Who is Winning in the AI Race?

Let’s analyze who is benefitting from AI today. It is clear that AI is revolutionary and will be here to stay. There winners though, aren’t completely clear.

Nvidia, which supplies GPUs (graphics processing unit) that power large language models, is now the most valuable company ($3.3 trillion market cap).

Nvidia’s stock has exploded in recent years. Most notably, there was a sharp rise in the beginning of 2023. This coincides with the launch of ChatGPT, which was released to the public on November 30, 2022. This was one of the first indicators of AI productivity and credibility to the public.

ChatGPT was the fastest to reach one million users.

ChatGPT reached 100 million monthly active users in January 2023. And as of June 2024, ChatGPT has over 600 million monthly active users. Clearly there’s something valuable there.

So now the question is, who is benefitting from AI?

AI revenues are quite low at this point. Outside of ChatGPT and a few other large language models, there aren’t many adoptions of AI just yet. We are still at the infrastructure point in AI’s development.

Phase I – AI Infrastructure

The current beneficiaries of AI are the manufacturers of GPUs, datacenters, power, and cloud. AI’s infrastructure is being built. GPUs are the brains or processing power behind the large language models. Datacenters are the massive mainframes where all the data and processing are being stored. Hyperscalers such as Microsoft, Meta, and Amazon’s AWS, are the ones purchasing the GPUs. This is where most of the money is currently being spent (aka capital expenditures). Super Micro and Dell are examples hardware server companies providing the backbone that’s housed within the hyperscaler data centers. And in order to power and cool these servers and data centers, companies like Vertiv Holdings have also seen growth in its revenues. See below for images of Meta and Google data centers.

One of Meta’s data center campuses

Rows of servers housed within a data center

More servers

Phase II – Edge AI

After the establishment of basic infrastructure to process data, compute, and infer, the next stage would be to mobilize AI. Instead of relying on a centralized, cloud-base system, local “edge-based” AI will offer several key advantages. Smartphones will enable people to carry AI in a miniaturized PC version of AI. There will be personalized user experiences where each AI smartphone will provide adaptive interfaces based on user behavior and context. Predictive text, app suggestions, messaging are all likely. AI generated smart replies for texts and emails with accurate context may be the norm.

Phase III – AI Applications

AI personal assistants will be accurate. Amazon already announced it plans to give Alexa an AI overhaul. Microsoft Copilot is an AI-powered tool that works with Microsoft 365 applications. SoundHound, a voice AI and speech recognition company, has already partnered with several drive-thru restaurants for AI powered drive-thru voice ordering. Many customer service jobs will be replaced with accurate AI models that sound and feel realistic.

Phase IV – Physical AI Products

AI Robots (Tesla’s Optimus, Boston Dynamics’s humanoids robots) will become more prevalent. Restaurants are currently replacing some labor with AI machines. For example, Miso Robotics provides AI-powered fryer robots and Sweetgreen is rolling out robotic salad creation kitchens. To extend this further, self-driving cars, robotic surgery, AI home appliances will be more prevalent.

Market Implications

As this relates to markets… Are we in a bubble? Seems not to be the case. Although individual companies that show promise for future AI applications may flop, the overall market seems to hold out for AI earnings. In Nvidia’s case, the Company trades at ~40x earnings, which is the same multiple as its last 5-year average. Instead of overvaluation (think 1999’s internet bubble valuations), it seems that there’s a pull forward of NVDA earnings. Rather than a market bubble, I believe AI will continue to drive the market. We are only in the early innings of this AI revolution and super cycle. There are and will be many opportunities to play the AI space in public markets.