Headline: Auron's canonical rating was 6.0 (Active DD), explicitly contingent on the data room substantiating Daniel Stepanenko's verbal claims from the 5/28 call. The data room arrived as a single 12-page deck refresh. Every forensic flag from the 5/13 deck review ($155M OTA framing without "Phase 1 proposal pending" qualifier, "Ex-Palantir" badge the founder himself walked back on the call, CTO title inflation, accelerators-plus-customers-plus-employers conflated as "PARTNERS AND INVESTORS") is unchanged. A forensic second-pass re-read of the 49 KB call transcript then surfaced three additional first-order inconsistencies stated on the same day: Intellias revenue (deck $25M vs transcript $125M, 5x), GD OTA value (deck $155M vs transcript $1.2B, 7.7x), and customer cohort drift across a 6-minute span. Rating drops to 4.0 watchlist, between Delivr (failed DD) and the original 5/13 hard pass. Pass at $20-26M cap. Reopen only on real documents.
What changed since 5/13
The 5/13 deck-only forensic review rated Auron 2.5 (HARD PASS) on padded founder bios, an unverifiable $155M GD/Navy OTA, DoD seal usage, and Webflow placeholder investor logos.
The 5/28 founder call with Daniel Stepanenko reset the picture. Daniel presented credibly on his Ukraine SOF background, his brother and family's active military duty, the 93rd Mechanized Brigade field-testing relationship, the $25K/mo burn, the $3M closed at a $20-26M cap, and the GD OTA Phase 1 proposal pending government award. Mark, Joe, and Nick reacted positively. Daniel declined LOIs as "the most BS I've ever seen," offered a family-office reference call, and acknowledged ITAR is not yet obtained. Daxos re-rated to 6.0 (Active DD), explicitly contingent on the data room substantiating the call.
The data room arrived June 6, 2026 as a single 13.7 KB pitch deck PDF. Nothing else. The contingency failed.
Claim-by-claim verification
For each material claim Daniel made on the call, we checked the refreshed deck and the broader data room for independent corroboration:
| Call claim (5/28) | Data room evidence | Verdict |
|---|---|---|
| GD OTA Phase 1 proposal accepted, pending government award (~$1.2B program; 4,000 units × $300K) | Deck p.3: "$155M OTA WITH U.S. NAVY · Prime led Phase 1 OTA Demo unit / $300K demo costs". No award letter. No SAM.gov number. No GD contracting officer name. No "Phase 1 proposal pending award" qualifier on the deck. | Partial · same flag as 5/13 |
| Chevron pilot announced 5/27 | Deck p.3: "Chevron $150K INVESTMENT + PILOT". No PO. No pilot agreement. No Chevron press release link in the data room. | Plausible · undocumented |
| Lockheed SBIR Phase 1 ~$3M under discussion | Deck p.3: "ACTIVE $3M SBIR DISCUSSIONS · LOCKHEED MARTIN · SBIR Phase 1". Not awarded. No solicitation number. | Confirmed as aspirational |
| SAM.gov entries (any contract number, UEI, CAGE code) | Nothing in deck or data room. | Absent |
| $3M lifetime raise; Mitsubishi, Hitachi, Locke Adventures are strategic VCs | Deck p.2 shows Techstars W26, Plug and Play (accelerators). No Mitsubishi. No Hitachi. No Locke Adventures. No cap table. No SAFE docs. No wire confirmations. PitchBook on 5/13 showed Techstars at $200K as sole investor. | Direct inconsistency |
| $500K bookable revenue | "$10M discussions" with Ukrainian brigades (not revenue), $150K Chevron pilot, $300K GD demo unit (not paid). No invoices. No AR aging. No customer POs. | Bookable ≠ booked ≠ collected |
| TRL 6 / 6 months Ukrainian field testing | Deck p.3: "BATTLEFIELD PROVEN / 93RD MECHANIZED BRIGADE / SOF & RECON UNITS / FRONTLINE & DEEP OPERATIONS". Daniel referenced prototype videos in the data room on call [29:13]. No videos in the data room JSON. | Plausible · unverified |
| Founder bio: 2.5 yrs Palantir, 3 yrs Intellias (Head of Defense, $25M revenue, 34-person team) | Deck p.2: "Ex-Palantir" / "Former Head of Defense Program at Intellias, built $25M in revenue with a 34-person team". Daniel softened Palantir on the call [18:39-19:01]: "a couple of years... mainly junior roles... was a PM." | Deck inflates what founder walked back verbally |
| DoD / Navy / Air Force seal compliance | Deck p.2 logo row includes "U.S. NAVY" / "U.S. AIR FORCE" / "DOD" under "PARTNERS AND INVESTORS". The structural problem (federal service marks under "PARTNERS AND INVESTORS" implies endorsement, 32 CFR § 507) is unresolved. | Same flag as 5/13 |
| Three software patents | No USPTO numbers in data room. No provisional application numbers. No patent docs. | Absent |
| Santa Venture / Fort Lauderdale family office reference (offered on call) | No introduction. No reference letter. No follow-up after the call. | Absent |
| ITAR / EAR registration | Daniel explicit on call [29:31]: "we do not yet have ITAR." Not in the data room either. He says they don't charge Ukrainian units to sidestep export controls, which is correct mechanically but caps near-term revenue. | Confirmed: not obtained |
Forensic re-read (6/9) · same-day quantitative contradictions
A second-pass close read of the 49 KB Fireflies transcript surfaced three first-order inconsistencies the deck-only review missed. These are not 5/13-vintage flags. These are new contradictions stated by the founder on the same call to the same audience within minutes of each other.
- Intellias revenue: 5x discrepancy on the same day. Deck p.2: "Former Head of Defense Program at Intellias, built $25M in revenue with a 34-person team." Transcript [19:46-19:55]: "We had 6 million in revenue when they joined. In three years we had 125 million in revenue when, when they exited." $25M deck vs $125M transcript. Same metric, same day. This is either fabrication or extreme carelessness with numbers under pressure.
- GD OTA value: 7.7x discrepancy on the same day. Deck p.3: "$155M OTA WITH U.S. NAVY." Transcript [17:18-17:22]: "this ota is for 4000 units. The single unit price is 300k for us" → $1.2B by the founder's own math. Deck and call do not reconcile. Both can't be right.
- Auron's take is only 10-20% of the OTA, per the founder himself ([14:47-14:49]): "they [primes] get 80% of the share. Usually we get 10, 15, 20." So the headline $155M (or $1.2B) is GD's number, not Auron's. Auron's lifetime program revenue, if every unit ships, is $15-30M on the $155M framing or $120-240M on the $1.2B framing. Either way the headline marketing number is not Auron's revenue.
- "35 OTAs in conversation with GD" framing is wrong. Transcript [17:55]: "35 OTAs with JD where we apply" — APPLIED to, not active engagement. The post-call memo framing of "35 in flight" overstates this.
- Burn doubling next month, runway is months not years. Transcript [34:38] + [35:06-35:10]: only 2 of 6 are paid; 4 are equity-only volunteers; plan is to flip everyone onto salary, doubling headcount to 10-12 in 6 months. Effective burn going from $25K/mo to ~$65K/mo. On $450K + $500K bookable that is 14-15 month runway, not "years."
- ITAR reasoning is incorrect under federal export controls. Transcript [29:31-29:49]: "In Ukraine there is no Itar. So like until we start charging them, we can test invalidate whenever we like." Under 22 CFR §120.10 the technical-data definition does not require payment for a "deemed export" to occur. Providing RF/EW software-as-a-service to a foreign military without ITAR registration is a regulatory exposure regardless of whether Ukraine pays. The "we don't charge so it's fine" framing is wrong, not just incomplete.
- Customer cohort drift inside a 6-minute span. Opening [00:08]: Mitsubishi, Hitachi, "Locke Adventures" named as both competing strategic VC checks and active pilot customers. Traction slide context: Lockheed, GD, Chevron framed as biggest customers. Two different customer cohorts named within minutes. "Locke Adventures" is not a known investor — likely Fireflies mistranscription of "Lockheed Ventures," which would conflict with the deck's separate Lockheed SBIR Phase 1 discussion framing.
- Intellias tenure narrative changed on the call vs prior 5/13 LinkedIn finding. 5/13 forensic review showed Daniel's only verifiable Palantir signal was a 2025 Fellowship finalist + a 2019-2020 outsourced BD contract. Transcript [05:11]: "Worked in Palantir for two years two and a half." If only the Fellowship is real, "2-2.5 years" is fabrication. If the 2019-2020 BD contract counts toward it, the deck's "Ex-Palantir" senior framing remains misleading.
Net effect: the qualitative corroborations the call DOES provide (technical fluency on IQ samples and SDR, software-only license model, field-iteration deployment loop, partner sentiment positive at Daxos) are real and prevent a deeper cut. But quantitative-claim contradictions of this magnitude (5x and 7.7x discrepancies stated by the founder himself in the same hour) outweigh qualitative-to-process corroborations. The "call credibility uplift" that took the rating from 2.5 (5/13 deck) to 6.0 (5/28 call) is mostly clawed back.
Key strengths
- Real founder, real Ukraine connection. Daniel served in Ukraine SOF (call [05:47]). Brother and family on active duty. Frequent battlefield access to the 93rd Mechanized and other brigades. This is not LARP. Mark flagged it on the call as the strongest single signal in the file.
- Software-only, hardware-agnostic positioning is genuinely defensible. Selective non-kinetic defeat at the RF layer, integrating into Palantir Gotham / Anduril Lattice / Lockheed C2 as the autonomy module, is a credible thesis. Deck pages 5, 6, 10 lay this out coherently.
- Capital efficient. $25K/mo burn with 2 FTE plus 4 equity-only contributors (call [34:38]). On $3M (claimed), runway is years not months.
- Customer pull pattern is consistent narrative-to-narrative. GD invited them after Plug and Play demo day [16:22], Chevron pilot announced [06:26], 35 GD OTAs in flight [17:54]. Even unverified, the pattern of inbound prime/customer interest is consistent across the call and the deck.
- Defense-tech tailwind is real. Counter-drone, non-kinetic defeat, infrastructure protection, Golden Dome adjacency. Sector wind at his back.
Key flags · unresolved
- "Data room" is a 12-page pitch deck. Nothing else. No financials, no contracts, no cap table, no award letters, no SAM.gov entries, no patent filings. This is the central problem.
- $155M OTA framing on deck p.3 is unchanged from 5/13 despite the call clarifying it is "Phase 1 proposal, pending government award." Deck still markets it as a dollar-figure award. Highest-stakes integrity item.
- Investor logo wall conflates accelerators, customers, integration partners, and former employers into "PARTNERS AND INVESTORS" (deck p.2). Same pattern flagged 5/13. Mitsubishi, Hitachi, Locke Adventures (the strategic round Daniel cited on the call) do not appear on the refreshed deck.
- CTO title inflation flagged on 5/13 is unchanged. Deck p.2 still says "Former L3Harris Lead RF / Electronic Warfare Engineer". 5/13 forensic review showed actual title was "RF/Wireless Engineer."
- "Ex-Palantir" badge on deck p.2 vs. founder's own walk-back on call [18:39]: "mainly junior roles... was a PM." Badge is technically real (he was there) but the deck implies seniority the call disclaims.
- ITAR not obtained. Regulatory exposure item for a company selling RF/EW software with defense end-use and Ukrainian field deployment. Caps near-term revenue.
- Three patents claimed verbally, zero USPTO numbers in data room. Easy to verify; not provided.
- Family office reference call (Santa Venture / Fort Lauderdale) offered on the call but not delivered. The cheapest piece of independent verification, still not produced.
Math · comparable companies
Final = Base + EV + Geo = 2.5 + 0.5 + 1.0 = 4.0
Base 2.5 · documents unchanged from 5/13 hard pass; the call uplift that took the base to 4.5 (then total 6.0 contingent) is clawed back by new same-day contradictions surfaced in the forensic re-read (Intellias 5x, OTA 7.7x, customer cohort drift, ITAR misreading).
EV +0.5 · $3M closed (claimed, unverified) sits in the $3M-$5M tier per Rule 21.
Geo +1.0 · USA (Miami HQ, Delaware C-corp).
| Co | Rating | Sector / shape | Why Auron sits where it does |
|---|---|---|---|
| Askari Defense | 8.5 | Defense, US, real awards | Documented program traction. Different tier. |
| Sophia Space | 7.5 | Defense-adjacent, US, NVIDIA cohort | Real Series A structure, disclosed. Better data package than Auron. |
| Kragon Space | 7.0-7.5 | Defense / dual-use, US | Multi-USG documented, NATO / EU pipeline. Auron's claims are similar shape, unsupported. |
| TeroAI | 5.5 | Defense, real product, transparent founder | Better documented than Auron at half the cap. |
| Delivr | 4.0 | Failed data room DD | Financials didn't reconcile. Auron sits at the same level: real founder, but contradictions surfaced under inspection. |
| Auron (today) | 4.0 | Defense SW, real founder, undocumented + same-day contradictions | Watchlist at Delivr level. Pre-call deck-only floor was 2.5; remains anchored there with EV and geo bonuses. |
| Auron (5/13) | 2.5 | Original deck-only HARD PASS | Forensic flags. Documents unchanged today. |
Decision recommended
Pass at $20-26M cap as currently presented. The cap implied by Daniel ([01:19], [02:16]) is at the top of Daxos's seed/pre-seed range. We are being asked to write a check on documentation that is materially the same as the deck that earned a 2.5 hard pass three weeks ago. The marginal uplift from the founder call (credibility, ground truth, Ukraine connection) is real but does not justify a $250K check at a $20M+ cap without a single contract, cap table page, or SAM.gov entry to anchor the marquee revenue claims.
Conditions that would reopen at $250K
- GD OTA Phase 1 documentation: proposal acceptance letter from GD's contracting office, SAM.gov entry under GD's UEI or Auron's as subcontractor, Phase 1 task order PDF.
- Chevron pilot PO or signed pilot agreement.
- Cap table showing Techstars, Plug and Play, Santa Venture, plus the Mitsubishi / Hitachi / Locke Adventures strategic VCs Daniel named on the call, with closed amounts and pre-money.
- USPTO numbers for the three software patents (or provisional application numbers).
- Reference call with Santa Venture (Fort Lauderdale family office) as offered.
- Corrected deck: $155M reframed as "Phase 1 proposal pending", investor wall separated from partner / customer wall, CTO title corrected, DoD / Navy / Air Force seals replaced with non-seal text.
If items 1, 2, and 3 land AND the Intellias revenue and GD OTA value discrepancies are cleanly reconciled in writing, rerate to 5.5-6.0 and write $150K-$250K. If items 1+2+3+5 all land plus the reconciliation, rerate to 6.5-7.0 and write up to $500K. Re-rate DOWN to 3.0 if any single one of items 1-5 comes back unverified or refused.
Walk-away triggers
- GD OTA Phase 1 documentation does not exist or cannot be produced.
- Cap table shows Techstars + Plug and Play accelerator equity only (no Mitsubishi / Hitachi / Locke Adventures actually closed).
- DoD / Navy / Air Force seal usage remains on the deck after explicit feedback.
- Pattern repeats: more verbal updates, no documents.
Sources:
Method: Daxos batch DD scoring formula (Rule 22):
Document version: 2026-06-09.
auron-solutions.json in the DRA project corpus, sources [0] refreshed deck (13.7 KB extracted, June 6, 2026), [1] Fireflies transcript "Auron <> Daxos" (49 KB, May 28 founder call with Daniel Stepanenko), [2] post-call rating update memo (May 28, 2026).
Method: Daxos batch DD scoring formula (Rule 22):
Final = Base + EV + Geo. Ratings anchored to Daxos benchmark distribution (Rule 8). Original 5/13 deck-only review and 5/28 founder call covered in reference_daxos_benchmark_distribution.md. This rerate triggered by (a) data room arrival 6/6 and the failure of the verification contingency, and (b) a 6/9 forensic re-read of the 49 KB transcript surfacing three new first-order same-day contradictions (Intellias revenue 5x, GD OTA value 7.7x, customer cohort drift). Note on the source count: DRA's UI may display ~45-47 "chunks" for this project (two-tier retrieval splits 67,885 content chars into ~45 chunks at 1500 chars each). That is chunk count, not source count. The Auron corpus is and remains 3 sources (deck, transcript, post-call memo). No additional files exist outside what is ingested; an exhaustive sweep of the droplet (Gmail caches, Drive folders, archive backups, all subdirectories) confirmed this.
Document version: 2026-06-09.