What Do I Think of NVIDIA at These Levels and Gen AI?
I believe that NVDA has the best Semiconductor CEO, perhaps the best CEO in all of broadly defined Technology given NVIDIA’s ability to deliver innovation year after year. I also believe that NVDA shares are overvalued.
NVIDIA’s Blackwell release ought to provide air cover for a while given order flow from MSFT, META, xAI, GOOG, AMZN, OpenAI, Anthropic and others.
However, I do not see MSFT, META, xAI, GOOG, AMZN, OpenAI or Anthropic generating Gen AI-related Revenue that is commensurate with investment levels or valuations.
Take Microsoft (MSFT) for example – Copilot is a failure as far as I am concerned.
Google (GOOGL) has a couple of cool Gen AI products with NotebookLM and Gemini’s image generation capability.
Adobe (ADBE) has done some cool stuff with image and video generation.
NVIDIA (NVDA) has leveraged Machine Learning as part of its video game development process for years (background rendering).
There are a number of use cases across Creative job functions. In addition, AI Agents are beginning to be deployed for QA and bits of code writing, although I’ve yet to see AI Agents deployed for building commercial grade, enterprise software applications at scale.
In short, few companies have been able to productize Gen AI at the application layer. Beyond Copilot, I can’t name another Gen AI product at scale.
It is only the companies at the infrastructure layer (chip and LLM companies) that have generated meaningful revenue, although those companies are overvalued as I mentioned at the top.