Venture capital fundraising in Southeast Asia has dropped sharply over the past several years, leading some investors to wonder when the “winter” will be over. Venture funding into Southeast Asian tech companies plunged by 79% between 2022 and 2024, according to CNBC citing Tracxn data.
Helen Wong, managing partner at AC Ventures, said that green shoots can be seen and that “the best companies are still able to raise funding” while on a panel at the Fortune Brainstorm AI Singapore conference on Tuesday.
But the discussion revealed an underlying sense of urgency: It’s clear that the region is lagging and needs to catch up with more mature markets.
“There’s a tremendous opportunity for a lot of much more specific, highly differentiated visions that could potentially be billion-dollar companies,” said Matthew Graham, managing partner of Ryze Labs, but they need to be “above and beyond trying to be the 19th LLM.”
Jixun Foo, senior managing partner at Granite Asia, said Southeast Asia has lagged in adopting new tech, but DeepSeek has democratized access to AI, allowing more applications to be built. Yet Granite Asia is waiting to see “the confluence of talent and market adoption, and then we will start investing.”
Foo also worries that companies are being held back by their own legacies and risk disruption from AI natives. “The tech companies are becoming more traditional in the face of AI,” he said.
Ren Yeong Sng, managing director of artificial intelligence strategy and solutions at Temasek, said it was important to “invest in the bottlenecks” that appear as technology becomes cheaper. For example, agentic AI needs containerization, persistent memory, and persistent identity to be able to work. “Whenever startups come to us with problems like that, they’re addressing a bottleneck we think is going to be very unique,” he said. “That’s something we need to look at” when it comes to investment, he added.
Graham said founders in Southeast Asia should focus on their “super fans” who need their product. He said many companies “try to find their first 1 million users before they find their first three.”
