In 2023, venture capital experienced something of a retrenchment from the exuberance of previous years. U.S. VC spending on early-stage companies reached a six-year low in the third quarter, dropping to just $8.5 billion across an estimated 1,341 deals. This decline reflects a broader trend observed globally, with VC funding dropping 13% quarter over quarter, marking the lowest level since the second quarter of 2020.
Yet the dip in VC investment isn’t entirely indicative of a sector in distress; many analysts view the decline as part of a market correction following an unprecedented investment boom in previous years. And as Ben Miller, CEO of alternative asset manager Fundrise, and Tomasz Tunguz, founder of the VC firm Theory Ventures, discussed in a recent “Onward” podcast, there’s another trend on the horizon that could shake up this downturn in spending: the advent of generative artificial intelligence startups.
This new wave of AI, powered by large language models like GPT, can create original, meaningful content and analysis by learning from unstructured data like text, images, or audio.
“In the last, say, two years, most of the innovation in AI has happened at the infrastructure layer. Nvidia has had one of the most incredible runs in the public markets in history because they have chips that everybody needs to use,” said Tunguz.
“Now there’s a wave of companies that are simplifying or building abstractions or layers on top, where if you are a great front end software engineer, an engineer who builds websites or builds web applications, you can start to use these products,” he continued. “At the same time, we’re starting to see the very first applications at the application layer. Historically, it’s been very expensive to get value from unstructured Information, and now we’ve figured out a way of mining that data to get value from it.”
“The amount of hype is just extraordinary,” added Miller. “There seems like there’s two ways to play it. These are figuring out where the value will accrue or selling picks and shovels.”
Venture Capital Investment in AI
Enthusiasm for LLMs and generative AI has led to a resilient market for AI and data infrastructure investment despite the broader downturn in general VC funding. The AI market is projected to grow to $407 billion by 2027, a significant increase from its $86.9 billion revenue in 2022. The global market for AI infrastructure, the data foundation for the burgeoning AI sector, is expected to grow from $33.66 billion in 2022 to $128.25 billion in 2027 at a compound annual growth rate of 30.8%.
These figures are reflective of a sector that’s not only expanding, but also becoming increasingly integral to a wide array of industries. According to some recent estimates, corporate spending on AI-centric systems could surpass $300 billion by 2026.
“AI is a generational breakthrough. Is it bigger than the internet? I think it probably ends up being bigger than the internet, definitely bigger than a personal computer. It’s so fundamental that it’s hard to actually get your mind around, but there’s patterns to it and it’s really exciting,” said Miller in an interview with The Motley Fool in October.
“There’s going to be a lot of money, growth, and excitement. Most people won’t be able to participate because it’s happening in the private markets. There’s 10 companies in the public markets that are leveraging AI, and Nvidia is one of them. But most of the stuff is happening with startups.”
It was this lack of access for individual investors that led Miller to start the Fundrise Innovation Fund last year. The Innovation Fund offers retail investors an opportunity to buy into a venture-style fund that is backing early-stage tech startups, especially those in the AI and data infrastructure spaces. Some of its early investments include Databricks, Canva, and Tunguz’s Theory Ventures.
‘The Equivalent of Investing Corporations’
Fundrise’s Innovation Fund was in part a response to a trend in VC of larger funds getting the majority of access to potential high value companies. The past decade or so of evolution in the venture space has seen the industry shift to centering on large firms like Sequoia and Andreessen Horowitz, firms that made huge profits off earlier tech booms and now have massive amounts of capital and reputational currency as a result. This leads to new deals and perpetuates a cycle of continuing growth and access to the best private companies.
“The venture capital asset class has grown from about 8 billion to about 250 to 300 billion in 12 years,” said Tunguz. “That’s a pretty huge change, and about 80% of the increase is from what you would call nontraditional venture capital. What we mean by that is venture capital grew up as an industry where there were three or four people around the table, then they put in a little bit of money, and then that led to a partnership. Those partnerships have been formalized and now we have the equivalent of investing corporations.”
Tunguz has been a part of this evolution of VC. With experience at both Google and Redpoint Ventures, he’s helped discover and fund seven unicorn startups, including Looker, which grew revenue by over 100 times and was acquired by Google for $2.6 billion. He started Theory Ventures last year with $230 million in funding.
Fundrise’s stake in Theory Ventures is $5 million. A letter to investors announcing this partnership explained that Tunguz’s approach to venture is aligned with that of the Innovation Fund, and that investing in a VC like Theory Ventures is a way to give Fundrise investors more access to a tech-investing fund.
“In our view, Tomasz ranks among the top 0.01% of emerging venture investors and has the potential to become one of the leading VCs in the coming AI and data revolution. We’re thrilled our investors are able to back Theory Ventures, an opportunity historically reserved only for endowments and ultra-high-net-worth individuals.”