Cheap to Build, Hard to Back

Philippine startups raised about twice as much money in the first half of 2026 as in the six months before. On a chart, that reads as a recovery. Look one layer down and the picture changes: fewer companies were funded than at any point since 2019.

Bilyonaryo News Channel put the tension in one line in a recent interview with Bit Santos of Kickstart Ventures: AI has made startups cheaper to build than ever, and that has not made winning investors any easier.¹ It is worth taking that seriously, because it changes what a Filipino founder should actually be optimising for.

What is happening

The numbers come from the H1 2026 Southeast Asian Startup Funding Report by Kickstart Ventures and DealStreetAsia. Philippine startups raised US$81 million in the half, up from US$33 million in the previous six months. Over the same period the number of funded deals fell from nine to seven, the lowest count since 2019.²

The money was also concentrated. Fintech took US$61 million, about three-quarters of the total. A single company, Salmon, accounted for US$50 million, and Salmon and Sprout together made up roughly 86% of disclosed funding.³ In other words, the "doubling" is mostly two rounds.

The region shows the same shape at a larger scale. Southeast Asia raised US$7.25 billion across 217 deals, its strongest half since early 2022. But one data-centre financing, DayOne's US$4.5 billion, was 62% of that, the top five rounds took about three-quarters, and deal counts were the lowest since 2018.⁴ The Philippines captured 1.1% of the region's funding and 3.2% of its deals.²

Why cheaper building doesn't mean easier funding

AI tools have compressed the cost of getting to a working product. A small team can now ship in weeks what used to take a funded engineering team months. That is real, and for founders it is good news.

But consider what it does to the investor's side of the table. If almost anyone can build a credible prototype, a credible prototype stops being evidence of anything. The thing that used to separate a fundable team from the rest, the ability to build, has become common. Investors see more products that work and fewer reasons to choose between them.

So the scarce thing moves. It moves to distribution: the ability to reach customers at a cost the business can sustain. It moves to evidence: revenue, retention and repeat usage rather than demos. And it moves to defensibility: data, relationships or licences that a competitor with the same AI tools cannot reproduce over a weekend.

Kickstart partner Jecky Pelaez made a related point about why startups stall: "they can't reach the customers that would prove the business."² That diagnosis matters more, not less, when building is cheap. Building is the step AI has shortened. Reaching and keeping customers is the step it has not.

What it means for Philippine founders

Read together, the figures suggest a market that is willing to write large cheques but is writing very few of them, and mostly to companies that already look like winners in regulated, scaled categories such as financial services. That has three practical implications.

First, the bar for a raise has moved from "can you build it" to "can you show that people want it and that you can reach them". A founder who spends the AI dividend on more features is spending it in the wrong place. The better use of a cheaper build is more time, and more money, on finding and keeping customers.

Second, cheaper building means many companies may need less outside capital to reach meaningful traction. Fewer deals is a problem for founders who need funding to survive. It is less of a problem for founders who can reach revenue first and raise on their own terms later. The cost reduction is real; the question is whether it is used to extend runway or to imitate the burn rates of a different era.

Third, distribution partners matter disproportionately in a market this concentrated. Corporate-backed investors in the Philippines often bring access to customers, not only cash. Pelaez described corporate participation as "a commercial opportunity as much as a capital one."² For a founder, an introduction to a large customer can be worth more than a slightly larger round.

The fair counterpoint

It would be easy to over-read six months of data. Seven deals is a small sample, and a single large round can swing a half-year total in either direction. The previous period's US$33 million was low, so "more than doubled" partly reflects a weak base. One strong half does not make a trend, and neither does one weak deal count.

It is also possible that fewer deals is partly a healthy sign: founders raising later, with more traction, rather than raising early on a deck. If AI genuinely lets teams go further on less money, a falling deal count could reflect fewer founders needing to raise at all, not only fewer investors saying yes. The data available does not separate those two explanations, and anyone claiming certainty either way is going beyond it.

What to watch

The useful signals over the next year are not the headline totals. They are whether the deal count recovers, whether early-stage rounds broaden beyond fintech, and whether the share of funding going to the top one or two companies falls. If building stays cheap and those numbers stay flat, the message for founders is clear: the product gets you into the room, but proof of demand is what gets you the cheque.