The ask. $20 million seed. Not cash in hand. Not secured.
The job. Get the first enterprise customers live on their hardware, prove the product in a six-to-sixteen-week install, and let those customers become the book that unlocks Series A.
The rule. Eighty-five percent is product, deploy, evidence, and customer-side credits. Five percent is the sales desk. This raise does not buy a public campaign.
How enterprise clients back the launch
We do not launch a product into a void and hope inbound arrives. The first customers sit a proof of value on their floor, with their data, on their hardware. Seed pays for the factory that can run that proof — and for credits on their side of the table so a CIO does not have to open a science budget to start.
| Step | What the customer does | What the $20 million pays for |
|---|---|---|
| 1. Named conversation | Fortune 100 or government committee signs an NDA | $1.0M sales desk — two account executives plus counsel |
| 2. Proof on their hardware | Six, ten, or sixteen weeks. Archive installed. Recall protocol. Fail-closed demo. | $7.0M factory + $3.0M enterprise hygiene + $3.0M air-gap path |
| 3. Credit, not a science favor | Proof fee is credited against the land | $2.0M customer-side pilot credits |
| 4. They sign. They own it. | CORE $2M · DIVISION $8M · ENTERPRISE $25M | The install stays in their building |
| 5. They expand. They back us. | More divisions, more seats, more of the archive | Those logos are the traction. They are the Series A story. |
The gate that says this worked: three signed logos or $15 million booked. Either prong. Until then, no public brand spend. If the gate misses, unused pilot credit is the first line we stop. It becomes runway — not outdoor media.
That is how enterprise clients help back the company. They do not write a check to our cap table. They write a check for an archive they own. Their install is the product launch. Their expansion is the traction.
The $20 million
| Line | Amount | Share | What it buys |
|---|---|---|---|
| Core R&D and proof engineering | $7.0M | 35% | The factory that can run a 6–16 week proof and keep the memory stack maintainable |
| Air-gap and classified-capable deploy | $3.0M | 15% | Trinity’s own disconnected install path — so a POV never sends the buyer’s data through a public cloud |
| Enterprise engineering | $3.0M | 15% | SSO, SIEM, access control, restore — Fortune 100 hygiene. The difference between a demo and a renewable contract |
| Validation and evidence | $2.0M | 10% | Independent re-run of the recall clocks and a threat model a Series A data room will accept |
| Enterprise pilot credits | $2.0M | 10% | Customer-side proof credits. First line we stop if the gate misses |
| Security and IP | $1.5M | 7.5% | Export, assignment, and custody hygiene — required before a government task order is even discussable |
| Sales | $1.0M | 5% | Two account executives and legal. A sales desk, not a street campaign |
| G&A and reserve | $0.5M | 2.5% | Thin operating reserve. Too small to absorb a nine-month zero-logo outcome — that outcome is a bridge |
| Total | $20.0M | 100% |
What seed does not buy. Billboards, street teams, city waves, influencer retainers, developer booths as the primary motion. Those belong after the contract gate, funded from Series A.
Line by line, in the room
$7.0M — the factory. The thing being sold is a runtime the buyer puts on-prem, not an API key. This line staffs the engineers who keep the memory stack maintainable and who can install it in six, ten, or sixteen weeks. Without it, the sales team has nothing to put on the floor.
$3.0M — air-gap. Sovereign tickets and government vehicles assume a disconnected install. This is Trinity’s own classified-capable path. Customer GPU rooms and SCIFs stay the customer’s capital. We are not bidding on the $58.5 billion defense AI request.
$3.0M — enterprise engineering. A Fortune 100 committee will not sign $8 million for a research daemon. Single sign-on, SIEM, role-based access, backup and restore. This is how a proof becomes an annual contract.
$2.0M — validation. Internal clocks are not a data room. This line pays an independent party to re-run recall, unique-role binding, and the tick — and to write the threat model. B-grade internal results become a folder an associate can hold.
$2.0M — pilot credits. How a takeaway sale starts without asking the CIO for a net-new science line. CORE / DIVISION / ENTERPRISE proofs are creditable against the land ($150k / $400k / $1.0M in the proposed book). Credits sit on the customer side. If Phase 0 fails, this is the first line we stop.
$1.5M — security and IP. Export-control counsel, facility procedures, patent and assignment hygiene. A vendor that wins a $323,090 SBIR and cannot touch the later task order has wasted the vehicle.
$1.0M — sales. Two account executives plus legal. Fully loaded commercial cost to open a named account is about $0.9 million. Sovereign pursuit costs more; the rest of Year-1 selling sits in operating expense once the company is converting. Seed does not pretend a million dollars buys a national campaign.
$0.5M — reserve. Entity, books, insurance. Intentionally thin.
What this raise must produce
| Output | How we know | Which lines pay for it |
|---|---|---|
| Proof factory live | A 6–16 week install we can repeat | $7M R&D + $3M enterprise + $3M air-gap |
| First customers signed | 3 logos or $15 million booked | $1M sales + $2M credits |
| First federal vehicle | SBIR Phase I and/or a prototype OT | $1.5M security/IP + air-gap |
| Evidence pack | Independent re-run of the clocks | $2M validation |
| Series A eligibility | $80M run-rate or 15 logos + one $15M government task order | Seed enables the trigger. Series A funds the scale. |
Year-1 base recognized revenue of $48 million on $96 million booked is the operating plan if those outputs exist. It is not a guaranteed result of spending $20 million. Downside recognized is $10 million. Upside is $80 million. Full tables: ../05-model/FINANCIAL-PROJECTIONS.md.
Use of funds is a stock. The Year-1 plan is a period. On the base path the company consumes $8 million of cash at 42 heads. On the downside it consumes $14 million. Leftover seed does not become a street campaign.
Series A — after the customers exist
| Round | Amount | Use | Trigger |
|---|---|---|---|
| Seed (this raise) | $20 million ask | Factory, first proofs, first logos | Document only — not cash on hand |
| Series A | $250 million ($200–300 million band) | Scale engineering, national programs, international sales, public brand | $80 million recognized run-rate or 15 logos plus one government task order of $15 million or more |
The $200–300 million band is raise size, not seed pre-money. Series A is where scale lives. Seed is where the first enterprise clients install the archive and become the book.