Venture Building

Scientific founders often lack entrepreneurial training, so frontier science stalls before it becomes a company. Embed venture builders inside universities and labs to pair scientists with seasoned operators.

The rationale

In Latin America a scientist who starts a company is usually asked to raise capital, enter international markets and keep pushing the technical frontier at the same time. University culture treats entrepreneurship as a niche path rather than a viable career, so few labs produce founding teams that already contain an operator. The capital data shows where that ends: up to 2022, 65% of LATAM deep tech startups were still pre-seed or seed with under US$1m raised, and 8% had reached Series B. The science is there. The companies are what is missing.

A venture builder is an institutional co-founder. It scouts frontier science and IP, pairs scientists with seasoned operators, structures the company across IP, governance and regulatory, provides shared product and business-development resources, and co-invests to the first meaningful technical and commercial milestones. Three regional builders already run the model. GRIDX has raised US$41.5m across two funds and built 81 companies; Vesper screened roughly 4,500 scientific projects to co-found 8; Odisea reports 100 teams that have collectively raised over US$100m. Universities join only if economic value returns to the institution.

The honest objection is that a builder cannot fix a funding gap it does not control. It cannot. Non-dilutive bridge grants before Series A and at least one dedicated corporate pilot lane per hub are the missing complements, and the corporate channel is thin: CVC into LATAM deep tech peaked at US$119m in 2022 and fell to US$36.5m in 2024, roughly 70% below peak. Without early money and somewhere to run a pilot, a builder produces well-structured companies with nothing to graduate into.

The evidence

First moves

1. 0–6 months

   Stand up two builder units inside existing labs

   Select two universities or public labs with active deep tech output and co-locate a venture-builder unit that runs a three-month scientist-operator trial phase, on the GRIDX matchmaking format, with a multilateral first-loss tranche covering formation costs.

   IDB Lab + national development banks + university TTOs
2. 6–18 months

   Open a bridge-grant window and one corporate pilot lane per hub

   Run a standing pre-Series A non-dilutive grant window with a decision time under 90 days, and sign at least one corporate per hub to a pilot lane with published scope, duration and payment terms.

   FINEP, CORFO, ANPCyT, SECIHTI + corporate venture arms
3. 18–36 months

   Rewire university incentives so value returns to the institution

   Adopt a standard spinout term sheet fixing the institution’s equity share, licensing terms and revenue split, and publish spinout counts and follow-on funding per institution every year.

   University TTOs + ministries of science and technology + the LADP coalition secretariat

How we would know it worked

• Companies formed by builder units embedded in universities or public labs, reported annually by the coalition secretariat, rising from a 2026 baseline.
• Share of builder-formed companies raising a priced round within 24 months of formation, tracked through the funders registry, above 40%.
• Corporate pilot lanes live per hub, counted quarterly, at least one per hub with a signed scope and a paying corporate.
• Median time from application to disbursement in the pre-Series A grant window, tracked by the funding agency, under 90 days.

What could go wrong

Who acts

Rests on

The data behind it