SEA Funding Q2 2025 Rebound: Three Reasons I'm Cautiously Reading It

Originally published: 2025-06 | Last verified: 2026-09-13 H1 2025 figures are from Tracxn’s June 2025 report, linked inline, and every percentage names its comparison base. Tracxn has since restated these periods upward — Q1 2025 to about USD 1.3B in its Q1 2026 report, H1 2025 to about USD 3.2B in its H1 2026 report — so comparisons here stay inside the June 2025 report rather than mixing versions. The exceptions — the 2023 baseline in Reason 2, from Tracxn’s December 2024 annual; the Q1 2025 figure used to derive Q2, from its March 2025 Q1 report; and the later restated quarterly prints — are flagged where they appear. Where a number is derived rather than reported, the piece says so. Quarterly funding data is volatile on individual large rounds; please refer to primary sources for current figures.

The numbers from the first half of 2025 read like a clean SEA venture capital rebound story - and the named late-stage rounds behind them are worth checking against the headline. Total tech startup funding for H1 2025 came in at approximately USD 2 billion — 7% above H1 2024 and 24% below H2 2024 — with late-stage funding up 140% over H2 2024 to ~USD 1.4 billion. Tracxn’s June 2025 report gives no standalone Q2 figure. Subtracting the USD 909 million Q1 print, which Tracxn published on 20 March before the quarter had closed, from the half-year total implies roughly USD 1.1 billion for Q2. That is arithmetic across two early cuts, not a reported number, and I flag it because it gets quoted as though it were reported.

The headline framing in the regional press has been “SEA is recovering.” It’s a tempting read. It is also, in my view, the wrong framing for what the data actually shows. The H1 2025 numbers fit much more cleanly into the funding reallocation framework I outlined in the SEA pillar piece — late-stage capital concentrating into a narrowing set of survivors with proven unit economics, against a backdrop of continued seed-stage caution — than into a “broad-based recovery” narrative.

Three reasons to be cautious about reading H1 2025 as a recovery, rather than as a deeper continuation of the reallocation story.

Reason 1: The recovery is concentrated in a small number of large rounds

The first thing the headline numbers obscure is how few rounds carried the dollars. H1 2025’s USD 2 billion did not arrive distributed across a healthy population of rounds. Tracxn counted five USD 100M+ rounds in the half, against three in H2 2024 and two in H1 2024, and the largest one it named, Digital Edge’s USD 640M Series D, was a single data-centre raise.

DealStreetAsia, which runs a different denominator, saw no year-on-year rebound. Its count had H1 2025 equity funding down 20.7% year-on-year to USD 1.85B across 229 deals, with both deal volume and value at their lowest since its report began. Within the half, though, its numbers moved the same way as Tracxn’s. Unlike Tracxn’s half-year report, DealStreetAsia’s splits the half by quarter: Q2 capital of USD 1.28B was more than double Q1’s USD 0.58B, which the ITBrief write-up of the report reads as investors “targeting larger deals with businesses that demonstrate strong fundamentals.” On the same count, the half’s ten late-stage deals raised USD 756M, a 70% increase on the previous half-year, at a median size of USD 60M. So the two houses disagree on the year-on-year direction of the dollars. They agree on the shape: late-stage money rising and concentrating in fewer, larger rounds, in the same half that overall deal volume hit its low.

What this means is that the capital was available, but it was being deployed selectively into a smaller set of companies that had survived the 2022-2024 funding winter and demonstrated durable unit economics. The companies that did not have that demonstrated economics, particularly at seed and Series A, continued to face a structurally tighter market.

This is what reallocation looks like in deal-count form. Capital concentrating in winners, not capital broadening across the ecosystem.

We turned down 4-5 times more pitches in Q2 2025 than we did in Q2 2022, but we wrote our largest individual checks of the cycle to the survivors that were unequivocally working. The aggregate dollar number is up; the breadth of the ecosystem participating is not.— Singapore-based Series B partner, regional fund

Reason 2: The late-stage surge is structural, not cyclical

The ~140% H1 2025 increase in late-stage funding versus H2 2024 looks dramatic, but 2024 as a whole was the trough — late-stage ran USD 562M in H1 2024 and USD 583M in H2 2024 on the same report — and a percentage rise off a trough flatters. The more useful question is how much of the lost ground has come back.

Tracxn’s December 2024 annual put 2023 late-stage funding at USD 4.1B — an earlier cut of Tracxn’s data, so treat the comparison as directional. 2023 is not a pre-correction baseline, though: on the same report, total SEA tech funding had already halved, from USD 14.2B in 2022 to USD 7B in 2023. Against that already-corrected year, H1 2025’s USD 1.4B of late-stage annualises to about USD 2.8B, roughly two-thirds of 2023. Measured from the 2024 trough — USD 562M in H1 2024 and USD 583M in H2 2024, about USD 1.15B for the year — that closes a little over half of the distance back to the 2023 level. (Both are my arithmetic across two report versions, not reported figures.) Against 2022, when total funding ran at twice the 2023 level, the gap is wider still. So the rebound has recovered roughly half the ground to a year that was itself a correction year, and far less of the ground to the prior peak. The right framing is that late-stage is reverting from extreme depression toward a new, lower equilibrium — not that late-stage is recovering toward the prior peak.

This matters for thesis underwriting. A “we’re back to where we were” framing implies that the broader ecosystem dynamics that drove the 2021-2022 cycle are returning. They are not. What is happening is that late-stage capital, which pulled back hard in 2024 — Tracxn’s December 2024 annual had 2024 late-stage down about 77% on 2023 — is now deploying into the post-correction survivors, at a deployment rate I expected to stabilise near H1 2025 levels for several quarters. (The next prints broke that expectation in both directions. Tracxn’s Q1 2026 report, which restates earlier quarters, has late-stage at USD 1B in Q1 2025, USD 650M in Q4 2025 and USD 2.2B in Q1 2026, most of the latter a single data-centre round — see the Q1 2026 read. Those are later, restated figures, so they are read against each other rather than against the June 2025 numbers above.)

The structural read: late-stage SEA funding is now carried by a handful of large rounds — five USD 100M+ rounds in the whole of H1 2025 on Tracxn’s count — and a single round can swing a quarter. That is a healthier position than H2 2024, but it is qualitatively different from the 2021-2022 cycle.

Reason 3: The seed and Series A stages are not participating

The third reason for caution is the most important: the H1 2025 rebound is conspicuously absent at the seed and Series A stages. Tracxn had seed funding at USD 87M, down 51% on H2 2024 and 68% on H1 2024, and early-stage funding — where Series A sits — at USD 464M, down 74% on H2 2024 and 53% on H1 2024, despite the late-stage surge.

This pattern has implications. The pipeline of companies that will reach late-stage maturity in 2027-2028 is being funded today at rates that are significantly lower than the 2021-2022 cohort. Even with the assumption that today’s seed companies have better unit economics on average than the 2021-2022 cohort, the absolute dollar volume of seed deployment limits how many late-stage opportunities the ecosystem can produce two to three years out.

What the 2025 late-stage surge will not do is create the next 2021-2022 vintage. The capital being deployed today is going into companies that were funded in 2019-2021 — the cohort that has weathered the cycle and demonstrated viability. The companies being seed-funded in 2024-2025 are a smaller cohort, with tighter selection criteria, that will produce a smaller late-stage opportunity set in 2027-2028 than the current cohort produced in 2024-2025.

The forward-looking implication for LPs is that the SEA late-stage opportunity through 2026-2028 may be relatively concentrated and competitive — fewer late-stage companies attracting more late-stage capital — but the period beyond that depends on what happens to seed and Series A funding patterns over the next 18 months.

What the cautious framework looks like

Pulling these three reasons together, the framework I’d use to read H1 2025 and forward is:

The “reallocation” reading is the framework I’d advocate. It accepts the data, doesn’t deny the late-stage activity, and reads the underlying texture honestly.

What this means for capital allocation

For LPs allocating to SEA-focused funds in 2026-2027, the practitioner takeaways are three:

Re-evaluate vintage exposure. The 2024-2025 vintage of SEA funds is being constructed in a meaningfully different market than the 2021-2022 vintage. Smaller deal counts, larger ticket sizes, more concentration into late-stage opportunities. The portfolio construction implications are substantial: more concentrated portfolios, longer holding periods, potentially better IRR profiles per company but lower portfolio diversification. LPs evaluating these vintages should expect different return distribution shapes than the 2021-2022 vintage produced.

Beware the “we’re back” thesis. Fund managers raising new SEA vintages in 2026 will be tempted to frame H1 2025 as “the rebound” and pitch portfolio construction implicitly assuming a return to broad-based ecosystem activity. The data does not support that framing. Underwriting should assume a continued narrow-and-deep capital deployment environment for the foreseeable future.

Watch the seed-and-Series-A pipeline. The leading indicator for SEA’s late-stage opportunity in 2027-2028 is the seed-and-Series-A activity in 2024-2025. If those stages remain weak through 2026, the late-stage opportunity set narrows further into 2027-2028. If they recover, the late-stage opportunity set broadens again. This is the variable I’d track most carefully over the next 12-18 months.

What might shift the cautious read

Two things that could change the framework, ordered by my probability estimate.

Continued AI-driven deployment expansion. The largest round Tracxn named for H1 2025, Digital Edge’s USD 640M, went to data-centre capacity — the infrastructure AI workloads run on. If AI deployment continues to expand the addressable opportunity set for SEA-focused investing, some of the seed/Series A weakness could reverse as new AI-native companies enter the pipeline. This is the most likely source of upside to my cautious read.

A China-adjacent capital reallocation. If a meaningful portion of the capital that has historically gone to China-focused VC redirects toward SEA — driven by mainland-China political risk perception or by China’s continuing economic deceleration — the SEA ecosystem could see a step-up in available capital that broadens beyond the current narrow concentration. This would be a multi-year shift but the directional pressure exists.

Neither of these is the base case. The base case remains: SEA H1 2025 is the reallocation story playing out at an absolute capital level slightly above H1 2024 (though below H2 2024) on Tracxn’s count, with continued narrowness in deal count and continued weakness at the early stages. Read it as such, and the SEA opportunity set looks different — and probably better-priced — than the recovery framing implies.