AI venture funding is accelerating and we see no sign of it slowing down. Deal count and capital invested in AI has risen year-over-year. In 2016 venture investments in US-based AI companies reached $4.9B, and as of the end of October, we have surpassed this number with $7B invested year to date.
The growth in deal count and capital in AI counters that of the overall market. On a quarter-to-quarter basis, early-stage VC activity has slowed. The infamous “Series A crunch” is a direct result of VC over-investment in early-stage deals. An abnormally high number of companies were able to raise seed rounds from 2012–2015, creating a large crop of companies that were weeded out at the Series A and Series B after struggling to demonstrate traction.
Given that AI is a relatively new investment focus, the sector appears to have been shielded from the pressures of the broader market. The number of Seed, Series A, and Series B deals have all risen year-over-year, demonstrating a strong appetite for AI companies. In addition, the proportion of Series A and Series B companies have increased, signaling a slight maturation of the sector.
Compared to valuations 5 years ago, we have seen a 2–3x increase in median pre-money valuations across all stages. This trend has continued for later stage deals, where we have seen a 30–50% increase in pre-money valuations in 2017. Surprisingly, however, median Seed valuations increased only marginally in 2017, and Series A valuations actually decreased. While this could indicate a softening in the market, we believe this may be a function of the volume of deals. In an active VC market, earlier and earlier stage companies are able to raise capital. The pure growth in the number of deals seems to suggest that the definition of “Seed” has not stayed constant, and that less progress may be needed to receive initial funding.
The hype around AI is not isolated to startups or venture. Google, Amazon, Microsoft, and Intel are market leaders in AI and have been active participants in the M&A market, acquiring numerous startups as they build out their AI capabilities. So far in 2017 there have been 32 acquisitions of US based AI companies, up from 28 deals in 2016 and 9 deals in 2015. Given the scarcity of AI talent, we do not anticipate acquisitions to slow down over the next few years.
In 2018, we expect the VC appetite for AI to intensify. More acquirers are likely to enter the market as fast-following technology companies attempt to catch up with Google, Amazon, Microsoft, Intel, and the like. We believe this will lead to an increasingly hyped market that will likely become overheated. In private markets where valuations are often determined by the highest bidder, pricing can easily become irrational. While we remain bullish on the potential of AI, it has become increasingly important to put our assumptions and expectations in check.
Laura Cain is a Venture Associate at Thomvest Ventures. This post first appeared on the Thomvest Ventures blog and has been reproduced here with permission.