Google Ad

Nvidia Analyst Day was overshadowed by the selloff in $NVDA of 7.5% and $TSLA from self driving car accidents. The company did announce the scale and end-to-end reach of the AI platform and gaming and growing datacenter market.

Nvidia AD PC

Analyst Day Highlights

  • Double memory now for V100 with new DGX-2 appliance with 2x the number of V100s plus new NVLink switches (showed 10x+ better performance vs. DGX-1).
  • Expect a $50 billion TAM (vs $30B) data center market which TensorRT4 with Tensorflow integration helps penetrate.
  • RTX (real time ray tracing) which offers material improvements to game and pro viz developers.
  • Estimate a $60 Billion LT TAM as every vehicle should be autonomous by 2035, highlighting their simulation solutions.
  • Volta chosen by all major server OEM/ODMs and all major public clouds
  • 2,000 customers adopt NVIDIA inference platform
  • 86 of Top 500 Supercomputers NVIDIA Tesla Accelerated
  • Top 15 HPC Apps Accelerated; 550 accelerated apps
  • CUDA developers up 75% YoY to 770K

Nvidia AD Cloud

The punters were expecting Volta for gaming to be announced, but it did not materialize given the continued success of Pascal boosted by recent supply constraints. Analysts now believe  this will launch mid-year, which is a further positive catalyst in case the crypto downturn in price affect interest there. Note that $AMD does not have this yet either leaving $NVDA remains in a strong position in AI.

Nvidia AD Operating Expansion

Barclays issued a note after the analyst day, key segments and their comments are below:

Barclays on Gaming:

Core gaming trends remain strong helped by recent titles PUBG and Fortnite on top of the continued rise of e-sports. NVDA gained share this year (per Steam and NewZoo) vs. AMD which should be boosted by Volta for consumer (we expect mid-year) with plenty of runway as only 30% of the installed base is on Pascal.


 Nvidia AD Gaming

Barclays on Datacenter:

The appetite for GPU acceleration continues to grow as the company believes the TAM is now $50B (from $30B last year), driven by larger inference but also just farther reaching AI apps. Specifically, NVDA is making inroads into new industries such as healthcare (Clara medical imaging) and manufacturing (Isaac for robotics). The generational improvements to V100 and DGX-2 only help access these opportunities.

Nvidia AD Datacenter

Barclays on Automotive:

NVDA believes every vehicle will eventually have some level of autonomous and sees a $60B TAM by 2035 ($40B from self-driving cars and $20B from robo-taxis). Xavier just began sampling and seeks to simplify the multiple ECUs in an ADAS car today into one. They also announced Orin, which puts the performance of Pegasus into just 2 chips from 4, and DRIVE SIM to simulate driving miles more safely. 



Nvidia AD AV


Source: NVidia, Barclays

Log in to comment
Discuss this article in the forums (5 replies).