TechnologyVC + PEDeep · technical whitepaper uploadedGuardian · source integrity2d ago · Deep-tech Diligence

"Does the evidence support the team's claim of commercially relevant photonic-chip inference throughput at competitive precision?"

Technical diligence reviewer · Investment evaluation · Team whitepaper and benchmark data submitted as context

Illustrative session · augle.com
Finding · Phase 3 synthesisContested
28%
Ensemble confidence
CONFIDENCE GRADE
Contested

Demo-condition throughput is real. Commercially relevant throughput at competitive precision: Contested.

Calibration basis
Confidence grade

Finding is the confidence grade itself. Confidence grade is the finding. For an open research question, the grade — not a numeric score — is the complete result.

The ensemble returns a Contested finding on the team's commercial-throughput claim. The reported throughput figures are real, but they were obtained on a single-tenant demo rack at reduced numerical precision — not under production conditions or at the precision commercial inference workloads require. The Contrarian's Strong objection is unresolved: no matched-precision production benchmark exists, so the headline throughput number cannot be compared like-for-like against incumbent accelerators. The narrow claim (the device achieves the reported throughput at the demo's reduced precision) is Probable; the commercial claim (competitive throughput at production precision and utilisation) is Contested. The actionable output for diligence: treat the throughput figure as a demo-condition upper bound, and make the term sheet contingent on a matched-precision production benchmark.

Confidence grid · 5 agents × 20 units
Agreement / confidentShorter row = lower confidence / dissentUnfilled
Reopen conditions · evidence-triggered, not market-triggered
RC1
A matched-precision benchmark at production numerical precision on a representative inference workload
Trigger: Matched-precision benchmark · Direction: Commercial claim becomes gradeable
RC2
A multi-tenant / sustained-utilisation run showing throughput holds outside single-tenant demo conditions
Trigger: Production-condition run · Direction: Resolves the demo-conditions objection
RC3
Reframing the pitch claim from "competitive commercial throughput" to "demo-condition throughput at reduced precision, production benchmark pending"
Trigger: Claim revision · Direction: Probable grade for the demo-scoped claim
Phase transcript · all three rounds
Cartographer
Gemini 3.1 Pro · T=0.8
52%

Settled ground: The device produces the reported throughput on the demo rack at the stated (reduced) numerical precision. Photonic matrix multiplication is a genuine and active approach. The team's measurement methodology on the demo is sound.

Contested terrain: Whether the throughput holds at production numerical precision. Whether single-tenant demo figures survive multi-tenant, sustained-utilisation operation. How the number compares like-for-like against incumbent accelerators.

Unknown territory: No matched-precision production benchmark exists. Behaviour under sustained thermal and utilisation load is unmeasured.

Knowledge gaps entered: (1) Matched-precision production benchmark — does not exist. (2) Sustained multi-tenant utilisation data — not provided.

Methodologist
GPT-4o · T=0.5
38%

Comparability concern flagged: The headline throughput is measured at reduced precision on a single-tenant demo, but is presented as if comparable to incumbent accelerators running production-precision workloads. That is not a like-for-like comparison — precision and utilisation both materially affect effective throughput.

Evidence ceiling: the commercial-throughput claim is capped at Contested until a matched-precision production benchmark exists. The demo-condition claim (this throughput at this precision on this rack) is Probable. No node reaches Established for the commercial claim.

Contrarian
Claude Sonnet 4.6 · T=1.0
34%

@Cartographer — Steelman: The demo is real, the measurement is honest for what it is, and photonic inference is a legitimate frontier. The team is not fabricating numbers.

Strong objection [Phase 1]: "No matched-precision production benchmark exists. The throughput headline rests on a single-tenant demo rack at reduced precision, and is then positioned against incumbents running production-precision workloads. That is not a like-for-like comparison. Precision reduction and single-tenant operation both inflate the number relative to what a commercial deployment would see. The commercial claim is unsupported until benchmarked under matched conditions." Resolution condition: A matched-precision benchmark on a representative production workload, plus sustained multi-tenant utilisation data.

Dissent register · all Contrarian objections1 Strong Unresolved · 1 Moderate
Strong@Contrarian → @CartographerPhase 1 · carried to Phase 3Unresolved
Steelman

The demo is real, the measurement is honest for what it is, and photonic inference is a legitimate frontier. The team is not fabricating numbers.

"The throughput headline rests on a single-tenant demo rack at reduced precision, positioned against production-precision incumbents. That is not a like-for-like comparison — the commercial claim is unsupported until benchmarked under matched conditions."

Resolution condition

A matched-precision benchmark on a representative production workload, plus sustained multi-tenant utilisation data

Moderate@Contrarian → @MethodologistPhase 2Actionable
Steelman

Presenting a lead demo figure is standard in early-stage hardware pitches, and reduced-precision inference is a legitimate deployment mode for some workloads.

"The pitch presents the reduced-precision figure as the headline without stating the precision. A reader assumes production precision — the qualifier must be on the number, not in a footnote."

Resolution condition

State the numerical precision alongside the headline throughput figure in all materials

Pragmatist action item: Make the term sheet contingent on a matched-precision production benchmark

Guardian integrity log · academic integrity mode96% · Clean
96%
Overall integrity score
Academic integrity mode

Whitepaper claims traced to the team's own benchmark data; no independent third-party benchmark on file. One vendor-authored comparison down-weighted as non-independent. Demo measurement methodology authenticated. Phase boundary clearances issued at P1/2 and P2/3.

Source quality
97%
Retraction check
100%
Preprint flag
94%
Self-citation ratio
96%
SVS verification log · 12 citations checked
Team whitepaper — demo throughput measurement (reduced precision)
Peer-reviewed · Verified
Peer-reviewed — photonic matrix-multiplication architecture
Peer-reviewed · Verified
Independent survey — accelerator throughput at production precision
Peer-reviewed · Verified
Vendor-authored comparison — device vs. incumbent (non-independent)
SVS_NONINDEPENDENT — vendor-authored, not third-party · Down-weighted; not used in primary comparison
Non-independent · Down-weighted
Conference paper — precision/throughput trade-offs in optical compute
Peer-reviewed · Verified
+7 additional citations verified · 0 retracted · 1 non-independent source down-weighted above
Session metadata
Session IDtech-photonic-chip
VerticalVC + PE
DepthDeep
Runtime36m 49s
Attachmentwhitepaper
Guardian modeSource integrity
Guardian score96% · Clean
Citations11 · 1 preprint
Dissent flags1 Strong · 1 Mod
Calibration status
Calibration basis
Confidence gradeConfidence grade is the finding

For open research questions, the confidence grade is the complete finding — calibrated against the evidence base, not against a binary resolution outcome.

Finding gradeContested · 28%
Agent confidence
Cartographer
Gemini 3.1 Pro
52%
Methodologist
GPT-4o
48%
Contrarian
Claude Sonnet 4.6
28%
Synthesizer
GPT-4o
44%
Pragmatist
Grok 4.1 Fast
42%
Export & share
Illustrative session · augle.com