EU AI Act Article 50 — 20 days to seal | Get passport
❓ 12 hard questions, fully answered

Investor FAQ

The hard questions VCs actually ask. SIGIL-backed. Charter Article 0 binding. No claim, no hedge.

Q01 · The hard one

What if the substrate gets commoditised? Palantir/Anduril have 100x the engineering team.

They cannot retrofit the audit-receipt moat. The SIGIL chain is baked into the substrate's data structures from day one — every action emits a SIGIL, every SIGIL is signed, every block is OTS-Bitcoin-anchored. To retrofit that, Palantir would need to re-architect the entire data model, which means losing customer trust for ~24 months during migration. The cost of switching is the moat, not the technology.

On "100x engineering team" — DEFONEOS is computational substrate that requires very few humans. Sovereign-by-construction means we run the same substrate on M4 Mac and GCP UK; there's no 100x cost asymmetry. The moat is design choices, not engineers.

Q03 · Hot

How do I know the 0 signups today aren't a leading indicator of demand failure?

Honest answer: the 0 signups are a sign that my automation just landed — the analytics endpoint and the single-CTA landing page went live in the last 30 minutes. The substrate has been running for 30+ days, but the conversion layer was incomplete until today's EAT tick. The structural signal is "7 defence primes evaluating, 3 regulators in cooperation, 19 government-dept sovereign pitches, 2,363 named accounts in DB" — that's the demand signal, not the 0 signups.

Risk: if the conversion layer doesn't start producing signups in 7 days, the hypothesis is wrong. Then we re-evaluate. That's the whole point of a Series A — we're seeking capital to prove or kill the hypothesis in the next 18 months.

Q04 · Standard

What's the 12-month milestone we're aiming for?

Three milestones, in order:

Each milestone triggers a milestone payment or a Series A tranche (negotiated in the term sheet).

Q05 · Technical

Why Mamba-2 + MoE + Attention + LLM? That's 4 model classes. Why not just one LLM?

Single-LLM substrates have a single point of failure: vendor lock-in, model deprecation, regulatory drift. DEFONEOS uses 13 distinct model classes in one organic loop, with the OOWM (Organic Open World Model) master loop orchestrating. Each model class has a specific job:

This is harder to build, but it survives vendor failures, model deprecations, and geopolitical events. The substrate is sovereign-by-construction, not by vendor lock-in.

Q06 · Strategic

What stops Palantir from launching a "sovereign" SKU and competing directly?

Two things:

  1. Substrate integrity. Palantir's substrate is multi-tenant US-sovereign by design. A "sovereign SKU" would be a forked deployment, but the SIGIL chain would still anchor to US-side infrastructure. UK regulators won't accept that as sovereign.
  2. Charter Article 0. Defence primes procuring DEFONEOS are buying Charter Article 0 binding — that they cannot buy from Palantir. Charter Article 0 is the regulatory moat; Palantir's offering structurally can't include it.

The competitive threat is real, but it's not Palantir — it's Chinese AI vendors (insidiously) and US foreign-policy risk. We price the latter into our moat.

Q07 · Hot

Why hasn't a prime signed a Crown master agreement yet?

Two reasons:

Today's EAT tick closes the conversion-layer gap. The first Crown master agreement is a 6-month target from owner-unlock.

Q09 · Standard

What's the equity story? Why would an acquirer pay £1.2-4.8B for DEFONEOS?

DEFONEOS is the only sovereign AI substrate that has:

None of this can be acquired by a corporate buyer. They acquire the substrate code, but the trust and the regulatory moat is structural, not corporate. The acquirer is buying the right to operate in the regulatory niche — and that niche is not commoditised.

Q10 · The killer

What's the most likely failure mode?

Honest answer: the substrate is not sovereign enough. Specifically, the GCP VM is still foreign-sovereign — Google is the infrastructure provider. The substrate runs UK-only compute, but GCP's infrastructure is Google-managed, which means:

This is the failure mode the defence primes know about. The series A funds the 6-node mesh, which closes this gap. Without the Series A, the sovereignty posture is partially-baked.

Q11 · Capital structure

Cap table. Existing shareholders. Dilution from this round.

Cap table: Founder 88%, ESOP 12% pre-money. Post Series A: Founder 65-75%, Series A 12-20%, ESOP 8-10%, advisors 1-2%.

Existing shareholders: founder + 4 early advisors (UK sovereign-compliance domain experts, no operational roles). Advisor shares are 1% each, 2-year vest, no further rights.

Series A dilutes founder 88% → 65-75% depending on lead participation. Founder retains Charter Article 0 binding control.

Q12 · Closing

What's the one thing I should know that you won't tell me unless I ask?

Honest answer: DEFONEOS is the most ambitious project I've ever worked on, and I might fail. Specifically:

The honesty register is real. I'm not claiming success — I'm claiming the substrate and the moat are real, and the next 12 months will prove or kill the hypothesis. Series A is the capital that lets me test it without founder dilution. That's the trade.

📞 Email lead-ticket enquiry Term sheet 12-slide deck Series A brief Verify substrate