MONTREAL — Leading AI startups are growing so fast that they’re forcing investors to rethink how they evaluate firms in the sector, venture capitalists say.

San Francisco-based Anysphere’s AI coding assistant Cursor reportedly hit US$100 million in annual recurring revenue in January, just 14 months after launch. The firm’s revenue run rate had topped US$500 million by June. Stockholm-based Lovable Labs, which also makes AI development tools, claimed in May to have reached US$50 million in annual recurring revenue in six months.

This rapid growth is making investors raise their expectations for startups, said Elizabeth Yin, co-founder and general partner at Hustle Fund, at Startupfest in Montreal on Wednesday. “The bar has gone up.”

Toronto-based Georgian has long used the same seven metrics to assess the software-as-a-service companies it’s considering backing, including how efficiently firms use their funding and how quickly they expand their customer bases. The financier is now reviewing its process for comparing companies because some AI startups are increasing revenue so rapidly, according to lead investor Emily Walsh. “There’s no clear playbook today,” she said.

For firms developing AI applications, rapid revenue growth now isn’t a guarantee of success in the future, the venture capitalists warned. “I do think it’s a bubble,” said Isaac Souweine, a partner at Pender Ventures. While some companies adopting AI tools are seeing real benefits from the technology, others are buying them so that they can say they’re using AI, he said.

Walsh said lots of sales to big businesses may be for pilot projects to test out the technology; those may not turn into long-term contracts. Still, Georgian’s regular survey of business AI adoption shows that more firms are now using AI in their day-to-day operations, compared to a year ago when most were just experimenting with the technology.

The AI application startups that are growing fastest, like Anysphere or the image-generation tool Midjourney, tend to target developers and other professionals. That may help them hang on to customers, according to Yin. Users are willing to pay for code assistants, image generators and other AI tools that increase their productivity, and may stick with those apps as new alternatives emerge.

“I’m not going to cancel my Midjourney account,” Yin said. “But I’m not going to go and buy 20 other pieces of software.” It will be harder for new firms to break into the market if customers are all stocked up on AI tools.

Still, it can be hard for investors to pick which contender to back. Silicon Valley’s Menlo Ventures held off on backing a code-generation startup at an early stage because “it was really not clear who was going to win,” Amy Wu Martin, a partner at the firm, said on Thursday. “These are really expensive rounds where you have to put a lot of capital at risk to make a bet.”

Anysphere, for example, has raised US$1.08 billion to date, according to PitchBook data. Much of that is from a US$900-million Series C round in June 2025 that valued the firm at US$9.9 billion, up from US$400 million in its Series A round just a year earlier.

AI is also changing how much capital startups need to raise, and how they spend that money. Firms can now access “this mega-computer brain” to help with their coding instead of hiring a lot more technical staff, Souweine claimed. “We do expect that you can do more with fewer engineers than ever before.”

That trend is borne out by data from Carta, which makes software that startups use to manage their fundraising and equity. Founders are waiting longer to hire their first few employees, head of insights Peter Walker said in a presentation at Startupfest on Wednesday.

Startups are also growing their workforces more slowly. A few years ago the typical firm raising a Series A round had 22 employees, according to Walker. As of 2024, the number was 15, and Carta predicts that will fall to 12 this year.

For all the noise in the AI market, investors are betting some startups will become sustainable businesses and deliver big wins for their backers and customers. “We’re in an era of technological transformation,” Walsh said, and it will produce “very large companies.”