- Aurify
- Posts
- The envy of processors - Today's stock deep dive
The envy of processors - Today's stock deep dive
Every stock has a story. Here’s one we are interested in. The year is 1992. Jurassic Park comes out next year, so do popular games like Street Fighter II and Disney’s Aladdin. Chris Malachowsky and Curtis Priem want to talk to Jensen Huang about a pitch. Jensen takes them to a his go to restaurant -Dennys. He used to work there while he was younger and its one his favorites. Chris and Curtis are talented engineers at Sun microsystems while Jensen has just finished his masters from Stanford and is an electrical engineer. He is a chip design product manager at LSI who make custom chips for computers.
They talk to Jensen about the inability of Personal computers to make high quality 3D graphics. They discuss the growing gaming industry and the limitations of a PC to run games. Specifically they discuss a customized graphics card. If they could build a custom graphics chip it would enable gaming on a PC. This is a new idea in 1992. Jensen is sold on it and talks to Will Corrigan, the CEO of his company LSI, about his idea. Wilf Corrigan, tells him that he can keep his job as backup and recommends him to talk to Don Valentine, founder of popular VC firm Sequoia. Jensen botches up the pitch at Sequoia. LSI being is one of the early success stories for Sequoia. Don tells Jensen since he trusts Wilf’s judgement and agrees to fund their company. Also if Jensen loses his money, he will kill him - a classic Don line. ‘Invidia’ Latin for envy becomes a real company. Jensen and team decide to drop the ‘I’, calling it NVIDIA.
The reason graphics card are more powerful than CPUs are because they are able to parallelize tasks. CPUs are good for performing a much wider variety of tasks well, while GPUs can perform some tasks much better.

Microsoft is not unaware of the gaming demands in the early 90s and comes up with their own gaming system - Direct3D. They use triangles as a building block for their 3d graphic creation. Nvidia uses polygons to render graphics and also manages to secure a deal with SEGA, a major gaming company at the time. The industry comes around to adopting Direct3d and NVIDIA’s deal with SEGA falls through. NVIDIA is now losing out on time and money. Jensen makes a tough call and changes their complete gaming architecture to use triangles as their building blocks. With a scrappy team of 35 engineers, they set out to out-engineer and out-ship more than 90 competitors out there making graphic cards. They cut out extensive hardware testing and seat a guy to do frame rendering in front of a computer to compress their production launch time from 2 years to 6 months! Backs against a wall and without a tried and tested manufacturer like TSMC they launch the RIVA 128.

NVIDIA RIVA 128
It’s buggy and doesn’t have a third of its processing power it states but it’s a beast. Developers love it. They sell 1 million units in 4 months. Jensen established 2 key working principles at his company from that day. Intellectual honesty and performance, these are the metrics they would build upon in the coming years.
NVIDIA, today is a 10000 pound Gorilla in the world of computer infrastructure, and has applications in data centers, self driving cars, medical research and AI. It owns 91% of the world’s GPU enterprise market share compared to it’s competitor AMD’s 8.5%.
Here’s a summary of their last 2 years -

We also listed out some of pro/cons of the stock. Take a peek.
Potential upside:
The global GPU market grew at a compound annual growth rate (CAGR) of 18.2% between 2010 and 2020. The market was valued at $14.7 billion in 2010 and reached $76.8 billion in 2020. The gaming industry accounted for 42%, AI and ML came at 201% and 18%. The market for GPUs is expected to grow at a CAGR of 25.4% from 2022 to 2032. The gaming market itself is expected to grow at a CAGR of 10.2% as per one report. The market is being driven by an increasing demand for GPUs in applications like gaming, artificial intelligence (AI), machine learning (ML), and data center computing. This puts NVIDIA in a unique position to establish themselves in the long run and increase their revenue through different applications.
Post ChatGPT era, the global AI market today is estimated to be 120 billion dollars and is expected to hit ~1590 billion dollars by 2030.
The bull case for NVIDIA is that the company is a leader in a rapidly growing market and is likely to remain so for the next 3-4 years.
There is also immense potential for growth in the cloud market and NVIDIA which it dominates already.
The downside:
The stock trades at over 150 times it’s earnings as of 10th April 2023 and remains overvalued.
Increased competition from the likes of Google, AMD and Intel in the coming years might make it tough for NVIDIA to further increase its market share. Google recently claimed that its TPU (Tensor Processing Unit which is designed more specifically for ML and AI tasks) was able to surpass NVIDIA A100’s performance benchmarks.
Rising inflation and interest rates may result in cutbacks in consumer and corporate spending affecting margins.
Continued supply chain disruptions resulting from the war in Ukraine and a potential conflict of Taiwan with China can present substantial headwinds for the NVIDIA which contracts Taiwan Semiconductor Manufacturing Company (TSMC) for its chip manufacturing.
Key fundamentals as of April 2023:

To say the least, NVIDIA is a company with an amazing origin story. With fundamentals which just keep on giving.