He has invested in over 600 companies and manages $72 million in assets. But you won’t find Vibe Coding Company in his portfolio of early-stage startups.
There is no room for more new entrants in the AI-assisted coding space, said Jussi Sarovara, co-founder and managing partner of startup accelerator Antler Asia.
“It’s great, but I don’t feel it makes sense to start another vibe coding company now,” Sarovara, who is based in Singapore, told Business Insider. “This is an exciting category, but the players are already there.”
He said he sees his colleagues at Antler and the founders he has invested in using AI to write code more efficiently and build tools. The company’s European fund was an early investor in the $6.6 billion Swedish vibe-coding darling Lovable.
Still, Sarovara said he doesn’t believe vibecoding is a viable category for new early-stage investing.
He said startups are always considering how to protect themselves from advances in the industry.
“A million people will do the same thing. How will you differentiate? Anthropic will ship a product next week. What will happen to you? The cost of the model will increase five times. What will happen to you?”
The future of vibecoding and the businesses it has begun to disrupt is one of the most hotly debated topics in the tech industry this year. Even as professional developers continue to say that AI-encoded software is risky and ultimately requires human input, startups like Lovable, Cursor, and Emergent are raising billions of dollars, closing big deals, and soaring valuations.
Sarovara said there could be some consolidation among existing AI coding companies.
“It’s going to be like a collective system of a few winners, and they’re going to stay and start buying smaller companies and continue to stay at the top,” he said.
But there’s one type of AI company that venture capitalists of the last eight years say they’re excited about.
Mr. Sarovara said he wants to invest in companies whose founders have deep expertise in areas that Antler has focused on for years.
“They are combining this advanced AI capability with industry capabilities, so the customers are going to be automotive and advanced manufacturing,” he said.
He said he is thinking about whether improvements to the baseline model will improve the product or replace it.
As an example, at one of his portfolio companies, IndustrialMind.ai, former Tesla employees are building AI agents to optimize factory operations, including manufacturing and process engineering. Another startup founded by a former professional filmmaker is building video editing tools for other professional filmmakers.
“We often see people not taking the time to build for the space,” he says. “That’s something that always bothers me, does it even make sense?”
Not all accelerator programs are putting the brakes on investing in AI coding startups. Y Combinator and Andreessen Horowitz’s Speedrun group of early-stage startup investors includes several companies building AI coding agents and AI app builders.