Anchor Public R&D
Public R&D budgets on an infrastructure footing
LAC spends 0.57% of GDP on R&D against a 1.92% global average, in short annual cycles. Put strategic deep tech on infrastructure footing and turn the R&D budget line into multi-year co-financing.
The rationale
Public money carries the majority of R&D financing in the region, and it arrives on the wrong clock. LAC R&D spending sits at 0.57% of GDP against 1.92% globally, the state underwrites roughly 60% of it, and it is disbursed through annual grant calls that expire before a hardware or biology program reaches a technical milestone. In Brazil, 36% of deep techs rely solely on public funds and 47% received no investment at all, so the grant serves as the entire balance sheet until the company runs out of runway, never as a bridge to private capital.
Three instruments already running in the region do the work once they are scaled and lengthened. Matching funds co-invest public capital alongside private money, typically covering 40 to 60% of eligible costs; Chile’s CORFO reimbursable grants cover 40 to 65% of private R&D cost, and Uruguay’s Innovation Hub has run a US$10m 1:1 program since May 2024. Upgraded Special Economic Zones trade fiscal perks for clocked permit timelines. A multi-year commitment horizon written into statute is what turns any of them from a subsidy into an anchor investment.
The standing objection is that governments cannot pick winners and should not try. Israel’s 1993 Yozma program answers it structurally: the state supplied capital and rules, private managers selected the companies. The fiscal objection is harder and deserves a direct answer, which is why the first move here re-times appropriations that already exist rather than asking for new ones. The composition gap is the target. In advanced economies business supplies more than 60% of R&D funding; in LATAM the state carries roughly 60% and the private share is printed at 35% in one chapter of the report and 43% in another, which is itself the argument for publishing it as a range and for lengthening the public commitment until private money has something to follow.
The evidence
What it unlocks
Public R&D money committed on the timescale hardware and biology actually take, with one matching-fund term sheet a venture can file across three countries.
It has been done before
Israel, the Yozma program (1993)
The state placed US$100m into ten venture funds structured 40% public and 60% private, with private managers selecting the companies. The funds Yozma seeded now manage more than US$10bn and are credited with more than US$80bn in value creation. It followed the 1985 R&D Law, which reimbursed up to 50% of corporate R&D and lifted Israeli private R&D spending to 4% of GDP. Figures are as published by the IDB in 2023, cumulative to that date, and are not independently re-verified.
First moves
- 0–6 months
Re-time the existing R&D line
Convert single-year grant calls in named deep tech verticals into three-year committed tranches with milestone gates, using appropriations already voted rather than new money. Deliverable: one amended call per agency, published with its tranche schedule and gate criteria, before the next budget cycle closes.
National innovation agencies (FINEP and BNDES in Brazil, CORFO in Chile, ANII and the Uruguay Innovation Hub) with finance ministries - 6–18 months
Standardise the matching-fund term sheet
Publish one regional matching-fund template with a single stated coverage band per country, common eligible-cost definitions and a 90-day decision clock, so a venture operating in three countries files once. Deliverable: a template adopted by at least three agencies, each stating its own coverage percentage rather than a regional average.
IDB Lab, national innovation agencies and the LADP coalition secretariat - 18–36 months
Legislate the infrastructure footing
Write named strategic deep tech fields into the same multi-year capital-planning statute that already covers ports, grids and telecoms, paired with SEZ authorities holding statutory service-level deadlines for permits. Deliverable: at least one country with a deep tech line ring-fenced across an election cycle and a published permit clock per approval step.
National congresses, planning ministries and SEZ authorities
How we would know it worked
- Share of national deep tech grant funding committed on multi-year tranches rather than single-year calls, reported annually by each innovation agency, rising above 50%
- Median days from matching-fund application to decision, published quarterly by CORFO, FINEP and ANII, falling below 90
- Private deep tech investment per country set against public disbursement, tracked in the EMERGE Radar and the LADP registry, with Brazil’s US$216m private 2024 figure rising toward its share of regional companies
- LAC R&D expenditure as a share of GDP in the UNESCO UIS release, moving off 0.57% toward 1%
What could go wrong
Capture by incumbents
Matching funds reward applicants who already hold private capital to match, which in practice means established firms and well-connected labs, and the ventures the instrument exists for stay unfunded. Containment: publish applicant and awardee lists with founding year and prior funding raised, cap the share of any call going to firms above a stated revenue threshold, and reserve a tranche for first-time applicants.
Election-cycle reversal
A multi-year commitment created by decree survives until the next administration, which is shorter than a deep tech development cycle, and a cancelled tranche kills the company rather than the policy. Containment: put the commitment in statute with an explicit sunset and renewal review rather than in a decree, and structure disbursement as contractual tranches to the venture so cancellation carries a cost to the treasury.
Who acts
Rests on
14 Research & Visibility An open data commons and a funding catalogue
08 Regulatory Portability 33 countries, one optional 34th regime
The data behind it
Business covers about 35% of R&D in Latin America, against 60% or more across the United States and Europe and nearly 80% in China. Thin private research funding is a structural brake on deep tech.
As published in the report, September 2025.
SourcesECLAC / CEPAL, Ciencia, tecnología e innovación: cooperación, integración y desafíos regionales (LC/TS.2022/156); UNESCO Institute for Statistics (UIS), 2026 R&D Data Release
Methodology: ECLAC Graph I.3, 2019 data (ledger #75/#97). us-eu government = 25 is the midpoint of the report's "20–30" range. Superseded framing: UIS 2026 gives LAC R&D at 0.57% of GDP vs 1.92% global, a stronger level comparison than this composition split.
| Category | Current |
|---|---|
| Latin America | 35% |
| United States & Europe | 60% |
| China | 80% |