n/a. There is no peer EV/Sales or P/E table — peers (Rocket Lab, the only public pure-play) trade on revenue Stoke does not have; a multiples comp would be fabrication. n/a for all earnings multiples.
No public stock, so "what moves the mark" = the milestone/funding event ladder. Pattern of value-moving events 2023–2026:
What the pattern reveals: the mark steps on (a) capital events and (b) hardware-test milestones, not on revenue. The dominant forward catalyst — Nova's maiden orbital flight, targeted late 2026 — is the single binary event that re-rates (or impairs) the entire enterprise.
Andy Lapsa — Co-founder & CEO. Began at Blue Origin developing the BE-4 engine, then led the BE-3 and BE-3U engine programs — i.e. he ran the propulsion for New Shepard and New Glenn's upper stage. PhD in Aerospace Engineering (Univ. of Michigan), BS Mechanical Engineering (Cornell). Track record: genuine, delivered, propulsion-deep — not a finance founder. Built real flight engines at one of two serious US new-space primes before leaving over a specific conviction (no one was pursuing full reuse).
Tom Feldman — Co-founder & CTO. Senior propulsion design engineer on the BE-4 (oxidizer pump + thrust chamber) at Blue Origin; M.S./B.S. aerospace (Purdue); prior internships at SpaceX and NASA. Met Lapsa at Blue Origin.
Archetype: founder-engineers, not professional managers — the right archetype for this stage, where the binding constraint is hard propulsion/thermal engineering, not sales. "Heads down doing the math in Lapsa's basement for six months" before raising is the right origin signature for a deep-tech moonshot.
insider-transactions.csv absent) → n/a — private.n/a — pre-revenue.Net: high-quality, mission-credible, appropriately scoped operators. The management lens is a clear positive; the risk is not the people, it is the physics and the calendar.
No audited financials exist → traditional forensic accounting (revenue recognition, receivables vs. revenue, SBC flattering non-GAAP, goodwill) is not applicable / not assessable; financials.csv is empty. Per the +private overlay, the analogous risk surface is program/financial-execution risk, not accounting fraud:
Regulatory findings (required sub-section). Per regulatory/regulatory-findings.md:
n/a — no 10-K exists (private).There is no EPS to project. Per the +private overlay, the question that matters is distance-to-tradeable-event.
Current readiness: 2/5 (growth) — among the least IPO-proximate names on the watch. An rNPV/DCF is not meaningfully sourceable (no revenue, no contracted backlog value disclosed, single binary technical gate) → n/a; any number would be fabrication.
Milestones that unlock an S-1 (the gating ladder):
Estimated tradeable window: `` an IPO is not a 2026–2027 event. Even on an optimistic path (maiden flight late 2026, recovery 2027, cadence 2028), a public listing is plausibly 2028–2030; secondary liquidity (Forge/EquityZen) is the only near-term access and is itself thin/discounted. Write-back: private-watch.json stoke-space dossier field should be set to this dossier path so privates.ts shows the name dossier-warm; readiness stays 2/5 (no event has moved it — funding ≠ readiness). (Per strict wave boundaries, I am not editing the JSON in this run — flagged for the central step.)
Brier forecast (binary, per overlay): the scoreable bet is "Stoke Space achieves Nova orbital insertion (any payload, any orbit) on or before 2027-06-30." p ≈ 0.40 — first flights of brand-new vehicles routinely slip 6–18 months and frequently fail on debut; a late-2026 target realistically reads as a 2027 event with real failure odds, hence sub-coin-flip on this window. (forecast.ts create skipped — --watchlist unattended rule; logged here for the record only.)
Bull case. Stoke is the only Western company outside SpaceX with a credible path to FULL reuse, and it is attacking the part of the cost curve everyone else conceded — the thrown-away second stage. If the metallic-heat-shield upper stage works, the economics aren't incrementally better, they're structurally different: a couple-hundred dollars/kg vs. a few thousand, plus uncontested downmass and in-space mobility revenue no first-stage-only competitor can touch. The franchise is real today (NSSL Lane 1 + OSP-4, US-sovereign, government desperate for a Falcon-9 alternative), the team genuinely built the engines they're now iterating on, capital is abundant ($1.34B raised, USIT national-security backing), and the hardware is converging on a concrete late-2026 maiden flight (Stage-1 proto-qual done June 2026). The contrarian's prize: a second viable full-reuse provider in a world that has exactly one.
Bear case (permanent-impairment risks). (1) The upper stage may simply not come home. The actively-cooled metallic heat shield is unprecedented at orbital reentry scale — elegant on paper, un-flown in fact. If it can't survive reentry flyable, the entire 20x-cost thesis evaporates and Stoke is just another expensive expendable-ish rocket. (2) The market may not need it. The Electron→Falcon-9 medium-lift "gap" is being attacked simultaneously by Neutron, Terran R, and Falcon 9 rideshare; Falcon 9's Starlink-subsidized cadence (98 launches in 2025) makes winning on price "nearly impossible", and whether demand sustains multiple new providers "remains an open question". (3) Funding/timeline risk: capital was scoped to first flight; a 12–18-month slip forces a raise into a stale mark → down-round, dilution, or distress.
Pre-mortem (it's Dec 2027, the thesis broke — what happened?). Most likely: Nova's maiden flight slipped to 2027, reached orbit, but the second-stage reentry failed or returned too damaged to re-fly — full reuse remained a slide, not a fact. With the differentiator unproven, the late-2026/2027 cash crunch forced a flat-or-down round; secondary marks fell from ~$40 toward the low-$20s; and the medium-lift market had meanwhile been absorbed by an operational Neutron and Falcon 9 rideshare. Stoke survives as a niche national-security launcher but the category-defining outcome is gone.
Are multiples too high? No public multiple exists. The ~$1.95B primary vs. ~$40 secondary gap suggests the private market already discounts the primary mark — i.e. sophisticated secondary buyers think the headline valuation is rich for the flight risk. That's the closest thing to a "multiple-too-high" signal available.
Contrarian view (what the market refuses to see). Two opposite blind spots coexist: space optimists refuse to see that full reuse is still an unproven physics claim, not a delivered capability — Stoke has flown a 30-ft hopper, not an orbital reentry; and space pessimists refuse to see that the downmass/in-space-mobility TAM is genuinely uncontested — if Nova works, Stoke owns a market with literally no other supplier, which no Falcon-9-comparison captures.
Dismantling the bull case:
A genuine launch-and-connectivity monopoly wrapped inside an unprofitable $2T+ aspiration stock — Starlink is the real business, but at ~110x sales the market is paying for Mars, orbital AI data centers, and a $60B Cursor bet that aren't earnings yet.
A $20M-revenue EO manufacturer trading like a $1B growth story on a 200% YTD re-rate — the cost-per-satellite edge and Tether-backed balance sheet are real, but the price already discounts a Merlin success that hasn't launched, and three customers are half the revenue.
A genuinely great company and a genuinely terrible price — the only Western full-stack launch+satellite pure-play, compounding at ~50%, but trading at ~64x EV/sales with the entire Neutron thesis still un-flown. Own the business, fade the multiple.