Tightrope balance balancing act

We are hearing increasing speculation that AI-focused start-ups may be engulfed in a market bubble. This week, Mistral AI, the French start-up launched just a year ago by former employees of Meta and Google DeepMind, raised $640 million from Nvidia, Salesforce and others, putting its valuation at roughly $6 billion, triple what it was in December.

Cognition Labs, launched last November, is reputed to be in talks to raise money from investors that would value it at up to $2 billion. That’s almost six times what the company was valued at after raising $21 million earlier this year. Cognition is developing a fully autonomous AI software engineer called Devin.

With no clear end in sight for the exit drought, toppy multiples could make some VCs put new AI bets on pause if not for fear of missing out, which never fully went away. Touring Capital, an AI-focused venture firm I spoke with recently, hopes to protect itself from inflated valuations by choosing to not invest in hardware or infrastructure, at the foundational layers of the AI tech stack, but instead to invest in B2B apps being built on top of them.

Touring, based in San Francisco, is currently raising its debut fund, the Oakley Touring Venture Fund. It is targeting $300 million, with a final close expected later this year, according to a person familiar with the fundraising. LPs include Oakley Capital and 30 founders that Touring’s partners previously backed, co-founder and general partner Priya Saiprasad told Venture Capital Journal last month.

The fund’s sweet spot is Series B rounds for AI-enabled B2B software companies earning $3 million to $10 million in annual recurring revenue and that have figured out the basics of their product-market fit.

Two reasons for the fund’s size are Touring’s desire to lead rounds with checks of $15 million to $20 million and wanting to “de-risk and diversify the portfolio” by making 12 to 15 investments. Most of those will be in the US, but the fund is also interested in companies in Europe, India and Australia.

Saiprasad and her co-founders and fellow GPs Nagraj Kashyap and Samir Kumar met at M12, Microsoft’s VC arm. Kashyap and Saiprasad also worked together previously at SoftBank’s Vision Fund. Collectively, they have backed 16 unicorns and helped steer 26 companies to successful exits, including Zoom, Kahoot and Livongo.

Touring has announced four of the six investments it has made so far. Last September, the firm led an $85 million Series C1 round for Pixis, which has developed codeless AI technology that provides marketers with plug-and-play AI products without their having to write any code. The new funding is enabling Pixis to further develop its capabilities, invest in R&D to refine and launch a generative AI-powered creative studio, and build strategic product and business partnerships with various social media brands.

Another portfolio company is Daloopa, whose AI-powered extraction engine and modeling co-pilot helps hedge fund analysts and equities analysts at large investment banks to update their investment models faster and with more accuracy. Touring led Daloopa’s $18 million Series B round last month. The company is shifting from a bottom-up approach that starts by selling to particular analysts to a top-down sales model that lets Daloopa target an entire team. Morgan Stanley invested in the latest round, and Daloopa is thinking about partnering with other large banks as distribution partners, too.

Touring sees less risk in backing companies whose software is “purpose-built for some sort of business user and has an ROI that delivers, let’s say, increasing ad spend or decreasing cross-selling here and there and direct, tangible ROI upfront,” Saiprasad said. She emphasized that each company must have some kind of data moat, or data-based competitive advantage, that “you can leverage in a very targeted, thoughtful way, and eventually once you get so much customer data and domain data that’s very exclusive to you, you can start to build your own customer foundation models in-house.”

Wary of the hype cycle that has consumed AI since the launch of ChatGPT in late 2022, Saiprasad and her co-founders deliberately chose a more conservative investment thesis.

“By the time we get [an investment opportunity at the Series B stage], you still have software valuation multiples, not AI multiples, really,” Saiprasad said. “But our thesis is, eventually over time, you prove out that AI moat, and then hopefully when you exit, it’s either at AI multiples if the market stays the same. If the market crashes, it’s software multiples that you fall back to. Because we’re entering at software multiples, we’re not really banking on the AI markets staying as hot as they are today.”

The fourth paragraph has been revised to reflect the accurate source for the fund size, and the seventh paragraph has been updated to show that only two of the three partners previously worked at SoftBank.

Share this