As the global and U.S. venture capital markets reach new heights in Q3 2017, one of the most popular sectors helping to drive that growth is artificial intelligence. Startups in the space also continue to attract acquisition offers from major companies looking to maintain their edge against plucky upstarts.
However, although dollar volume into AI startups is hitting records, it also appears the barrier to entry for AI startups to secure their initial funding may be growing.
To get a clear picture, we will look at artificial intelligence startups as categorized by Crunchbase, including seed-stage deals, early-stage deals, and exits by way of acquisition.
VCs Reap What They Sow
Overall, early-stage dealings in US-based AI startups are seeing an increase in funding. However, compared to the same quarter last year, round counts inclusive of all stages are down.
In fact, Q3 2017 is the worst-performing quarter in terms of known deals counts since Q4 2016.
That said, for startups that managed to land a deal in the most-recent quarter, the funding amounts were likely favorable. In terms of dollar volume, Q3 2017 hit a record with $1.165 billion in known fundings. And while deals counts in the third quarter hit record lows, dollar volume compared to the same quarter in 2016 increased by over 200 percent.
If we rely on dollar volume only, Q3 2017 is performing above par. But given that the AI sector is, in general, quite hyped, what could be leading to a decrease in overall deals?
The Early Bird Is Sleeping In
Following the pattern set by our general reporting the US VC market, we are seeing a steady decline in the number of seed and angel-stage deals. However, early-stage deals (Series A and B) now command a larger share of the funding pie in terms of deals and dollar volume, pointing to a maturation in the AI sector. Here’s how each stage of funding breaks down.
Seed Stage Loses Footing
As the chart below shows, seed-stage round counts peaked in Q1 2017. Since then, the total number of reported seed-stage deals in AI startups have dropped by approximately 53 percent.
And in percentage share of deals, we see a similar pattern:
Despite the precipitous fall in the number of seed-stage deals, startups still walked away with a good chunk of investor change.
Olono, which landed $4 million in seed funding in early September, uses machine learning to help sales reps close more deals. And eventually, Element Data’s AI wants to shape how you make decisions. To do so, it has raised $3.5 million in seed funding.
Overall, the decline in seed-stage deals could imply that VCs are making smarter bets, requiring entrepreneurs to dig deeper for market opportunities. But for those who have already raised an angel or seed-stage round, going for a series A or B shows encouraging signs.
Early Stage Makes Gains
While seed-stage deals go down, early-stage deals are seeing a consistent uptick in terms of deal volume counts and percentage share (as outlined in the charts above). In fact, Q3 2017 has set a record with 56 known early-stage deals compared to 41 known early-stage deals in the same quarter of last year.
In terms of percentage of dollar volume, early-stage AI companies in Q3 2017 are also performing well, as the chart below shows:
However, this record in early-stage deals is largely attributable to Plenty’s $200 million series B. Backed by Softbank’s Vision Fund, the startup is looking to reinvent indoor farming with the help of AI and machine learning. Following Plenty is HouseCanary, a real estate analysis startup, with a $31 million series B.
Overall, the trend for early-stage startups is positive, and the uptick in early-stage deals in dollar volume and deal counts could be pointing to a maturation of the sector.
However, with seed-stage deals falling, the pipeline for series A and B rounds is possibly dwindling. If investors and startups can continue the growth of the AI startup sector with early and late-stage dealmaking is not known today. But in regards to measuring the health of AI startups, it’s an important metric we will keep note of for future quarterly reports.
Of course, millions don’t get poured into AI startups for fun. Investors want an exit, and given that the AI category is relatively new, acquisitions are likely the most effective way to get it.
As we determined in past reporting, AI startups are attractive commodities for the Big Five (Microsoft, Apple, Facebook, Alphabet, and Amazon). And although Q3 2017 didn’t set any records in acquisition count, a few familiar names make our list of acquisitions.
In July, Facebook acquired Ozlo for an undisclosed price. The Palo Alto-based company, which raised $14 million in funding with Greylock Partners and others, builds conversational chatbot APIs. The team, according to TechCrunch, will be working on Facebook’s Messenger platform.
Only one other company that is a part of the Big Five acquired a U.S.-based AI startup in Q3 2017. For an undisclosed price, Amazon acquired AI-powered data presentation startup GraphIQ. No price was disclosed.
Google and Apple, meanwhile, decided that their needed AI talent resided across a pond or two. As Ingrid Lunden of TechCrunch put it: “talent doesn’t have to be located within spitting distance of US 101 to get noticed.”
Google has backed up that statement by acquiring two international AI startups in Q3 2017. Its first acquisition of the quarter, AIMatter, processes images using AI. The company’s home is in Belarus, which has recently become known for its unique startup culture. Meanwhile, Halli Labs, a deep learning systems startup based in India, was acquired by Google after only four months of in operation.
For Apple, image recognition and analysis also appeared to be a priority. The company acquired Paris-based Regaind, which uses AI to help developers sort masses of images effectively.
Notably, Microsoft did not make the list. This is despite the fact that it is one of the most active acquirers of AI startups, placing behind only Apple and Google.
So what does this all mean for AI startups? Right now, it’s a toss up. While funding may be increasing, fewer deals in seed-stage startups could be pointing to a decreased appetite in new AI ventures. It also could mean that the hype for AI is beginning to peak, and investors are being much more careful about the early bets they place.
If only there was a bot VCs would fund that could tell us.
Illustration: Li-Anne Dias