If you flip the AI narrative upside down, the headline becomes “The S&P493 has increased by 1% year-to-date”. Only 7 stocks have driven the entire 9%+ performance of the US market so far this year.

One of them particularly stands out: NVIDIA, which produces GPU (Graphics Processing Units) aka graphic cards. To respond to the growing need for processing power, NVDIA is offering various combination of GPUs packaged as “Data Centers”.

According to the company’s 1Q24 results “Data Center revenue was a record, up 14% from a year ago, led by growing demand for generative AI and large language models using GPUs.” But the line that reached popular fame was “Revenue is expected to be $11.00 billion [in 2Q24]”, which blew the $7.3bn revenue expectation off the water.

Why is it so famous with AI? In short, NVIDIA’s API allows programmers to utilize the high number of parallelized processing cores present in the GPU hardware for use in machine learning algorithms. Yes I know, it is a mouthful.

If like me you reason like a 4-year old, see NVIDIA as the conductor of a massive line of singing GPUs in a warehouse relentlessly browsing data 24/7 to compute correspondence between vast troves of data.

NVIDIA is at the confluence of buzzy trends: it is US-based, it serves the promising AI use case, and it enables the transformation of companies to become software-centered.

It receives vast investment amounts, notably from big tech companies. Apple, Amazon, Alphabet, Meta and Microsoft invested close to $400bn in Capex and R&D in 2022.

For these reasons, NVIDIA has enjoyed an eye-watering rise of 352% over the past 3 years.

As an aside, note that NVIDIA’s sales in FY23 (ending Jan23) came from Taiwan for 26% and China for 21%. So much for a “US-based” success story, which carries a massive geopolitical risk. Taiwan Semiconductor Manufacturing (TSM) is a key supplier of NVDIA’s wafers, and is having a renewed bout of interest, rising ~23% in May23, despite being in the eye of the US-China cyclone.

Having said that, NVIDIA is probably the case in point of the AI rush of the moment, which has blunted the interest for any other sector. Is that a right call?

AI is no less than a buzz word that encapsulates the growing reality that software is eating the world. Take automotive, it represented a 114% increase in Nvidia sales in 1Q24 for “self-driving platforms and AI cockpit solutions”. In a radical transformation, automotive companies are front-loading the heavy cost of a software-led infrastructure. Hard to tell which one of BMW, Ford or Renault would be rivaling with Tesla and BYD in the years to come. But one thing is for sure, they too will undergo their AI revolution. AI will go viral.

Like in any gold rush, those who sell pickaxes and pans are the surest winners, and the market has been favoring those who ostensibly invested more in AI. Now, you still have to reconciliate the massive valuation discrepancy between the 7 leading companies of the software movement, and the 493 others. Something’s got to give.

In my view, the trade is to rebalance out of the top 7. The AI trade is young and certainly very promising, but with risk-free rates where they are, the equity risk premium looks very very skinny from a historic viewpoint.

In layman’s terms, you don’t get much more bang for your buck at this level and it is typically in these circumstances that markets start to turn risk-off.

I would keep an AI-ye on this if I were you.