Where is the new money going inside Latin American deep tech? Two figures answer that between them, and they only work as a pair. Biotechnology accounts for 61% of the region’s deep tech companies and artificial intelligence for 11%, 72% of the mapped population between them, on the IDB’s 2023 mapping Deep Tech: The New Wave. That split is current rather than historical: EMERGE and Cubo Itaú’s Radar Deep Tech LATAM 2025, presented on 11 September 2025 across 1,316 companies, reproduces the same shares by company count. Then, on 8 July 2026, Dealroom’s Latin America deep dive reported that AI-core companies founded between 2023 and 2025 captured roughly 70% of the rounds those cohorts closed.

Read as one number, the two look incompatible. Read as two measures, they fit. A company count is cumulative by construction: every firm founded across the span a mapping covers contributes one unit, whether it raised a round or nothing at all, so biotechnology’s 61% records accumulated formation. A share of rounds is a period measure that resets and responds to recent investment committee decisions. The first tells you which companies exist in the region; the second tells you which ones are clearing rounds right now.

Biotechnology built the base, and the base is worth owning

The 61% is not an accident of classification. The IDB attributes the region’s biotech density to endowments no policy can conjure and no rival region can copy:

Abundant specialized talent in biological sciences, the international competitiveness of the agricultural sector, and the remarkable biodiversity that serves as a resource for researchers.

That base has produced companies that clear global bars. Establishment Labs, founded in Costa Rica in 2004 and listed on NASDAQ, makes the Motiva implant line and carried a market capitalisation of about US$2.59 billion on 4 July 2026, well above the US$1.8 billion our report takes from the IDB’s 2023 read; it remains the region’s most valuable deep tech venture. In Brazil, brain4care’s non-invasive intracranial-pressure monitoring made it EMERGE’s largest Brazilian deep tech recipient of 2024, with US$23.6 million raised in total, and Symbiomics, out of Florianópolis, was the third largest at US$2.7 million in 2024. Argentina’s Puna Bio, backed by the Gates Foundation and SOSV/IndieBio, and Uruguay’s Antarka, a Universidad de la República spinoff, work the same endowment through extremophile microbes. Brazil’s Future Cow, one of the country’s first precision-fermentation startups, raised US$885,000 in 2025 from Antler and Big Idea Ventures. Company totals in this post come from the report and from named third-party sources rather than a fresh pull, and stand as reported unless we date them.

Beyond the two leaders, the IDB’s 2023 split puts nanotechnology at 6%, cleantech at 5% and spacetech at 4%, shares we carry as extracted and not yet re-checked. Spacetech outweighs its count: Satellogic, out of Buenos Aires, has raised about US$249 million across seven rounds, most recently a US$35 million registered direct offering that closed on 27 January 2026.

The newest cohorts are clearing their rounds in AI

The Dealroom figure repays close reading, because its basis is what makes it usable. The roughly 70% is a share of rounds rather than of dollars, it is explicitly approximate, and it covers only rounds closed by companies founded between 2023 and 2025, a cohort defined by founding year rather than a share of all deep tech financing in the region. Stated strictly: among companies young enough to be raising their first institutional capital, the AI-core ones are the ones getting rounds away. Restating that as “AI took 70% of Latin American deep tech funding” changes the measure, the universe and the currency in a single move, and Dealroom’s Latin America page is a live dashboard, so the value is as of the date it was pulled.

The names are already visible. Carecode, in Brazil, raised a US$4.3 million pre-seed in 2025 led by Andreessen Horowitz with QED Investors, for AI agents that run healthcare operations. Tractian took a US$120 million Series C led by Sapphire Ventures, with General Catalyst and Y Combinator behind it, for industrial AI that puts predictive maintenance on factory floors. In Chile, Splight raised a US$12 million seed led by noa, with EDP Ventures and the UC Berkeley Foundation, for grid-operations AI that cuts curtailment and congestion so more renewable power reaches load. Chile’s largest deep tech recipient by cumulative funding is NotCo, which EMERGE puts at roughly US$466 million, about 75% of the Chilean total, backed by Tiger Global, Bezos Expeditions and Kaszek.

Corporate venture funds are naming AI first

Corporate investors point the same way. As of 2025, 41% of Latin American corporate venture funds put AI first, against 31% for electric mobility and 31% for sustainability and the energy transition. That reaches us through a secondary read rather than a primary filing, and the year comes without a month, so carry it as reported.

What the same corporate money did inside deep tech is a separate series, and it belongs on the page in full. Sourced to Dealroom, our report publishes it as:

2020: US$9 million

2021: no figure published

2022: US$119 million

2023: US$73.9 million

2024: US$36.5 million

We carry 2021 as no figure rather than as zero, because printing a nought there would publish a zero-CVC year nobody claimed. Our report reads the series as roughly 4× the 2020 baseline, which is our own arithmetic on the endpoints, and as about 70% below the 2022 peak. Both are true of one series, which is why the series belongs on the page and neither ratio does. Two constraints: it is extracted from the report and not yet re-verified, and it stops at 2024. The all-sector corporate venture market and this deep-tech-specific series are different universes, so the AI priority share and the cheque sizes do not stack into one line.

What we are measuring next

The best objection to the pair deserves stating in full. A cohort of 2023 to 2025 foundings is small and self-selected, and young AI companies taking most of the rounds available to young companies is true of many markets in this cycle. A count of rounds says nothing about cheque size, so 70% of rounds can sit inside a modest share of the money. The figure may also be reporting founding rates: if a large share of new companies in that window are AI-core, they will close a large share of the cohort’s rounds without anyone having reallocated anything.

Our record cannot settle that last reading yet. AI sits at 11% in both the IDB’s 2023 mapping and EMERGE’s 2025 Radar, which argues against a surge in AI formation, but the two are built on different universes and taxonomies: the IDB’s 2023 edition covered 340 companies as our report records it, EMERGE’s 1,316. LADP’s own database breaks AI and cryptography out as separate verticals, which the IDB taxonomy does not, so the 11% is not a share of our base either. Two readings taken with instruments that count different things cannot separate investor selection from a wave of new companies.

That is a measurement job, and our report already prescribes the fix: adopt a common sector taxonomy at the source, and build an open, versioned map of the region’s ventures carrying sector, sub-sector, TRL, location, funding stage and investors, so that the community can inspect entries and cross-check against existing mappings. A shared taxonomy is the cheapest public good available to Latin American deep tech, and it is what we are building next.

What to do with the pair

For allocators, the move is to screen on the flow and diligence on the stock. The 61% and the 11% name the companies that exist and where the region’s scientific advantage is real; the recent-cohort round share names where committees are already moving. A pipeline built only on the sector mix oversamples accumulated formation, and one built only on the AI flow misses Establishment Labs, brain4care and everything like them.

For founders outside AI, scarcity is not a moat. Sitting inside the 61% does not by itself attract a round while the rounds in recent cohorts concentrate elsewhere. What worked for Symbiomics and Future Cow was a named lead with sector conviction and, where the technology allows, a strategic backer who needs the thing in production.

For ministries, the binding constraint sits outside deep tech. Fintech took about 58% of Latin American venture capital in the twelve months to Q2 2026, US$3.4 billion of it, and about 61% of 2025 venture funding on Dealroom’s reading, which we carry as reported; our report published 55% from Sling Hub for 2024, so fintech’s share has held or risen. The pie is growing, with Dealroom putting regional venture capital at US$4.8 billion in 2025 and a projected US$6.2 billion for 2026. Deep tech policy competes for a residual of that. Instruments built around a biotech-heavy company base alone will keep producing company counts, the one measure that held flat here while the financing moved.

The rotation we can document is happening inside deep tech rather than into it, and that is an opening rather than a warning. The scientific base has produced the most valuable deep tech venture in Latin America, and the newest cohorts are already pulling most of their rounds into AI. We publish the stock and the flow side by side, each with its source, its date and its basis, so a reader can size the opportunity. If you are allocating into the region, building inside it or writing policy for it, tell us which of the two measures is doing the most work in your decision, and we will keep sharpening the one that matters.