I don’t know where we are in the AI hype cycle, but I do feel confident in one thing: We’re turning some kind of corner, or have.
And if you look at PitchBook’s recently released Q3 AI and ML report, there’s actually a lot of evidence that AI companies and investors are in a pretty different spot now than they were when this year began. Consider, for example, the trends around deal value (the total amount of capital invested in a given sector, in this case) for foundation models.
I honestly may have expected this number to take a dive in 2024, but it’s up and seemingly growing (we’ll know where it lands, as with all these numbers, in the full-year report). Broadly, part of what’s going on is that, actually, lots of AI companies have begun to prove their long-term worth, particularly when it comes to enterprise applications, said PitchBook senior emerging technology analyst Brendan Burke.
“While AI as a technology is still early in its societal impact, the enterprise GenAI hype cycle can see the end of a trough of disillusionment and shift to a plateau of productivity,” said Burke via email. “While this market size will fall short of the peak of inflated expectations, which foretold universal economic disruption from the current generation of large language models, large businesses can be justified from this new technology as with cloud computing before it.”
Alright, good news, right? Sure, but there are still a lot of questions, particularly when it comes to what exits can look like. (Exit value, by the way, refers to the aggregate value generated when an investor sells equity in a portfolio company.)
So, sure, buyouts aren’t much of an option, and acquisitions are more sparse for AI companies than they have been in recent memory. That leaves, of course, the toughest and most high-reward path for AI companies seeking an exit: an IPO. It’ll be interesting to see how this data shifts next year, should the IPO window blast wide open again as so many hope and expect.
I was particularly fascinated by the declining deal value at the later stages, which appears set to hit a four-year low right now. Burke says part of the late-stage drag is that pre-generative AI boom companies have struggled to find their way, especially as investors have struggled with their own 2021 hangovers.
“Most pre-GenAI companies have struggled to find their next chapter as the underlying technology races by them and investors look for new ideas,” he told Fortune via email. “Of the 546 AI & ML companies that raised a mega-deal of at least $100 million from 2020 to Q2 2022, only 287 have closed a deal since the VC downturn beginning in Q3 2022, including acquisitions.”
I can’t help but think it’s a reminder: Sometimes what’s old is new again, but sometimes what’s old is, well, old. And regardless, it’s clear there are still many corners yet to turn in the AI story.
Next week…Fortune Brainstorm AI is next week! It’s my first Brainstorm conference as a co-chair, and I’m so grateful to all the incredible investors and firms who will be joining me onstage and in my Term Sheet Breakfast roundtable (which, by the way, is already at capacity, as I learned yesterday). If you’re curious about the conference, you can learn more here. And those of you who are coming, thank you. I genuinely cannot wait to see you.
Elsewhere…Olympus Partners, a middle market private equity firm, has returned $3 billion to its investors this year, a person familiar with the situation told Fortune. Most, or $2 billion, came from Olympus’s sale of staffing company Soliant earlier this year. Olympus also sold Rise Baking in September, which contributed another $1 billion in distributions, the person said. Olympus has produced over $6 billion in distributions since January 2022. –Luisa Beltran
See you tomorrow,
Allie Garfinkle
Twitter: @agarfinks
Email: alexandra.garfinkle@fortune.com
Submit a deal for the Term Sheet newsletter here.
Nina Ajemian curated the deals section of today’s newsletter. Subscribe here.
