,
23 tweets,
3 min read
Read on Twitter
1) $NVDA. Most important variables to me as a long term investor here are/were 1) whether publishers would adopt DLSS and 2) whether TensorFlow would emerge as the “one framework to rule them all” in AI.
2) Answers appear to be 1) yes after last nights CES event and 2) no given the growth of Pytorch, which substantially increase the attractiveness of $NVDA.
3) DLSS adoption was/is critical because ray tracing is only viable in the next few years if publishers adopt DLSS. And if they adopt DLSS, then $NVDA is on its way to having its deepest competitive advantage ever in its core graphics market.
4) DLSS is important because without it, very few gamers will use ray tracing in the next few years and $NVDA just made an enormous bet on ray tracing.
5) Ray tracing makes games more beautiful – and is the future of graphics – but beautiful graphics are a secondary consideration to most buyers of leading edge GPUs.
6) Most gamers are buying leading edge GPUs to have a faster frame rate in competitive FPS games, which gives them a competitive advantage and improves their K/D. Essentially, a faster frame rate means you see your opponent before they see you.
7) Ex DLSS support, games only run at a ~35% faster FPS on the RTX 2080 Ti relative to the GTX 1080 Ti when prior new architectures offered a ~70ish increase in FPS. With DLSS, the RTX 2080 Ti runs games 70% faster. This meant the RTX series was not a good “value” without DLSS.
8) Reason is that $NVDA spent valuable silicon real estate in the RTX architecture on both ray tracing and DLSS at the expense of more traditional GPU silicon, which was an enormous bet they could afford to make given effectively zero competition from $AMD at the high end.
9) If developers did not adopt DLSS and raytracing, then $AMD would have been able to get back into high end GPU competition by focusing on traditional graphics technologies (TAA, etc.).
10) However, if developers adopt DLSS, $NVDA will have their biggest advantage over $AMD ever as it is a proprietary standard. Prior attempts at this by $NVDA (PhysX) were largely unsuccessful.
11) It also lets $NVDA have their cake and eat it too by putting the tensor cores to use in graphics.
12) Last night Nvidia announced that both Anthem and Battlefield V, as rumored, would support DLSS. I suspect this will result in most AAA games following suit over the next year.
13) This makes life very difficult for both $AMD and $INTC (on their second attempt at a discrete GPU following Knightsbridge) going forward.
14) Not nearly as important, but bringing Gsync to Freesync monitors eliminates $AMD’s only current competitive advantage (much cheaper monitors with adaptive sync to eliminate tearing). Icing on the cake.
15) As a sidenote, $AMD may work big from here, but outperformance will be driven by their CPU business. The fact that $INTC has fallen behind in process development for the first time in decades (ever?) is seismic.
16) Literally cannot believe that $AMD, $NVDA, $AAPL and $QCOM have a process advantage over $INTC via Samsung and $TSM. Why it makes sense for $QCOM to take another run at Windows.
17) On the AI side, Tensorflow’s dominance was becoming a significant risk to $NVDA as $GOOGL has never guaranteed backwards compatibility with Tensorflow to my knowledge.
18) Diversity in both frameworks and algorithms are a significant part of $NVDA’s competitive advantage in training (and data center inference to a lesser extent). Diversity in both makes life very tough on all the startups.
19) However, $GOOGL is not a startup and even though TPUs are inferior to Tesla’s (chip not car), $GOOGL does not have to pay $NVDA’s margins on top of the foundries which meant they could make TPU’s cost competitive with Tesla’s on Tensorflow both internally and externally.
20) And controlling both Tensorflow and TPU development gives $GOOGL significant advantages, especially if they ever broke backwards compatibility such that TensorFlow would only run on TPUs ($NVDA could engineer around but painfully).
21) Community would’ve revolted but was still a risk. Therefore, emergence of PyTorch as the preferred framework for research post $FB open sourcing was super powerful for $NVDA. $GOOGL TPU support for Pytorch effectively an admission of defeat.
22) Stock may continue going down near term due to crypto overhang, but the 3-5 year outlook is very powerful. At the center of graphics, VR, self-driving cars, AI and a free option on crypto ever recovering.
23) $NVDA large position for me. Feels good to end this on MJ’s number. Also very bullish on $GOOGL. Tensorflow dominating AI frameworks isn't at all critical to the equity story there.
