The Thesis · Altazor Ventures Desk

Summary

Latin American venture is a USD 4–5 billion a year market that briefly became a USD 16 billion market in 2021 and has spent four years finding its floor. The floor appears to be in: 2025 was the first year of growth since the peak, and the first half of 2026 ran roughly a third ahead of the same period a year earlier. But the recovery is narrow. It is concentrated in a handful of late-stage fintech rounds, in Mexico more than Brazil for the first time in over a decade, and in a small group of managers who can still raise. Seed and Series A activity is at multi-year lows. That combination — capital returning at the top, drying up at the bottom — is what makes the next two years interesting for an investor who can underwrite early-stage companies with a view on who will fund them next.

1. Where the capital is

Year	VC invested (USD bn)	Deals	Series
2019	5.0	479	LAVCA
2020	4.3	548	LAVCA
2021	16.0	971	LAVCA
2022	8.0	1,199	LAVCA
2023	4.2	826	LAVCA
2024	4.5	751	LAVCA
2025	4.1	681	Crunchbase
H1 2026	2.4	—	Crunchbase

2019–2024 from LAVCA [1]; 2025 and H1 2026 from Crunchbase-based series [4][5][6][7], which run roughly 10–20% below LAVCA's methodology. LAVCA's own 2025 total is not public; its summary implies about USD 4.2bn, with early-stage at USD 2.2bn or 52% of the total [3]. The two series agree on direction: 2025 grew low-double-digits over 2024, and H1 2026 (about USD 2.4bn) grew roughly 30% over H1 2025 (USD 1.8bn) [2][6][7].

Three structural features follow from that table.

The recovery is late-stage. In 2025, late and growth rounds were USD 1.6bn and early-stage about USD 2.0bn, while seed and angel fell 22% to USD 540m [4]. In Q1 2026 late-stage took 74% of dollars; in Q2 2026 late and growth reached USD 991m, up 84% year on year, while the number of seed and early rounds fell both sequentially and year on year [6][7]. Five companies — Plata, ADDI, Klar, Omie and Kavak — took about a quarter of all 2025 venture dollars [3]; the top ten deals took about 30% [5]. This is a market of larger cheques and fewer deals.

Brazil and Mexico dominate, and Mexico is now ahead. In 2024 Brazil took 44.6% of dollars and Mexico 26.0%, with Colombia at 11.4%, Argentina 8.8%, Chile 3.9% and Peru under 1% [1]. In 2025 Brazil remained first at roughly USD 2.0bn and Mexico second at about USD 1.0bn, up 53% [4][5]. In 2026 the order flipped: Mexico raised USD 404m in Q1 against Brazil's 240m, and USD 944m in Q2 against 350m — the first time Mexico has led the region in a quarter since 2012, bar one [6][7]. The caveat matters: strip out Clip's USD 500m and Plata's USD 405m and Mexico's Q2 was small. A regional thesis that is not explicit about country allocation, and about the difference between a country's headline number and its deal flow, is not a thesis.

Unicorn creation has slowed to a trickle, and the exit window has only just cracked open. The region minted one new unicorn in 2024 (QI Tech), two in 2025 (Plata and Kapital, both Mexican fintechs) and one so far in 2026 (Enter, Brazilian legal AI, at USD 1.2bn) [8][9][10]. The IPO route was effectively shut from 2022 until January 2026, when PicPay listed on Nasdaq at about USD 2.5bn, followed by Agibank on the NYSE in February — both fintech, both listed in New York, not São Paulo [11][12]. The realistic exit for most of the current cohort remains strategic M&A: Visa's USD 1bn purchase of Pismo (2024) and Visma's roughly USD 300m purchase of Conta Azul (2025) are the templates [13][14]. LAVCA counts around 60 Latin American tech companies that have raised more than USD 150m and have yet to exit [1].

2. What the 2021 cycle taught

The 2021 cohort rewarded distribution and growth at any cost: neobanks, delivery, quick-commerce, used cars, B2C marketplaces. Many of those companies were built for a cost of capital that no longer exists. Kavak, valued at USD 8.7bn in 2021, raised at USD 2.2bn in 2025 before a USD 300m Series F led by a16z in early 2026 on the back of its first profitable month [15]. SoftBank, which deployed USD 8bn across two Latin America funds, has made two new deals in the region in two years; Latin America is now about 2% of its equity holdings, down from 5% [16].

The lessons that survive: unit economics denominated in local currency need to be tested against a devaluation scenario, not a base case; regulatory arbitrage in one country is not a moat in another; and a cap table that has already absorbed several rounds at 2021 prices is a liability, not a signal. The companies now raising the largest rounds — Plata at a USD 5bn valuation with Bicycle Capital [23], Kavak with a16z, Ualá with Allianz X — are the ones that got to profitability or close to it.

3. Who is still writing cheques

The manager base is consolidating. LAVCA's private-capital data show the share of fundraising going to experienced managers (sixth fund or later) rising from 51% in 2020 to 80% in 2024 and 88% in the first half of 2025, with first-time managers at about 4% [2]. The caveat: that series covers all private capital, not venture alone, and the same deck notes that 174 venture managers raised their most recent fund between 2021 and mid-2025, so the venture manager base is young rather than shrinking. What is clear from the fund-level evidence is that even established names are raising less: Canary closed USD 150m for Fund IV in February 2026 against a Brazilian funding drought [17]; Monashees is reportedly targeting about half its previous fund; Kaszek, with close to USD 1bn of dry powder from its USD 975m 2023 raise, is preparing a Fund VII of around USD 540m [18].

Among global funds, the picture is bifurcated. a16z, Sequoia, Founders Fund, Insight and General Atlantic have led or co-led large Latin American rounds in 2026; SoftBank and Tiger Global have not [6][7][16]. The global money that stayed is concentrated in late-stage fintech and, increasingly, AI. The gap it leaves is at seed and Series A, where local funds with smaller new vehicles are the marginal cheque.

4. Why European capital, why now

Three arguments. Nearshoring is real and is being financed. Mexico's foreign direct investment reached a record USD 23.6bn in the first quarter of 2026, up 10% year on year, and new investment (as opposed to reinvested earnings) tripled in 2025 [22]. The operating software for that industrial shift — logistics, trade finance, cross-border payments, compliance — is being written now, and the buyers are companies European investors already know.

Latin American founders are building from Europe. About 30% of the roughly 300 startups Madrid's city incubators take on each year have Latin American founders, and the city runs two landing programmes exclusively for them [20]. Spain has 4.2 million Latin American immigrants and a scale-up landing programme run with ICEX, IDB Lab and Endeavor now in its fifth edition [21]. A growing share of regional founders operate with a Delaware entity, a team in São Paulo or Bogotá and a desk in Madrid or Lisbon, which puts them within reach of European funds that have never flown to the region.

European capital has started to move, but thinly. Mundi Ventures reached a first close on a USD 100m Latin America fund in April 2026 with IDB Invest and Cofides as anchors [19]; Allianz X led Ualá's USD 195m round in Q1 2026 [6]. Against European venture funding of about EUR 73bn in 2025, Latin America's EUR 3.5bn is a rounding error [21] — which is the point. Entry valuations after four years of reset sit well below what comparable companies command in Europe, and the competition for early-stage deals from global funds is the lowest it has been in a decade.

The gap is not capital. It is a map: who the local funds are, what a good cap table looks like, which regulators matter, and which exits have actually happened. That is what this desk is for.

5. What makes sense now

The categories worth underwriting share a profile: they sell into the infrastructure or the balance sheet of large, slow institutions; they earn in dollars or in dollar-linked revenue; and they have a credible strategic acquirer set inside the region or among the global players that already own Latin American assets. Fintech infrastructure took 61% of 2025's dollars on 29% of its deals [5], and the Pismo and Conta Azul exits show software acquirers paying software multiples. The first memos on this desk follow that profile — cross-border payments and FX infrastructure, housing finance and proptech, and industrial and energy-transition software in the Andean economies. Each starts from a market the author has worked in.

6. What I look for

Revenue that survives a 30% devaluation of the local currency, either because it is dollar-linked or because the cost base devalues with it.
A wedge into a regulated or infrastructure-level workflow that is expensive to replace once embedded.
Founders who have already operated in the region, or a team that pairs a regional operator with a builder.
A named list of realistic strategic acquirers, and evidence that at least one has bought something similar.
A cap table where the founders still own enough to matter after the next two rounds — and a view on which fund writes the next cheque, because at seed and Series A that fund may not exist yet.

Sources

Where LAVCA and Crunchbase-based figures differ, the series used is named in the text. Figures from press reports of official data (Mexican FDI) are cited to the press source. LAVCA's fund-level fundraising data is members-only and is not reproduced here.

Changelog

Date	Change
6 Sep 2026	v1.0: sourced figures for funding, country split, stage mix, unicorns, exits, fund closes and the Europe–LatAm bridge. Section "Who is still writing cheques" added.
6 Sep 2026	v0.1: structure and qualitative view.