Economic analysis · Introductory

Build versus license

Two routes to the same capability question: license access to a platform that already works, or fund a programme that might produce one. This page sets out the comparison honestly, including what it cannot yet support.

Document status: Introductory, not a business case
Figures: Internal assumptions, no quotation

Proposed licence

$20M

Three-year term · customised experience · no source code, no transfer of ownership

Working hypothesis

Independent development scenario

$100M–$300M

Illustrative range · full ownership only if the programme succeeds

Illustrative scenario

Dimension	Licensed access	Independent build
Capital committed	$20M across three years, staged by agreement	$100–300M illustrative scenario, before overruns
Time to first working capability	Integration and evaluation, then a pilot, with no research phase	Multi-year research, training and infrastructure programme
Ownership	Right to use. No source code, no weights, no transfer	Full ownership, if the programme succeeds
Depth of customisation	Customised experience and integration layer	Architecture-level modification, including the model itself
Operating burden	Vendor operates the platform; we operate the integration	Permanent ownership of training and serving infrastructure
Talent required	Integration, data and evaluation engineers	Frontier research, distributed training, large-scale infrastructure
Downside if the thesis is wrong	Bounded to the licence term plus integration spend	Large sunk capital with limited salvage value
Position at end of term	Renew, renegotiate or exit. No residual asset	Asset continues to be owned and improved internally
What the money buys	Access to capability that already exists	A capability that may or may not materialise

What these numbers are not

The $100–300M independent development figure is an illustrative scenario. It needs detailed cost modelling of compute, data, talent, evaluation and time before it can carry weight in a decision.
A three-year licence is not equivalent to owning source code. It buys access for a defined period, with defined limits, and it ends.
The $150M source-code valuation hypothesis is a preliminary internal frame for negotiation, not a valuation opinion or a market price.

Source-code valuation hypothesis

$150M

Preliminary internal anchor for negotiation. Not a valuation opinion and not a market price.

Working hypothesis

Detailed economics dashboard: planned

The full dashboard is a Phase 2 deliverable. In the interim, the scenario envelopes and path analysis are held in the strategic assessment.

Interim scenario analysis →