AI… and Everything Else

Written by Christopher Walsh, Partner @ 7GC & CO

Back in 2023, we discussed the fundraising environment of AI compared to all other technology verticals. While the data points have changed, the overall status quo has remained incredibly constant.

We have watched AI continue to evolve and believe the technology has matured to the point that it is driving a transformational technology platform shift similar to those seen once per decade previously:

  • Personal computer in the 1980s

  • Internet in the 1990s

  • Mobile in the 2000s

  • Cloud in the 2010s

Like previous compute cycles, venture capital quickly looked to deploy capital rapidly across generative AI despite, at the time, an inflationary environment, frozen capital markets, and distressed selling from GPs following a collapse of ZIRP-era (Zero Interest Rate Policy) investing. Unlike other previous hype cycles, the incumbents, known today as the "magnificent 7", were well-read and coached with Clayton Christensen's Innovator's Dilemma playbook. While OpenAI delivered the catalyst with the release of ChatGPT, the response from Amazon, Google, Microsoft, and Meta was swift and came with a war chest that accumulated over $251B in free cash flow in FY23.

Cloud Platform Capex is Expected to Grow Over 40% this Year
Source: CapitalIQ; as of 08.15.2024

This has dramatically magnified investment volumes and innovation within generative AI, which has considerably sped the cycle's "build" phase. This spending has been visible in the public markets, not only in the semiconductor industry but also in the downstream "picks and shovels" supply chain that helps power semiconductor production. With history often repeating itself, we were under the assumption back in the second half of 2023—when we wrote our initial AI piece—that public markets had a strong handle on future expectations. We were wrong.

Market Consensus in July 2023 was that Nvidia Capex Cycle Peaked; the Opposite was True
Source: CapitalIQ; as of 08.15.2024

As the public markets continue to acclimate to generative AI, private markets continue to show fundraising KPIs that differ dramatically from those of other VC-backed technology cohorts, beginning with total capital deployed. Cumulative funding across private AI companies has continued to be parabolic in the last 12 months, doubling in just 12 months since 2023 and now eclipsing $70 billion.

Cumulative Private AI Funding Has Increased by 33% in Just Two Quarters
Source: Pitchbook data; as of 08.15.2024. Foundational models include 01.AI, Adept, AI121 Labs, Aleph, Anthropic, Character.AI, Cohere, Deepseek, Hugging Face, Midjourney, Mistral, MosaicML, OpenAI, Reka, Runway, Stability, Together

When stripping out magnificent 7 investment into Generative AI, the results sober the environment slightly but still represent aggressive demand for top deals, driving premium valuation relative to non-AI peers. We would be amiss not to be mindful of the "hype cycle" valuation risks that surround AI and the corresponding regulatory risks. And while the AI revenue opportunity remains in its infancy and will take time to mature, technology investors see clear evidence that it is already leading to the creation and destruction of large profit pools, driving clear "power law" opportunities for fund managers over the next 10 years.

Magnificent 7 Continue to Invest Heavily Across AI / ML Applications Deals
Source: Pitchbook data; as of 08.15.2024. Includes cumulative round size of rounds participated in and not the specific investment amount for each transaction; Magnificent 7 is defined as Amazon, Apple, Tesla, Google, Nvidia, Meta, and Microsoft
The greatest shortcoming of the human race is our inability to understand the exponential function.
— Albert Allen Bartlett

Peter Thiel famously said, "We don't live in a normal world; we live under a power law." For larger funds, this is a mathematical certainty. Those firms have already begun to amass an extensive portfolio across the entire generative AI stack.

The "power law" implication is that 1 in 20 deals may produce two-thirds of all returns, and 1 in 100 deals may return more than all other deals combined. This skew of outcomes is the shielded reality for the new wave of venture capital megafunds born out of the ZIRP era.

“Power Law” Drives Volume for >$1B Venture Funds
Source: Pitchbook data; as of 08.15.2024 

Given this reality, we have seen valuations and round sizes exponentially higher for these generative AI deals than for non-AI deals. In 2024 YTD, valuations were ~5x higher for AI deals, and deal size was ~6x higher versus non-AI venture deals, respectively.

Comparison of AI vs. Non-AI Rounds in the United States in 2024
Source: Coatue Management; As of June 2024

Given this disequilibrium, we continue to see opportunistic air pockets within other verticals, including FinTech, Infrastructure SaaS, Gaming, Sports, and AdTech. We see these verticals as underfunded and undervalued relative to historical levels, with many tourists packing up and moving to frothier pastures.

"Non-AI" software remains a significant sector with hundreds of high-quality businesses. Traditional SaaS has suffered since 2022 primarily due to deteriorating fundamentals, but not a major paradigm shift. For many verticals, the simplest explanation of deteriorating fundamentals is the combination of the extinction of free cost of capital and the maturation of the S-curve, marking more challenging directives to drive incremental revenue for many businesses. Because companies lost the ability of free capital, options like 1) spending more to bring incremental revenue and 2) reckless product road mapping for short-term upsell opportunities quickly evaporated. This "tough macro climate" equates to a new reality for many of these verticals. The dispersion of idiosyncratic companies provides clearer visibility for investors in evaluating who is executing versus who is not.

Execution is Still Being Rewarded Outside of Generative AI for Best-in-Class Technology
Source: Pitchbook data; as of 08.15.2024

These companies do not have generative AI in their business description but continue to drive best-in-class KPIs, showcasing clear new incremental ARR, net revenue retention, and innovative new product deployment, which drives strong upselling motions with customers.

While Opportunities Exist Elsewhere from Generative AI… Investors Follow the Money

Supply and demand are two sides of every equation for any business. While we have made clear that the supply of new generative AI startups is rampant, where is customer demand? While 2023 was about exploring these new possibilities for customers, 2024 has been the year where results matter.

Bain & Company has conducted quarterly surveys to assess readiness and deployment capacity among enterprises.  At the beginning of 2024, 87% of companies surveyed said they were already developing, piloting, or deploying generative AI in some capacity. The average annual spend across the cohort was $5 million annually, with an average of 100 employees dedicating at least some of their time to generative AI. Among large companies, about 20% are investing up to $50 million per year, with more than 60% of companies surveyed seeing generative AI as a top three priority over the next two years. While that figure is astonishingly high, only 35% have a clearly defined vision for how they will create business value from generative AI.

Investment Concerns Have Subsided for Enterprise and Replaced by Org Readiness 
Source: Bain & Company 

This prioritization, coupled with the fact that enterprise AI represented less than 1% of cloud spend in 2023, suggests that the greenfield opportunity for winners remains astronomical. While first-wave new entrants will be challenged to gain traction, the early winners will be able to provide a novel and differentiated offering that allows enterprises to validate proof of concept, ROI, and continued investment.

Key Barrier for Generative AI Over the Last 12 Months Was ROI
Source: Menlo Ventures; June 2023

When looking back previous technologies in their first decade, the cloud reached 30%, the internet reached 45%, and mobile reached 80% of enterprise software spend. For ChatGPT, between 25% and 33% of the entire developed world's population has tried the product in the first 18 months since launch, suggesting an extremely fast trajectory. However, there is still a question on utilization and stickiness.

OpenAI ChatGPT Utilization in the Developed World
Source: Reuters Institute; May 2024

The combination of high stakes, consumer exuberance, and broad accessibility of the technology has driven a Goldilocks effect for generative AI. This has led to a quick race to the bottom, making LLM technology ubiquitous faster than previous technology cycles. During the mobile revolution, building out an application on the app store allowed you to differentiate your product; now, a mobile application is required to remain competitive ("table stakes").

This rapid commoditization ties back to our previous statement that incumbents have displayed they can offer and aggressively market the same AI technology to their existing user base. With the most valuable companies in the world leveraging their pre-existing distribution and investing aggressively, we are streamlining utilization to a point where better engagement from gen AI applications should be just around the corner.


Previous
Previous

What is it like out there? Our Insights on the VC market

Next
Next

7gc summer reading list