As AI technology and programs like ChatGPT evolve, the way venture capitalists think about investing in startups is changing.
Investor Leah Solivan, the founder of freelance marketplace TaskRabbit, which sold to Ikea in 2017, has been working as a venture capitalist for the last eight years. She currently works with startups building AI products as a general partner at early-stage fund Fuel Capital.
The process to build an AI company is "very expensive," she says.
Leah Solivan. (Photo by Chance Yeh/WireImage)
"[AI] is a big game-changing technology, but the costs are still so high to launch something," Solivan told entrepreneur Jeff Berman last week. "Startups need to raise a lot more money to get started right now."
Related: How to Start a Multi-Million Dollar Company, According to an IBM Engineer Turned Founder
AI models can take upwards of $100 million to develop, according to Anthropic CEO Dario Amodei.
Solivan says the cost of AI is changing where a smaller, early-stage fund like Fuel Capital invests. Big industry players like Microsoft and Nvidia, which have invested billions of dollars into AI companies, can afford to invest in expensive AI startups — but smaller, early-stage funds might not see the return on investment they're looking for.
So smaller funds could strategically choose to pass on AI startups because of the steep price, even if those startups are developing cutting-edge technology.
"It's almost like when we used to look at hardware companies and we were like whoa this is going to take way too much capital, the ROI on our investment, the math just doesn't work for our fund," Solivan explained. "You need really, really deep pockets to be successful. I think it's harder for the small funds to play here."
In 2023, AI was one of the best industries for growth in unicorns, or startups with at least a billion-dollar valuation.
AI was also the sector with the biggest jump in funding last year, with AI startups collectively raising $50 billion, even though the year was tough as a whole for startup fundraising.
Related: ChatGPT Is Writing Lots of Job Applications, But Companies Are Quickly Catching On. Here's How.