As we see unicorns slash staff and the prevalence of down rounds spike, it may seem that the startup ecosystem is chock-full of bad news and little else. That’s not precisely the case.
While AI, and in particular the generative AI subcategory, are as hot as the sun, not all venture attention is going to the handful of names that you already know. Sure, OpenAI is able to land nine and 10-figure rounds from a murderer’s row of tech investors and mega-cap corporations. And rising companies like Hugging Face and Anthropic cannot stay out of the news, proving that smaller AI-focused startups are doing more than well.
In fact, new data from Carta, which provides cap table management and other services, indicates that AI-focused startups are outperforming their larger peer group at both the seed and Series A stage.
The dataset, which notes that AI-centered startups are raising more and at higher valuations than other startups, indicates that perhaps the best way to avoid a down round today is to build in the artificial intelligence space.
What the data says
Per Carta data relating to the first quarter of the year, seed funding to non-AI startups in the U.S. market that use its services dipped from $1.64 billion to $1.08 billion, or a decline of around 34%. That result is directionally aligned with other data that we’ve seen regarding Q1 2023 venture capital totals; the data points down.
Series A rounds in Carta’s database from non-AI startups also tumbled in value, raising $2.08 billion in Q1 2023, off 54% from $4.52 billion in the final quarter of 2022. In contrast, AI-focused startups saw their total capital raised at the seed level in the first quarter rise 2.5% to $199.2 million, while the value of Series A deals for the same cohort ticked up 58% to $754 million compared to Q4 2022.
Simply put, more capital went to AI startups in Q1 2023 than in the sequentially preceding quarter. In contrast, non-AI startups saw their capital inflows constrict over the same timeframe.
The data gets even better for the U.S. AI-focused startups, raising a median $3.4 million in the first quarter of the year at a $16 million median pre-money valuation. In contrast, other seed-stage rounds that Carta saw in the first quarter raised less ($3 million) at a lower pre-money valuation ($13 million).
Series A was much the same, with American AI-focused startups snagging a median $12.5 million (up 23% from Q4 2022) at a median pre-money valuation of $52 million. In contrast, over the same timeframe non-AI Series A rounds that Carta handled raised a median of $5.9 million (off 25% from Q4 2022) at a pre-money valuation of $37.5 million, with that final number representing a gain of 7% or so from the last quarter of last year.
Every founder wants to raise money at a comfortable up-valuation. Given that AI-focused startups are able to raise more, and at higher prices, we can infer that venture capitalists are willing to pay more for revenue and corporate progress at those companies.
There is material risk in the choice to build AI products and services, however. Many of the leading LLMs that tech companies of all sizes are levering are built by others and hosted atop a third-party cloud service. That means that there is immense platform risk at play for startups busy using someone else’s models in their work.
There are open source models in the market, and they are improving at a seemingly rapid pace. But because those models are themselves open for general use, it may prove difficult for startups leveraging them to build a long-term moat. In other words, if you are building a house with the same bricks as your competitor, it’s harder to stand out than if you had an advantage at the unit-construction layer.
Those concerns appear to be less powerful than the sheer excitement thrumming through the technology landscape for new AI services; long waiting for the next platform shift in tech, venture investors are betting that this is it. It was supposed to be crypto, but as that revolution remains stuck in somewhere between neutral and first gear, AI is powering ahead not only in terms of founder and investor interest, but also in terms of where it is finding rapid adoption.
And where would that be? Everywhere, it seems. Startups want a chunk of that market. The only question we have now is whether this revolution will be captured by a handful of players oft bone-grafted to the public cloud of one mega-tech company or by startups.
Comment