"Sell To" vs "Sell Through" - NVDA, TSLA, and more

One of the senior partners on the early stage tech company investment team I was part of for many years used to have a mantra about “sell to” vs. “sell through”. “Sell to” is when you sell your product to your customer. “Sell through” is when your customer uses your product to satisfy the needs of their own customers. You can have a lot of “sell to”, but unless that somehow translates into creating lots of value for your customers – for example, by enabling them to sell a lot more to their own customers – you’re not going to have much room for your sales to stay strong, let alone grow.

For those of us eyeing the heights to which NVDA has been climbing, this is a hugely relevant topic. NVDA currently has more demand than they can satisfy with their current production – but much of that demand would go away quite suddenly if the buyers weren’t able to translate their investments in NVDA chips and systems into increased revenue of their own. “Sell to” vs “sell through”.

With that in mind, here are three bits of interesting news on the “sell through” front:

  1. Tesla has long been working on enabling their cars to drive themselves, a program they call Full Self Driving (FSD). This weekend, I listened to the most recent BG2Pod (thanks to the person who pointed me to the podcast, I’ve lost track of who that was, but it’s highly recommended!) where they talk about the new version of Tesla FSD. Tesla threw out all their previous work on FSD and moved to a purely AI/neural net version, trained entirely on video from highly curated sources. They showed the AI what excellent drivers do, and the AI learned from that material how to be a good driver. Brilliant!

  2. AI has long been a part of synthetic biology, but it’s now taking an even more important role, as discussed on this recent podcast where Eric Topol (a prominent medical innovator) interviews Daphne Koller, founder/CEO of insitro about where AI is currently in synthetic biology and where it’s going to take us. Carl Zimmer wrote a piece the other day in the NY Times that also lays out some of progress happening with AI in life sciences. Clearly there’s a lot more room to run here!

  3. Companies like ServiceNow (NOW) are reporting that their customers are shifting their strategies to be better positioned to take advantage of the potential of generative AI:

What’s really happening, and I can say this after 186 CEO meetings in the last 6 months, the CEOs are now getting very involved with the gen AI revolution. They realize there has to be architectural adjustments to their environment and the manner in which they manage their data and the platforms they’re beholden to, to actually take advantage of gen AI.

Bill McDermott on the latest earnings call

I’m going to continue to look for evidence that the bloom is off the AI rose, focusing on places where the “sell through” should be evident, looking for clues that people are having second thoughts. Meantime, I’m a happy holder of NVDA, NOW, and, as of today, TSLA as well.



Great notes on sell to and sell thru.

This is Jensen’s recent speech at Stanford. 1) at 9:20 he talked about not only create the technology but also create the market. And that leads to NVDA’s early explortion and entrance into DL, human oriented robots, autonomous driving, etc. 2) 46:52 how he visions his technology could advance/assist other areas to succeed, for example, computational biology
In short, it sounds that NVDA is not only creating new sell to technology, but also sell thru markets for further development


Yes, NVDA has been enabling pioneers in many different markets for a long, long time. That’s a big part of what has made it a “30 year overnight success story”: supporting the inventors of many new industries and new applications.

To support the growth of the company from the current lofty heights, we’re going to need to see multiple mass markets develop. Jensen will tell you that essentially all computing markets will become AI-driven, GPU-driven markets. I don’t have the vantage point he has to see into all those different markets, so I’m looking for all the clues I can find to see how right he is.



I felt like Jensen Huang reallY showed a different side to himself and Nvidia, in this Stanford Interview. It was by far my favorite so far.

My note I took the day I watched it…
3/6Jensen Interview@Standford

How Nvidia is Absolutely DeepTech, why Humanoids will happen ‘very soon’…spoiler, Jensen is better than Elon keepimg his company employees focused on Mission!

Never sell!

28% Nvidia, having doubled my initial 14% position DCA @$447/share.