Nvidia, its parabolic growth and making sense of it

Never have so many depended so much on so few or rather on just one company – Nvidia! Goldman Sachs’ trading desk termed it as ‘the most important stock on planet Earth’ before its earnings release last Wednesday. Options pricing data implied expectations of greater than 11 per cent or around $200 billion of market cap move in the stock on either direction and linked to it was the fate of the entire AI rally that has been driving global markets since May of last year.

It would be worth recollecting that it was not the release of ChatGPT in CY2022, but Nvidia’s quarterly results reported in May 2023 that ignited the massive AI rally, which has continued unabated till date.

 As investors and traders awaited with bated breath, Nvidia did what it has been doing quarter after quarter in recent times —  delivering a beat and raise quarter, triggering a fresh round of trillions of dollars of market cap getting added across the world in a single day, sending major indices in Japan, India, Germany and the US to all-time highs last Thursday.

Nvidia on its own added $277 billion, creating a record for the largest market-cap addition by a company in a single day, beating the record Meta Platforms had created only a couple of weeks back in early February. What is driving this frenzy around Nvidia — the latest company to cross the exclusive $2-trillion market cap (equal to 45 per cent of the entire market cap of India) level? Five years ago, its market cap was $90 billion, while 10 years ago, it was $9 billion, which indicates how outstanding its growth has been in recent years.

Mind-blowing numbers

To begin with, the company’s fundamentals are so strong, it is hard to identify when in history a company has had it so good. As we explained in our Big Story in bl.portfolio edition dated February 11, 2024, currently, it is estimated that Nvidia has anywhere between 70 and 90 per cent share in the GPU chips used for AI training and inferencing. Although, its market share will decline over the medium term, for now its moats are very strong, given huge first-mover advantage in making AI chips, and more importantly because of its proprietary computing platform CUDA that runs on all its AI chips.

With customers building their AI training and inferencing models on Nvidia hardware and software, while competitors eventually will come out with competing GPUs, they may still depend on Nvidia for the software part as customers are spending millions of dollars to build their applications on this, making migration to any competing software ecosystem challenging.

– ‘It will take half a decade to put a real dent in Nvidia’s monopoly’. -says an analyst at a foreign brokerage called Susquehanna. This appears to be a view shared by many other analysts as well, that is driving the bull thesis on Nvidia.

According to Nvidia, coming to this stage in chip development was a 10-year process starting with re-orienting its approach to chip design a decade back.

These fundamentals are well reflected in its FY24 (ended January) and Q4 results released last week. In Q4, it reported a revenue of $22.1 billion, a 265 per cent increase versus Q4 FY23. More specifically, its Data Center (AI GPUs) segment revenue zoomed a staggering 409 per cent to $18.4 billion, accounting for 83 per cent of consolidated revenue. Operating profit was at $13.65 billion and net profit at $12.28 billion, up by 983 and 765 per cent respectively y-o-y. It’s next to impossible to see such high growth for companies with revenue and profit in the billions. Equally rare are the profit margins reported by Nvidia, with an operating profit margin at 61.7 per cent and net profit margin at 55.5 per cent!

Post an unprecedented full year FY24 (ended January) in which revenue grew 126 per cent to $60.9 billion and net profit grew 582 per cent to $29.7 billion, consensus estimates now project further revenue growth of 74 per cent and earnings growth of 97 per cent for the current FY.

To put it simply, without any exaggeration, these are mind-blowing numbers. After all these, the stock trades at one-year forward PE of 33.7 times. Appears inexpensive? Well, here are the two sides of the coin.

Cyclicality

Semiconductors are a notoriously cyclical industry, repeatedly going through boom and bust. As recently as CY2020, there were massive chip shortages, as smartphones, PCs, tablets and gaming products saw a huge spurt in demand during lockdowns/work from home. This turned into a supply glut by the year 2022. In fact, Nvidia itself, faced the brunt of this glut when its shares crashed by 65 per cent in 2022, and it was the AI wave that brought about a paradigm shift.      

The bears point to this and believe the Nvidia ‘hype’ might dissipate. They also point out how Intel was in Nvidia’s current position during dotcom boom, a time when Intel dominated the personal computing chip business (a huge beneficiary of the nascent internet boom then), but the stock has never since crossed its peak levels reached in the year 2000.

However, as things stand now, in the case of Nvidia, the bulls may have a stronger case than the bears. Nvidia’s recent growth far exceeds anything Intel delivered then. As compared to Nvidia’s growth in just one year mentioned above, Intel took five years to double its revenue between 1995 and 2000. Its net income margin was at 35 per cent, compared to Nvidia’s 55 per cent. Nvidia’s superior margins reflect stronger moat, and its superior growth reflects stronger demand. Nvidia’s CEO Jensen Huang highlights a robust demand scenario driven by existing base of data centres across the world transitioning to GPU based computing, and a demand from new AI focussed data centres.  

More than demand and competition, the larger debate is, what is the right valuation to buy Nvidia. Investors can keep it on their radar and consider it when broader market-driven corrections present better entry points.

COMMENT NOW