Why I am not writing a Deep Dive this month

I was working on this month’s Deep Dive on Nvidia. I was excited to finally study the largest company in the world with focus and patience it deserves, but after spending the last three weeks on the business, I have decided to give up on writing the Deep Dive. I still feel too much gaps in knowledge to be comfortable writing a Deep Dive on the company.

While this is unfortunate and I apologize to you for the inconvenience, I think it can be instructive to understand how I came to realize this.

I listened to all the Acquired series on Nvidia (Part 1, 2, 3, interview with Jensen Huang). Huang’s very recent interview with Joe Rogan was also helpful to appreciate some of the near death moments Nvidia faced. I was particularly touched by Huang’s upbringing and I came away quite appreciative of his innate ability to evolve and adapt to almost any situation Huang (and Nvidia) may find themselves in. I listened to a bunch of episodes (here, and here) on Founders podcast to get to know Nvidia and Huang’s story better.

After having a reasonable grasp on the history, I started going through their SEC filings. I read Huang’s all the shareholder letters. There are too many good bits to quote, but let me share a couple that stood out.

“Foreseeing the importance of energy efficiency, we set out half a decade ago to build high-end parallel and mobile processors. These groundbreaking initiatives sought to address two of our fundamental convictions: that power will limit the number of computers in large data centers, and that energy efficiency will define our experience with mobile devices”

“There is already well over 1 exabyte of images and videos in the cloud, more than 100,000 times the books in the Library of Congress. Hundreds of millions of photos are uploaded daily. And the amount of data created will rise 50-fold this decade, according to IDC. Our GPUs can make a real contribution to processing this deluge — enabling computers to learn how to help us search a world filled with images. It is an opportunity that could potentially require millions of our GPUs.

Here’s a quiz: when do you think Huang wrote these sentences in his shareholder letters?

The first quote was taken from 2011 shareholder letter, and the latter from 2013. This was my reaction after reading these quotes...

I built a shell excel model for the company and the way their revenue and earnings ramped up, it is fair to say I have never seen anything like it. I wanted to see what some analysts were saying about Nvidia before some of these eyepopping numbers showed up in their financials.

I read a bunch of them, and I was pleasantly surprised how Scuttleblurb’s 2021 series on Nvidia (Part 1 and Part 2) touched on so many key areas that became part of the conversation in the next few years (The Bitter Lesson, Jevon’s Paradox, competition from ASICs, detailed explanation of CUDA moat etc.). I let Scuttleblurb know how much I enjoyed reading his pieces, and fortunately, he took paywall off his Nvidia pieces later. So, you can all go and read it yourself. I also want to highlight this piece by Fabricated Knowledge back in 2020 who was admirably early in outlining the “future”.

To get a sense of the current competitive landscape of Nvidia, I went through a bunch of work done by Semianalysis (see their work on AMD, TPU, Trainium, for example). Gavin Baker’s recent podcast on ILtB was also helpful in understanding the rivalry between Google and Nvidia.

I probably mentioned only ~30% of the content I went through during my research process, but you can sense there is a deluge of content on Nvidia. After consuming a good chunk of it and feeling reasonably good about my understanding, I started writing the Deep Dive myself. After making 20% progress on writing the Deep Dive, I only started appreciating the gaps of my knowledge when I got to more meaty parts of the Deep Dive.

Frankly speaking, if I didn’t try to write a Deep Dive, it would have been lot easier to convince myself that I understand the business reasonably well. But writing is a forcing function in making you realize your gaps in understanding. AI didn’t quite help because AI’s answers were only making it more apparent to me that the gaps may be wider than I appreciated. It also didn’t help that there is some news flow about Nvidia that is happening on a daily basis which further eroded my confidence in my ability to write a Deep Dive that can maintain its relevance even a year form now.

So, after writing a Deep Dive every month since September 2020, I came to this unfortunate decision to pause writing the Deep Dive for this month.

However, since I have an “okay-ish” foundation of knowledge on Nvidia now, it will be marginally easier for me to follow the company going forward. I would like to follow and cover their quarterly earnings, and major events such as GTC more closely for a couple of years before making another attempt of writing a Deep Dive on the company.

While everyone is talking about semis, the whole experience was a stark reminder that the technical nature of the industry remains quite challenging for not only to get up to speed but to keep up with the industry. I am not giving up on covering semis in the future, but just appreciating the patience it will require to get there.

It’s not ideal, and I wish I had a better update for you. Thank you for your understanding.

Disclaimer: All posts on “MBI Deep Dives” are for informational purposes only. This is NOT a recommendation to buy or sell securities discussed. Please do your own work before investing your money.