The past 18 months have witnessed the dominance of a small group of tech stocks in the U.S. equity market. Collectively dubbed the Magnificent Seven (Mag 7)1, these stocks as a basket have appreciated by 131% since January 2023, as investors flocked to capture the potential of generative artificial intelligence (AI). Investors have good reason to want exposure to this technology. AI is expected to drive a sharp increase in productivity and ultimately be a deflationary force.
While these tech stocks have benefited from AI in the near-term due to the infrastructure buildout, it is important to look beyond the “picks and shovels” phase. We believe the recent pullback in public stock prices is the market questioning the return on investment of this capex buildout. Alternatively, Venture Capital looks to benefit from the formation of next phase generative AI companies that will be created over the long-term.
Since late 2022 when OpenAI released ChatGPT and became the fastest application to 100 million users, conversations surrounding technology have revolved around generative AI, which transforms user prompts into computer generated text, code, images, audio, and video. The technology powering ChatGPT, known as large language models ("LLM"), holds the potential to fundamentally transform how we create, work, interact, and live (think on the order of the advent of electricity).
As such, value creation stemming from generative AI will likely accrue across all industries and to companies small and large, including legacy incumbents, who have been the immediate beneficiaries thus far. NVIDIA's revenue has skyrocketed from selling computer chips (GPUs) primarily to other mega-cap technology companies (Alphabet, Amazon, Microsoft, Meta) in what has become an arms race to outspend each other. These four companies, among others, are investing unprecedented amounts to build the infrastructure required for generative AI to function (and with hopes of future revenues). Public investors, expecting meaningful increases in revenue from sales of AI infrastructure, have piled into the Mag 7:
However, some analysts more recently are questioning whether these companies will be able to generate the vast amounts of revenue required to justify their respective investments into AI infrastructure. Sequoia Capital, the preeminent venture firm, recently estimated that in order to payback the aggregate 2024 capex infrastructure investments, major tech companies would need to generate $600B in annual AI revenues2.
Media coverage of the arms race to develop AI infrastructure often obscures the true impact and evolution of the technology. The infrastructure buildout, or capex cycle, by the big tech players is the railroad upon which generative AI start-ups will operate. We believe new start-ups stand to benefit from the investment buildout and enjoy the efficiency gains and reduced hosting prices over time (see chart below)3. To contextualize just how much railway has been laid, Barclays estimates that aggregate AI capex by 2026 will be sufficient to support 12,000 AI products of the scale of ChatGPT.
Similar to how the internet, cloud, and mobile super cycles unfolded, entrepreneurial founders and venture capitalists are today prognosticating, experimenting, and investing into new market categories and companies which will use AI to disrupt incumbent business models. And so, while investing into big tech players may enable investors exposure to generative AI, primarily in the picks and shovels phase, we believe the next class of generational companies – where significant value will be captured – are just getting started.
- Alphabet, Amazon, Apple, Meta Platforms, Microsoft, NVIDIA, and Tesla
- Sequoia Capital: https://www.sequoiacap.com/article/ais-600b-question/
- https://www.coatue.com/blog/perspective/robotics-wont-have-a-chatgpt-moment
