Shaping Tomorrow: The Role of Investment in AI Advancement

AI is changing the world. Nvidia, the world’s leading manufacturer of GPUs, the engine that driving AI applications today, has released its Q1 2024 revenue numbers and has yet again crushed earnings forecasts. Rewind the clock to Q1 2021, Nvidia’s revenue was 3.1billion, no paltry sum. The main applications were graphics cards from computers and crypto mining. This past quarter Nvidia did 26 billion USD of revenue valuing the company at a staggering 2.6 trillion behind Apple (2.9 trillion) and Microsoft (3.2 trillion) and ahead of google (2.2 trillion) and Aramco (1.9 trillion). Going from 100 billion market cap to 2.6 trillion in 5 years in unheard of. If you invested 10,000 dollars in Nvidia in 1999, it would now be worth close to 25 million dollars, a whopping return of 250,000 percent!

What’s more important than the valuation and revenue number is what these numbers  represent. The capital investment on chips, on related infrastructure a.k.a the rails, and on power generation represents a quarter to half a trillion to bring AI to the masses. Exciting times!

The interesting thing to remember is during the PC revolution in the 90s, intel peaked at a market cap of about 200 billion US dollars and from 1998 to today the average growth has been -1.4% each year. The nature of a competitive marketplace compels other players to jump in to try to capture value when any one company has outsized returns. And when other players invest and the industry catches up, value reallocates usually moving higher in the chain. For intel, that meant networking and the infrastructure on which that compute capacity would move i.e. switching, routing, security etc hence Cisco had its big moment and intel slumped. Thereafter value continued to move higher up as networking became commoditized (Huawei and others made cheaper equipment and Cisco stagnated) and the valued moved to browser companies, app companies, with SAAS and media companies (Netflix, Spotify, social media etc) together device and mobile (Apple, Samsung) companies which became more and more valuable and remain valuable today.

AI and the applications it will unlock represents a new age of compute. We are seeing the infrastructure build out along with language models and some early applications. Inevitably, and it may well take longer this time given increased complexity, the value has to reallocate up the stack resulting in new players entering the fray with novel methods for compute, infrastructure, and applications.

The key takeaway is that any company competing effectively in this arena is likely to continue to rip, specially those in the higher end of the stack such as Athropic’s Claude, Meta’s Llama, or OpenAI GPT4o and others.

The potential for significant productivity gains from Artificial Intelligence remains untapped, but so do the risks of its malicious use. Threats range from minor cyber scams to large-scale social disruptions, from the creation of harmful materials to catastrophic scenarios akin to Hollywood films. Prioritizing ethical, value-driven development and safety is essential to prevent irresponsible use. Specially as the systems evolve towards autonomous reinforced learning and Artificial General Intelligence (AGI). Given the rapid pace of advancement, self-regulation currently appears to be the chosen approach to balance progress with responsible growth but standards and regulations will certainly follow. Not all AI applications are equal in this regards and some stand out from the pack.

At S2015, we are committed to responsibly investing is start ups and technology companies that create positive social impact. We are allocating capital in order to continue to advocate for responsible growth and development. To this end, we made our first AI investment in Anthropic after carefully evaluating their ethos. Anthropic released its Responsible Scaling Policy (RSP) in Sept of 2023, the 22 page Anthoropic RSP Policy Document highlights their commitment to AI Safety Levels and rigorous red-team testing under adversarial conditions. We believe this represents an unambiguous commitment to safety and given that Claude 3 in consistently ranked in the top 5 Large Language Models (LLMs) in  independent testing, we are confident that Anthropic is a good fit in our portfolio.

We are moving into a brave new world and ground is shifting beneath us. Whether its in AI, health tech, SAAS, or anti-fraud technologies, we at S2015 are committed to bold responsible action, backing our words with strategic investments. Borrowing from Robert Frost,

“The woods are lovely, dark and deep. But I've promises to keep. And miles to go before I sleep. And miles to go before I sleep.”