Crunchbase’s numbers for the first half of 2026 are the kind that are supposed to make an entire industry feel good. Global venture funding hit US$510 billion, beating the whole of 2025 in six months and smashing the previous half-year record of US$375 billion set in the second half of 2021. Startup capital, by the headline, is back.
Read past the headline and the story curdles. OpenAI and Anthropic alone accounted for US$217 billion of that total — 43 per cent of every venture dollar deployed on the planet in six months. Anthropic’s US$65-billion second-quarter raise by itself was close to a third of all global venture funding for the quarter.
Also Read: Anatomy of a shakeout: what 7K+ deadpooled startups reveal about Southeast Asia’s new tech reality
AI’s share of capital jumped from under 50 per cent a year earlier to more than 70 per cent by Q2. Deal count, meanwhile, barely moved. This was not a broader boom lifting more founders. It was two companies, and a handful of AI-infrastructure bets around them, absorbing a record-breaking pool of capital that the rest of the world’s startups mostly watched pass by.
What Southeast Asia actually got
Set that number against what Southeast Asia raised over the same stretch and the gap stops being abstract. Startups in the region pulled in US$2.81 billion across 98 equity deals in the first quarter of 2026, the lowest quarterly deal count in at least eight years, according to DealStreetAsia. More than 70 per cent of that quarter’s value came from a single transaction: Singapore-based data centre operator DayOne’s US$2-billion Series C. Strip that one round out and the region raised roughly US$800 million in three months.
For the whole of 2025, Southeast Asia’s tally was US$5.37 billion across 461 deals. OpenAI and Anthropic’s combined first-half haul is more than 40 times that entire annual figure, extracted from global markets in half the time.
We wrote e27‘s own retrospective on this drought a few days ago: 7,538 Southeast Asian tech startups have deadpooled since January 2020, with peak closures in 2021 and 2022 and attrition still running through 2025. The instinct in the region has been to read that as a home-grown correction, a hangover from pandemic-era excess, high interest rates, and investors demanding a path to profitability that many grocery-delivery and social-commerce plays never had.
All of that is true, but none of it is the whole story. The bigger truth is that the capital pool available to everyone who is not building a frontier model has been quietly shrinking as a share of the total, even as the total itself hits records.
Concentration is not a US problem you can watch from a distance
It is tempting for Southeast Asian founders and investors to treat this as someone else’s bubble — a Bay Area story about two labs, a handful of hyperscalers, and a debt-financed data centre build-out that has already drawn warnings from the IMF and the Bank of England about opaque leverage. But the region is not a bystander. GIC and Temasek-linked vehicles are direct participants in the AI mega-rounds reshaping the market — Temasek-backed Xora led a US$53-million seed round in Hang Ten earlier this week, and sovereign capital from the region sits inside several of the infrastructure deals now competing for the same limited pool of late-stage dollars that used to flow more evenly across sectors.
When four transactions can account for roughly two-thirds of a quarter’s global venture dollars, as insights4vc’s analysis of Q2 2026 found, every LP with exposure to venture as an asset class is making a portfolio decision about how much of that concentration it wants, whether it says so explicitly or not.
Also Read: Analysis: SEA’s June funding spike masks a narrow recovery in VC funding
Southeast Asia’s own funding data is starting to rhyme with the global pattern, just at a smaller scale. Five megadeals accounted for 93 per cent of June 2026’s US$4.22-billion regional total, a four-year high built almost entirely on DayOne, Supabase, AI startup Acrab, Airwallex, and Vietnam’s Vinpearl.
The region’s headline funding numbers are increasingly a story about a handful of outsized rounds, mostly in data centres, payments infrastructure, and AI, rather than broad-based conviction in the next generation of SEA founders. Concentration, in other words, is not just something happening to the region from outside. It is becoming how the region’s own capital behaves.
The uncomfortable trade-off nobody in SEA wants to name
None of this means Southeast Asia is being deliberately starved. Foundation-model economics are genuinely different: training frontier systems requires sustained, enormous capital in a way that a fintech Series B never did, and some of the money flowing into regional data centres is itself SEA’s cut of the AI infrastructure build-out, not capital diverted away from it.
Malaysia, Indonesia, and Thailand now have 31 planned data centres above 100 megawatts, versus just two a few years ago — real, if capital-intensive, participation in the cycle.
But participation in infrastructure is not the same as participation in venture. A data centre lease is not a Series A term sheet, and the jobs and equity created by hosting compute for someone else’s model are structurally different from the jobs and equity created by building the model, or the application layer, yourself.
If the next decade of technology value accrues overwhelmingly to two or three frontier labs and the infrastructure landlords around them, Southeast Asia’s founders need a funding strategy that does not depend on a rising tide that has, for the first time in venture-capital history, stopped lifting most boats.
What actually needs to change
The honest response is not another op-ed lamenting a “funding winter” as if it were weather. It is building permanent, regionally-controlled capital — sovereign funds, corporate venture arms, and family offices willing to underwrite unfashionable sectors precisely because global LPs have stopped bothering to look at them.
Also Read: The end of Southeast Asia’s unified startup funding story?
It is also being candid with founders that the pitch of “just build something AI-adjacent and the capital will find you” is a bet on scraps from a table two companies are eating alone.
Southeast Asia’s startups do not need to out-raise OpenAI and Anthropic. They need capital that was never going to chase them in the first place, and right now, that capital is nowhere near enough of it.
The post When 43 per cent of the world’s funding goes to two firms, where does that leave SEA? appeared first on e27.