SEA's Funding Reallocation Story (Not the "Recovery" Narrative)

Originally published: 2023-11 | Last verified: 2026-05-06 Statistics in this article have been verified against Tracxn, DealStreetAsia, and Tech Collective reporting current as of May 2026. Funding data revises frequently; please confirm against primary source reports for the most current quarterly figures.

The narrative on Southeast Asia’s startup ecosystem heading into late 2023 is “recovery.” You see it in fund manager letters, in conference panel framings, in the Tier 1 VC marketing decks that have started circulating again after eighteen months of silence. The pitch is simple: 2022 was the peak, 2023 is the trough, and we’re calling the bottom now because [insert macro signal of choice].

I want to push back on that framing. Not because I think SEA is in trouble — I don’t — but because “recovery” is the wrong shape for what the data is actually showing. The data is showing reallocation. The capital that left the early stages of the ecosystem is not coming back; it is being concentrated in fewer, later, larger checks at the top of the stack. Calling that “recovery” obscures who wins and who loses, and it obscures the underlying structural change that an LP (Limited Partner) ought to be reading from.

This piece lays out the framework I’ve been using internally to think about that reallocation. It is built from publicly available Tracxn and DealStreetAsia data through the recent reporting cycles, anchored to the H1 2025 read which is now the cleanest cut of the post-2022 environment.

The shape of the data, briefly

Three numbers do the heavy lifting here.

Across full-year 2024, SEA tech funding came in at roughly USD 2.84B, a 59% decline year-on-year. Seed-stage funding for the year was USD 373M (down 52.4% from 2023’s $783M); early-stage was USD 1.5B (down 28.6%). (Source: Tracxn SEA Tech Annual Report 2024.)

In H1 2025, the picture inverted at the top of the stack. Late-stage funding surged 140% versus H2 2024 to USD 1.4B. Seed-stage in the same period collapsed a further 50% to USD 50.7M. Early-stage fell another 27% to USD 167M. (Source: Tracxn H1 2025 reporting; Tech Collective summary.)

For the full year 2025, DealStreetAsia recorded USD 5.37B across 461 equity deals — up modestly in dollars but with one of the lowest annual deal counts in more than six years. H2 2025 funding value of USD 3.51B came largely from a small number of outsized late-stage transactions. (Source: DealStreetAsia 2025 SEA funding report summary.)

Sit with those numbers. The dollar headline is recovering. The deal count is at a multi-year low. The composition has shifted dramatically toward the top of the stack. That is not a recovery curve; it is a barbell.

Why “reallocation” is the better frame

Three concurrent things are happening, and they are easier to read separately than together.

Reallocation 1: From breadth to depth at the top of the stack

The late-stage surge in H1 2025 is real, but it is structurally narrow. Tracxn’s read shows late-stage funding concentrated in fewer rounds with larger ticket sizes. DealStreetAsia’s late-stage deal count tells the same story from the deal-volume side: 18 late-stage equity deals across the first nine months of 2025, the first year-on-year increase in late-stage activity in four years, but still very small numbers in absolute terms.

What this means: the capital that’s flowing into SEA at the top of the stack is being directed at companies that have already cleared a particular bar — proven unit economics, demonstrated revenue scale, line of sight to either profitability or strategic exit. There is no broad-based late-stage market opening up. There is a narrow, well-developed market for a small number of late-stage companies that the major Asia-focused funds and global growth investors have already underwritten.

For an LP evaluating an SEA-focused growth fund right now, the relevant question is no longer “is the late-stage market opening.” It clearly is. The relevant question is “does this manager have access to the small number of companies that are actually cleared for that capital.” The dispersion in fund manager outcomes over the next 24 months will track that access question more than anything else.

Reallocation 2: From early-stage breadth to late-stage selection

The seed and early-stage decline is the other half of the same story. Capital that historically would have funded SEA seed cohorts at $300–500M per year is no longer present at that scale. The H1 2025 seed number ($50.7M) is not just a slowdown — it is a structural reset. Seed cohort sizes that the SEA ecosystem priced for in 2021–2022 simply do not exist as fundable populations at current capital availability.

This isn’t unique to SEA — global seed has compressed everywhere. But the SEA effect is sharper because the pre-2022 seed thesis was heavily dependent on the assumption that follow-on Series A and B capital would be reliably available. When the follow-on dried up, the seed cohort got quietly stranded. The lesson the next cohort of SEA founders is internalizing is that capital efficiency, not growth optionality, is the binding constraint at every stage.

What this means for the ecosystem shape: the talent and entrepreneurial pipeline that SEA built between 2017 and 2021 is being structurally redirected. Some founders are pivoting to bootstrapped or revenue-funded models. Some are leaving for the US or UAE. Some are stepping out of the founder track entirely and joining the regional offices of the few remaining well-capitalized companies. The “founder ecosystem” of SEA in 2027 will be smaller and more concentrated than the founder ecosystem of 2022.

Reallocation 3: From multi-country breadth to Singapore concentration

The geographic reallocation is the third moving part, and it is the one that gets the least honest treatment in the press. Singapore captured roughly 92% of SEA tech funding and 88% of fintech funding in H1 2025. That is not a small Singapore premium; that is near-total concentration. (Source: Tracxn H1 2025 reporting.)

Two readings of that number compete:

1. The hub story. Singapore is the rational booking jurisdiction for SEA-region investment, the operational base for management teams running multi-country businesses, and the regulatory home for fund vehicles. Of course the capital flows through there.

2. The concentration risk story. A regional ecosystem in which 92% of tech capital is booked through one city is structurally fragile in a way that the headline does not communicate. It also obscures the underlying question of where the capital is actually being deployed (Indonesia? Vietnam? Singapore-domestic?) versus where it is being booked.

Both readings are true. The honest version is that the 92% figure tells you almost nothing about deployment geography. It tells you about booking geography, fund domicile, and the operational reality that running a $50M+ growth round through anywhere other than Singapore is currently friction-heavy. For an LP, the relevant follow-up question on any SEA fund pitch is “what does the deployment pipeline look like by country, and how does that map to the booking jurisdiction.” The two are routinely conflated, and the conflation flatters the diversification story.

What “recovery” misses

The recovery framing implies that the prior shape was the equilibrium and we are returning to it. The reallocation framing says: the prior shape was the disequilibrium, and what we are seeing now is the new equilibrium settling in.

If you accept the second framing, several conventional reads of SEA become problematic.

The “wait for seed to come back” position. A meaningful subset of regional VCs have been telling LPs that the current seed environment is cyclical and that returns at seed will normalize once “macro stabilizes.” The data doesn’t support that as a base case for SEA specifically. The seed pullback is global, but the SEA seed pullback is sharper because the supporting follow-on capital infrastructure has structurally thinned. Without follow-on capital available to a meaningful share of the seed cohort, even successful seed picks face a structural exit problem before they reach a Series B-able scale. A manager telling you “seed is the entry point now” needs to also be telling you what their plan is for the follow-on bridge.

The “Indonesia is back” position. Indonesia has been a perennial “next year is the year” market for SEA VCs since the late 2010s. The Q1 2025 data shows Indonesia at USD 119M of funding — well below its 2021 peak — and the share of SEA capital that touches Indonesian deployment versus what gets booked through Singapore is a number that the SEA reporting infrastructure does not consistently break out. Indonesia may indeed be due for a structural inflection, but the case has to be made on demand-side fundamentals (rising urban consumer spending, payments infrastructure maturity, B2B SaaS pricing power), not on the assumption that capital will rotate back to it because it’s “underweighted.”

The “Singapore captured 92% so the regional story is fine” position. This is the one that bothers me most. Singapore capture of regional booking does not equate to regional deployment health. It also does not equate to a sustainable equilibrium — at some point, either deployment geography catches up with booking geography (which would mean meaningful capital starts flowing into Indonesian, Thai, Vietnamese deployment from Singapore-booked vehicles), or the booking advantage starts to be questioned by LPs who realize they are paying for hub costs without diversified country exposure. Both outcomes are possible. Neither is being priced honestly into the current pitch decks.

What this means for an LP today

Three implications follow from the reallocation read for an LP allocating to SEA right now.

Implication 1: Late-stage funds are the right place to look, but with discrimination

The late-stage market is the part of the SEA stack that is functioning as a market right now. Capital is being deployed, deals are being underwritten, exits (slowly) are happening. The question is which managers actually have access. The well-known regional growth funds (think: the Asia practices of the major US-listed growth investors, plus the established SEA-native growth funds that have closed Fund III or Fund IV) have the operating infrastructure and the founder relationships to compete for the small number of fundable late-stage companies. Newer or smaller managers pitching late-stage exposure into SEA need to be evaluated on access, not thesis. The thesis is roughly the same across the cohort; the access is not.

Implication 2: Seed funds need a credible follow-on story

A pure-play SEA seed fund without a clear answer to “what happens to your portfolio companies at the Series A/B bridge” is a structurally fragile bet right now. The fund may pick well at seed, but if 70% of the cohort can’t access follow-on capital before the seed runway runs out, the picks don’t matter. The seed funds that should be on the LP shortlist are the ones with explicit co-investment relationships, structured follow-on pools, or tight integration with later-stage vehicles. Pure-play seed without that bridge is a thesis that worked in 2019 and 2020 and does not work in the current environment.

Implication 3: Country-specific deployment funds may be a better risk-adjusted route than regional vehicles

This is the contrarian implication of the reallocation read. If the regional booking story is increasingly a Singapore-concentration story, then the diversification benefit of “SEA exposure” is being delivered by funds with Singapore booking and unclear country deployment exposure. A fund that is explicitly Indonesia-focused, or Vietnam-focused, with investment professionals on the ground in those markets, may offer cleaner exposure to the country economic fundamentals than a regional vehicle whose actual deployment geography is opaque. The trade-off is concentration risk versus exposure clarity, and reasonable people can disagree on which they want. But the disagreement should be conscious, not implicit.

What practitioners get wrong

Two reads I think are wrong on the current SEA narrative:

“The 2025 dollar rebound proves the market is back.” It proves that the dollar number is back; it does not prove the market is back. A market with one-fourth as many deals at three times the average size is not the same market as one with the prior deal count at lower average size. Median-deal-size analysis tells you whether the market is functioning broadly or whether a small number of large transactions are doing all the work. In current SEA, the answer is overwhelmingly the latter, and that has implications for portfolio construction that the dollar headline obscures.

“The seed pullback is just resetting valuations.” That’s part of it, but it’s not the structural story. The structural story is that the supply of follow-on capital has compressed, which means the option value of any seed pick has compressed regardless of valuation. A correctly-priced seed bet still doesn’t work if the path to Series A is unavailable. The pricing reset is real and necessary, but it does not by itself fix the funnel.

How this maps onto fund construction

If you’re inside a fund formation conversation right now — either as a GP raising or as an LP being raised — the reallocation framework is also a check against the Fund III/IV positioning that’s quietly proliferating across SEA managers.

Every SEA-region growth fund that closed Fund I or Fund II between 2018 and 2022 is now back in market for the next vintage. The pitch decks I’ve seen over the last six months tend to converge on a few common moves: rotation toward “AI-native enterprise” or “fintech infrastructure” thesis lift, a re-baselined fund size that is materially smaller than the prior vintage (often 40–60% of the previous fund), and a pivot from breadth-first portfolio construction toward concentration-first.

Two of those moves are sensible. The fund-size right-sizing is honest engagement with the deployment environment — a USD 300M growth fund deploying into a SEA market that produces 18 late-stage deals across nine months of 2025 is structurally over-funded for the addressable opportunity. Cutting to USD 150M makes the math work. Similarly, concentration-first construction (six to eight portfolio companies rather than fifteen to twenty) maps cleanly onto the late-stage thesis that fewer fundable companies means higher conviction per check.

The thesis lift is where I’d push back. “AI-native” or “fintech infrastructure” framing is the marketing layer, not the deployment layer. Underneath, the question is whether the manager has access to the small number of companies in those verticals that are actually fundable at growth stage in SEA, and whether the manager’s diligence framework is calibrated to operate companies in the SEA regulatory and customer-acquisition environment rather than ports the global thesis. Some managers have that calibration; many do not. The thesis lift is easy to write into a deck; the operating capacity behind it is harder to build.

For an LP, the diligence question on these next-vintage funds is less about thesis and more about: (a) does the manager have an honest read on the deployment environment that matches the data above, (b) is the right-sized fund actually right-sized for the deployable opportunity, and (c) does the team have the operating muscle to execute concentration-first construction without fund-of-fund-style dilution back to a broad portfolio. Two out of three is a real fund. One out of three is a marketing exercise.

The narrative tax

There is a soft cost to the “recovery” framing that gets less attention than it should. Recovery narratives let LPs and managers avoid having to underwrite what the reallocated equilibrium actually looks like. If 2026 is just “the bottom + recovery,” then portfolio construction stays anchored to the 2019–2021 mental model — diversified by stage, broad by country, balanced between seed and growth. If 2026 is the new equilibrium with a barbell shape and Singapore-heavy concentration, portfolio construction needs to be different.

I’m not arguing the new equilibrium is better or worse than the old one. Different markets, different playbooks. What I’m arguing is that calling it “recovery” lets the conversation skip the part where you actually decide what to do differently. The reallocation framing forces that conversation, which is why I prefer it even though it sounds less optimistic.

Worth tracking next

Four signals I’d watch over the next four to six quarters:

• Series A/B deal count trajectory. The seed-to-Series-A bridge is the part of the funnel where the reallocation pain is most visible. If Series A deal count starts ticking up alongside late-stage, that’s a real signal that the broader market is healing. If it stays flat or declines while late-stage continues to surge, the barbell is structural and not transitional.
• Singapore deployment geography breakdown. If DealStreetAsia, Tracxn, or one of the regional VCs starts publishing the deployment-geography decomposition of Singapore-booked capital, that’s a meaningful disclosure event. Right now the absence of that breakdown is a feature for managers and a bug for LPs.
• Indonesia mid-market activity. Indonesia funding through Singapore vehicles is the perennial “is this finally the year” question. The data point I’d watch is mid-market activity at Series B and growth stage in Indonesia-deployed capital — if that picks up, it’s a real Indonesia rotation. Top-line “Indonesia funding rose X%” headlines tell you almost nothing without the stage breakdown.
• Fund formation slowdown. SEA VC fundraising hit a 7-year low in H1 2025 by DealStreetAsia’s count. If the LP commitments to new SEA funds remain compressed through 2026, the late-stage capital available for the small number of deployable companies will eventually thin. The reallocation pattern in the deployment data is downstream of the fundraising data, with about an 18-month lag.

The recovery story will keep getting written because recovery stories are what fund managers pitch at the bottom of every cycle. The reallocation story is the one that I think actually fits the data, and it’s the one that an LP underwriting SEA exposure now needs to be operating from. That’s the read I’m having.