SEA's Q1 2026 $2.8B Surge: Reading the 146% QoQ Spike Through a Reallocation Lens

Originally published: 2026-01 | Last verified: 2026-09-13 Figures are linked inline to the report each comes from. The quarter’s dollar figures and stage splits are Tracxn’s (rounded as Tracxn reports them); the deal count is DealStreetAsia’s. The two houses publish different SEA aggregations, and this article keeps them apart rather than blending them. A dated September 2026 update in the LP section closes the Q2 test the piece originally set; where the half-year data overtook the April seed run-rate in Implication 1, the text says so. Funding data revises across reporting cycles; please confirm against primary Tracxn or DealStreetAsia source reports for current figures.

The Tracxn Q1 2026 numbers landed in mid-April and the Asia tech press took the headline first: SEA tech funding at USD 2.8B, a 146% quarter-on-quarter jump from Q4 2025’s USD 1.1B and a 110% year-on-year jump from Q1 2025’s USD 1.3B. Those are the right numbers, and they got pulled into the predictable narrative: SEA is back, the trough is behind us, late-stage capital has resumed flowing.

I want to push back on that read — but not in the way the contrarian commentary has tried to. The contrarian read has been “it’s all DayOne” — meaning the USD 2B Series C raise by DayOne, the data centre operator, single-handedly carried the quarter and the underlying ecosystem is unchanged. That read is technically defensible but analytically lazy. It treats the DayOne round as noise to be subtracted, rather than as evidence about what kind of capital is willing to deploy at scale into SEA right now.

The more useful read is structural. The Q1 2026 numbers are not a recovery quarter and they are not a single-deal artifact. They are a reallocation acceleration quarter — the same reallocation pattern I laid out in my 2023 piece on SEA funding, now visibly compressed to its limit case. Capital is concentrating at the top of the stack, the bottom of the stack is not refilling, and the deal count is at historic lows. Reading the quarter through any other lens distorts the actual picture.

The numbers that matter

Pulling the relevant figures from Tracxn’s Q1 2026 report, via TechNode Global (15 April 2026):

• Total funding USD 2.8B, up 146% on Q4 2025’s USD 1.1B and 110% on Q1 2025’s USD 1.3B.
• Late-stage funding USD 2.2B, a 243% jump versus Q4 2025’s USD 650M.
• Seed-stage funding USD 105M, down 30% versus Q4 2025’s USD 149M — but up 39% on Q1 2025’s USD 75.3M.
• Early-stage funding USD 487M, a 40% rise versus Q4 2025’s USD 349M.
• Five USD 100M+ rounds. Tracxn names four of them: DayOne’s USD 2B Series C, EPG’s two Series B rounds totalling USD 200M, and Amity Solutions’ USD 100M Series D.

DealStreetAsia counts the same quarter separately, and counts deals: 98 equity deals, the lowest quarterly count in at least eight years — DealStreetAsia’s figure, as compiled by Second Talent. The two houses run different denominators, so I keep them labelled separately throughout and never divide one house’s dollars by the other’s deal count.

The deal-count number is the one that the headline reads underweight. 98 deals across an entire quarter, in an ecosystem the size of SEA, is not a recovery signature. On DealStreetAsia’s own count, full-year 2025 produced 461 deals — Second Talent’s compilation gives it as “$5.37bn across 461 deals” — about 115 a quarter, in a year Second Talent’s compilation calls “among the lowest annual totals in more than six years.” 98 is below even that quarterly average. That number is screaming “fewer companies are getting funded at all,” and the funding total is masking it.

The reallocation pattern, now in extremis

In my late-2023 piece on SEA reallocation, the framework was: capital is moving from the bottom of the stack (seed / early) to the top of the stack (growth / late), and the “recovery” framing obscures the seed-stage collapse underneath. That pattern has continued and accelerated through 2024, 2025, and now Q1 2026.

Look at the percentage of total quarterly funding that landed in late-stage:

A ~78% late-stage share is not a recovery print. It is a top-of-stack concentration print. The capital that came back into SEA in Q1 did not return to the ecosystem broadly — it returned to a small number of growth and late-stage rounds where the underlying companies have already proven the unit economics. Tracxn names four of the quarter’s five USD 100M+ rounds — DayOne, EPG’s two Series B rounds and Amity — and those four alone come to at least USD 2.3B. Take out the fifth as well and less than USD 500M of the quarter’s USD 2.8B is left for every other round in the region.

The reallocation thesis from 2023 said: late-stage will recover, seed will not. That was directionally right. The Q1 2026 print extends it: late-stage has not just recovered, it has been pulled forward dramatically — and seed has not come back. USD 105M for the quarter is down 30% on Q4 2025; it is up 39% on Q1 2025, but only from USD 75.3M, a very weak base. For the entire SEA region, that is the marker of an ecosystem bumping along a low floor in the first quarter, not one visibly producing a larger next layer of fundable companies. (The second quarter moved seed off that floor in dollar terms; see the September update below.)

What “reallocation acceleration” actually means

Reading the quarter as reallocation acceleration rather than recovery has three concrete implications.

Implication 1: The 2026 vintage may be smaller in number

The April version of this argument ran on the Q1 run-rate: annualized, Q1’s USD 105M of seed puts full-year 2026 around USD 420M — roughly level with the already-compressed 2024 figure of about USD 446M, the sum of the two 2024 halves in Tracxn’s June 2025 report. On that quarter alone, seed looked like it had found a floor. The half-year data has since overtaken that run-rate. Tracxn’s H1 2026 report puts seed at USD 328M for the half (details in the September update below), a pace of roughly USD 650M a year, above 2024 rather than level with it. So the dollar case for a structurally smaller 2026 vintage no longer holds. The count case still might: the same report has total funding rounds across all stages down to 127 from 153 a year earlier, and Tracxn gives no seed round count, so larger seed cheques going to fewer companies is possible but unconfirmed. If that is what is happening, the 2026 vintage is smaller in number even as seed dollars rise, and by the time these companies reach Series B in 2028-2029 the pool of late-stage candidates will be thinner. If it is not, the seed layer is recovering faster than this piece assumed in April.

If the count read holds, LPs (Limited Partners) underwriting SEA-focused funds with 2025-2027 vintage exposure should adjust portfolio construction expectations accordingly. The “you’ll find more graduates than you can fund” assumption that made SEA growth-stage attractive in 2018-2021 may be inverting. By 2028, the constraint on SEA growth-stage deployment may not be capital availability — it may be the supply of fundable graduates.

Implication 2: The mid-stage layer is lumpy, not broadening

The interesting space in the Q1 data is what happened at early-stage. Early funding rose 40% quarter-on-quarter to USD 487M, and Tracxn puts it 111% above Q1 2025’s USD 231M — a real number, but Tracxn does not break it down by round, and a quarterly total at this level can be moved by a handful of deals. The quarter’s own figures suggest how few. Tracxn’s four named USD 100M+ rounds come to at least USD 2.3B, more than the entire USD 2.2B late-stage line, so they cannot all be late-stage. On Tracxn’s own stage definitions, which put Series A and B rounds in early-stage and Series C onward in late-stage, EPG’s two Series B rounds, USD 200M together, would sit in the early bucket — roughly two-fifths of the quarter’s early-stage total, and most of its USD 256M year-on-year increase. Tracxn does not say which bucket each named round landed in, so treat that as an inference from its definitions and totals rather than a reported split. DealStreetAsia’s monthly barometers show how lumpy the Series A and B layer was, though they understate January: EPG’s roughly USD 100M Series B, announced on 27 January, does not show up with an amount in the January barometer, whose top five deals stop at USD 11M and whose one Series B is unsized: five Series A deals totalling about USD 40.4M in January, plus a single Series B deal the barometer does not size, three Series A rounds totalling USD 21.9M and one USD 12.5M Series B in February, and USD 108M of Series A and USD 165.5M of Series B in March. March’s USD 273.5M was several times the Series A and B amounts the January and February barometers put numbers on, but with EPG’s January round sized in, the layer looks carried by a handful of rounds (EPG’s two, plus one large Series A in March) rather than by a single month; and the DealStreetAsia barometers give no prior-year A and B baseline, so they cannot show whether that layer is broadening year on year. Tracxn’s year-on-year figure says the stage total doubled; its own totals suggest much of that may be one company’s two rounds. My read, which the quarter’s numbers neither confirm nor refute, is that local seed funds have been retrenching their early-stage allocations, leaving a gap between the seed-stage exit and the growth-stage entry.

That gap is where most of the 2024-2025 fund-of-fund LP commentary has focused; Q1 2026, with a handful of rounds carrying the layer, does not settle it either way. Companies that successfully raised seed in 2023 are now reaching the natural Series A timing window with, on my read, fewer funding sources than they would have had in 2020. On my read, the companies that secure A rounds tend to do so at higher dilution and on tougher terms. The companies that don’t secure A rounds get acqui-hired or wind down quietly. Either outcome reduces the pipeline that feeds the late-stage flow that is now driving the headline numbers.

Implication 3: Singapore’s capture share is now structural

Singapore captured 93% of Q1 2026 funding on Tracxn’s count, in a series that has held in a narrow band since mid-2025: 92% in H1 2025, 91% in H2 2025 and 94% in H1 2026. A quarter in which total funding more than doubled on the year barely moved it — which is the point. It is no longer a statistical anomaly; it is a structural feature of how SEA capital is now being booked. (It has not always been this high: Tracxn’s December 2024 annual put Singapore at nearly 67% of 2024 funding.) I’ve worked through whether that capture rate is sustainable or a statistical artifact in detail.

The reasons are mechanical rather than ideological. Look at who led the quarter’s named large rounds: Coatue led DayOne’s Series C, with Indonesia’s sovereign fund INA alongside; EPG’s March release lists Decarbonization Partners, the BlackRock-Temasek joint venture, as leading the round alongside Alibaba Cloud and other strategic investors; EDBI, the investment arm of SG Growth Capital, led Amity’s Series D alongside Asia Partners and SMDV. International and state-linked capital of this kind tends to invest through SG-domiciled holding companies, because the SG legal infrastructure, fund vehicle availability (VCC, in particular), and tax treaty network make that the dominant rational structure. The portfolio company can be operating anywhere in SEA, but where the equity round closes through an SG holding entity, Tracxn books it to SG. DayOne and EPG are both Singapore-headquartered. Amity is the booking exception: Thailand-founded — though the same announcement says more than 75% of its 2025 EBITDA came from European business units — and Bangkok’s 4% of the quarter is roughly the size of its USD 100M round.

This is not Singapore “stealing” deal flow from Indonesia. It is the consequence of capital structure decisions taken at the LP level, made before the deal was even sourced. As long as the late-stage layer is dominated by international and state-linked capital using SG-domiciled vehicles, the SG capture share is likely to stay high. It has sat in the low nineties on Tracxn’s count since H1 2025 — but “floor” would be too strong a word for a ratio that Tracxn’s 2024 annual put nearer two-thirds. I’d write a separate piece on whether this is sustainable politically; for the purpose of reading the Q1 2026 print, treat it as a fixed feature.

What this means for the LP read on 2026

Three positions I set out in April. The second was a test, and the second-quarter data has since come in.

First, do not extrapolate the Q1 print into a “SEA recovery” thesis. Tracxn’s dollar count produced the USD 2.8B headline; DealStreetAsia’s deal count produced the 98. They come from different houses, but they point the same way: a genuine ecosystem recovery would show dollars and deal counts rising together, and only the dollars are. Treat Q1 2026 as evidence of capital concentration discipline, not as evidence of broad-based ecosystem health.

Second, watch the Q2 deal count rather than the Q2 funding total. If Q2 2026 prints another <120 deal quarter on DealStreetAsia’s count, the reallocation acceleration thesis hardens into a longer-term structural pattern. If deal count rebounds to >150 while funding stays flat, the seed-stage compression may be bottoming. The funding total is the noisier signal in the current environment.

Update, September 2026: the deal-count test came in on the first branch; the seed read did not hold as cleanly. DealStreetAsia’s monthly reviews put Q2 2026 at about 105 equity deals — 33 in April and 31 in May, counts that sit behind DealStreetAsia’s paywall and could not be re-checked for this update, then 41 in June, which is on a freely readable page. On those counts, a second quarter well under 120. Funding did not stay flat. Tracxn’s H1 2026 report put the half at USD 7.4B; less Q1’s USD 2.8B from the April report, that implies roughly USD 4.6B for Q2 (derived across two report versions, not reported). DayOne’s second Series C close, USD 2.5B in June, is more than half of that implied Q2 total on its own. The rest of the quarter was not flat either: net of DayOne, roughly USD 2.1B against about USD 0.8B in Q1 (both derived). But DealStreetAsia’s separate count shows how concentrated June was: five megadeals made up 93% of the month’s USD 4.22B. Tracxn counted 127 funding rounds in H1 2026 against 153 a year earlier, and The Independent’s write-up of the report describes “a growing concentration of capital in a smaller number of large investment rounds, particularly those linked to infrastructure projects.”

Seed is the part that complicates the April read. The same report puts H1 2026 seed at USD 328M. Less Q1’s USD 105M, that implies roughly USD 220M of seed in Q2, about twice the first quarter (again derived across two report versions, not reported), and a full-year pace near USD 650M, above the roughly USD 446M of 2024. So seed dollars rose while total rounds fell year on year and the deal count stayed near its lows. That fits the reallocation pattern holding — capital going into fewer, larger rounds — but it is not the same as seed compression continuing. One possibility is larger seed cheques going to fewer companies; another is that the seed layer began to recover in Q2. Tracxn’s release gives no seed round count, so the data cannot yet tell those apart. What it does rule out is reading seed as stuck at the floor the Q1 run-rate suggested.

Third, the SG capture rate is not a number to take at face value when constructing a country-level SEA exposure. The deal-source geography matters more than the booking geography for understanding actual market dynamics. Indonesian, Vietnamese, and Filipino startup ecosystems are not as marginalized as the SG-share number suggests; they are increasingly under-represented in the booking statistics rather than in the underlying business activity. An LP underwriting a country-specific SEA fund should look at the deal-source numbers, not the capture-share headline.

The 2026 SEA ecosystem is doing fewer things, with more capital, through a smaller geographic footprint of booking entities. That is a different ecosystem from the 2021 one. Whether it is a healthier ecosystem is a much harder question than the Q1 headline suggests.