The Frontier Batch: Four Scores, Two Gate Rejections

Six candidates from the frontier — space, quantum, warehouse automation. We scored four and rejected two at the gate. The process itself is the story: this is what a scoring model looks like when it refuses to invent what the filings don't show.

5 min read · 4 scores · 2 gate verdicts · Not investment advice

The screening round

BlitzLab's coverage has mostly lived in familiar territory: AI infrastructure, SaaS, marketplaces. This batch went further out — orbital satellite networks, quantum hardware, rocket launch, warehouse robotics, eVTOL aircraft, small modular reactors. Frontier sectors are where blitzscaling should be most at home.

Six candidates entered the screen. Four emerged with scores. Two never made it past the model's gates — G1 (product-market fit or a second growth curve) and G2 (a ten-year market ceiling at least 10x current revenue). Here a rejection is a feature, not a failure of generosity: a model that cannot say "insufficient evidence" is a model that invents numbers.

The scoreboard

None of the four clears the 65-point success threshold — though one comes within two points. Each score tells a different story about what kind of blitz is actually running.

Rocket Lab 63/B — two points short, on no debt

Rocket Lab is the rare blitz that runs without a leverage clock. Total debt of $695M against $3.49B of equity, interest income of $16.5M a quarter against $0.58M of interest expense — the demerits section scores a near-perfect zero (all figures SEC-filed, Q2 2026). The blitz is funded by equity instead: share count up from 515M to 630M, $1.94B in September ATM issuance (press-reported). The trade: no fixed-cost trap, but a dilution bill.

It's also a different kind of "build ahead of demand." Capex is decorative ($53M in H1 2026); R&D is the fuel ($163M). This is a development blitz, not a construction blitz, and gross margin is improving as the company grows (34.4% → 36.1%). The whole score is, frankly, a bet on Neutron: the medium-lift reusable rocket hasn't flown, and the launch segment is already stagnant (−4.4% YoY).

Symbotic 54/C — the P&L is efficient, the structure is not

Symbotic is the oddest structure we have scored. The P&L is past blitz: GAAP profitable (+$55M last quarter), free cash flow positive (+$243M over nine months), capex just 3% of revenue. But 84.6% of revenue comes from a single customer — Walmart (SEC-filed, FY2025) — and "the vast majority" of the $22.5B backlog (SEC's own phrase) is Walmart plus its Exol joint venture. Growth is decelerating (52% → 26% → 22%).

So this is efficiency built inside a single-customer blitz. Symbotic never spent like a blitzer — it spent Walmart's money: the customer buys the systems, Symbotic builds and sells them. We call it a customer-capital blitz. The score's ceiling is structural: one retailer's capex cycle decides the company's fate until the second blitz ramps.

AST SpaceMobile 51/C — the orbital land grab

AST SpaceMobile is the textbook "build before demand" case: a contracted backlog of ~$1.3B against trailing revenue of ~$115M — an eleven-fold ratio — and H1 capex of $859M against ~$46M of revenue, roughly 19x. The scarce resource is spectrum, orbital slots, and carrier partnerships (60+ mobile operators, a claimed 3 billion covered subscribers). Debt is pro forma ~$4.1B (~36x trailing revenue), but low-coupon convertible structure with net interest near zero and ~$3.7B of pro forma cash.

The honest qualification is in the revenue itself: Q2's $31.5M (+2,626% YoY) was gateway hardware deliveries and government milestones, not commercial direct-to-device service — which slipped to early 2027. The gate verdict was a conditional pass, applying the Cipher Digital precedent: quantified contracted demand standing in for pre-commercial product-market fit. The bottleneck is not manufacturing (six satellites a month, 95% built in-house) but the launch vehicle — after the New Glenn explosion, the company depends on SpaceX, its direct competitor.

IonQ 47/C — the acquisition blitz

IonQ's headline growth is real — Q2 revenue $80M (+287% YoY), remaining performance obligations up 297% — but the 10-Q itself attributes part of that growth to acquisitions. Management's claim of "100% organic growth" is, in our scoring, unverified: the filings say otherwise, and the score reflects it. The company is effectively debt-free, yet the capital structure's real cost is dilution: $3B+ raised in 2025, a $1.8B SkyWater acquisition paid partly in 24M new shares.

The scoring note that matters most is G2. The gate passed on arithmetic — a $2.47B bar against a quantum market that even conservative estimates exceed — but we marked the confidence as low in the verdict itself. Multiple qubit modalities compete, hyperscalers are entering, and a winner-take-most structure is unproven. The model asks whether the market is big enough; quantum is the case where the real question is whether it is a market yet.

Where the rankings disagree — and why that's honest

An external ranking by "promise" orders them Rocket Lab, AST SpaceMobile, Symbotic, IonQ. Ours flips the middle two. Not a contradiction — two different questions.

The promise ranking values option value: AST SpaceMobile's asymmetry — direct-to-device connectivity for billions of unconnected people — is the largest prize of the four, even though commercial service hasn't started. Our score values recognized reality: v1.0 measures filed revenue, filed margins, filed contracts. SYM has $2.6B of recognized revenue and a GAAP profit; ASTS has $31.5M of hardware deliveries and milestones. Promise rewards the ceiling; the score rewards the floor. A reader who wants the biggest upside and a reader who wants the most-demonstrated business are asking different questions.

The two views should sometimes diverge. If scores always mirrored the excitement ranking, one of them would be redundant.

The two rejections

Joby Aviation and Oklo never made it to scoring. Both failed G1 — and the failure is instructive, because it shows exactly what separates a conditional pass from a rejection.

Joby. Commercial service revenue: zero. The Q2's $38.6M is entirely the acquired Blade business (helicopter brokerage — a different company, a different market) plus Department of Defense contracts. FAA type certification is at Stage 5 with Joby at 20% and the FAA at 10%. Dubai, Virgin Atlantic, and Delta partnerships are real but appear in filings as frameworks without quantities or dollar amounts — no quantified demand for the actual blitz product, where AST SpaceMobile passed conditionally on $1.3B of it.

Oklo. H1 2026 revenue of $1.21M — entirely acquired engineering services; power revenue is zero, operating powerhouses are zero. The NRC combined license application has not even been filed ("evaluating the timing and form"). The strongest demand evidence is a 12GW Master Power Agreement with Switch and a 1.2GW prepayment from Meta — but the prepayment funds fuel procurement, not a power purchase commitment, and the rest is non-binding letters of intent. The company itself says it is still "converting preliminary agreements into PPAs." A gate that cannot distinguish a letter of intent from a take-or-pay lease is not a gate.

Both verdicts carry documented re-examination triggers. For Joby: the type certificate, a quantified backlog in SEC filings, or the first paid commercial flight. For Oklo: an NRC license application, a binding fuel contract, or the first powerhouse breaking ground. Rejection here is a statement about evidence, not a verdict on the companies.

What to watch

Five indicators matter most for readers tracking this batch: 1. Neutron's maiden flight — the key variable in Rocket Lab's expansion score; 2. AST SpaceMobile's beta and early-2027 commercial launch — when contracted demand must become recognized revenue; 3. the ramp of Symbotic's non-Walmart revenue; 4. separated proof of IonQ's organic growth; 5. the re-examination triggers for Joby and Oklo — certification, binding contracts, ground broken. These are observation metrics for analysis, not investment guidance of any kind.

BlitzLab scores businesses against a scholar's standard — not investment advice. No price targets, no buy/sell recommendations. Scores cited: RKLB 63/B, SYM 54/C, ASTS 51/C, IONQ 47/C; JOBY and OKLO gate-rejected at G1 (2026Q4, model v1.0 frozen). Financial figures are press-reported or filing-based as labeled; estimates are marked as estimates.