Not values. Rules that constrain behavior. The set of decisions we have already made so we can spend operating time on the ones that are left.
We back founders with deep operating scars in their vertical: serial founders with 20-year vertical depth, first-time founders with a decade of being the operator who actually shipped, intrapreneurs who have lived inside the buyer. Pedigree is a poor proxy for that. Operating scars are the cheapest form of risk-reduction at the seed stage.
We build AI engines, not AI features: vertical-specific models that get smarter with every customer transaction. The flywheel: customer N's data trains the model that closes customer N+1, faster and at higher margin. B2B because the data feedback loop and pricing leverage are simply better than consumer.
We optimize for time-to-revenue, not time-to-product. A venture that ships in 4 months but takes 24 to find a paying customer is a 24-month venture. Speed without revenue validation is theatre. Capital deploys in tranches that reward each market signal. The next dollar follows the customer signature, not the calendar.
We do not write flat seed checks against vesting periods. Capital advances in tranches tied to specific product, revenue, or pilot milestones. The next dollar earns its way in. This protects the founder, the cap table, and the studio from running 12 months on auto-pilot.
The first 90 days after launch separate ventures with real ICP fit from ventures with deck fit. We expect first paying clients, or signed LOIs in regulated verticals where compliance gates the contract. The form of the signal flexes by industry; the requirement that one exists, does not.
We deliberately cap throughput. Studios that scale faster than their operating stack matures debase the very advantage that justifies the model. Discipline at the top of the funnel is the cheapest form of selection.
Not on a quarterly call. Not on a Slack channel that goes silent. In the ICP doc, the unit-economics spreadsheet, and the first ten hires. The studio earns its equity by sitting beside the founder where the work happens.
A specific commitment, not a slogan. The first ten enterprise customers must be closed by the founder personally: name on the contract, direct line for escalations, in the room for renewals. The studio brand never appears on a customer call. Studio support is operational behind the scenes, never relational on the front.
The São Paulo team operates the venture in market. The San Francisco team brings the bar: product reviews, hiring filters, capital discipline calibrated against US-tier outcomes. Neither side dilutes the other. Both anchors are real.
A studio that cannot articulate its negative space cannot articulate its positive. We turn down 95% of inbound: services-disguised-as-software, AI-feature wrappers, hype-first markets, founder profiles that score on charisma instead of judgment. Saying no is the discipline.
These principles are public for one reason: a studio that cannot publish its operating doctrine is operating one.
Read the Operating Stack →