This article was originally published on Fool.com. All figures quoted in US dollars unless otherwise stated.
Key Points
- Nvidia stock dropped 12.6% in November as investors weighed growing concerns of an AI bubble.
- Google's new frontier model was trained on its own chips, challenging the idea that companies need to shell out for Nvidia's pricey GPUs.
November was a tough month for Nvidia (NASDAQ: NVDA) investors. Despite reporting another blowout quarter on November 19th, the artificial intelligence (AI) juggernaut's stock fell 12.6% from the closing bell on Oct. 31 through the end of trading on Nov. 28.
November was marked by growing fear of an AI bubble with Nvidia at its very heart. While Nvidia itself is making incredible amounts of money selling picks and shovels to AI gold miners, investors are wondering just how much gold there really is and if it's enough to justify buying Nvidia's extremely expensive equipment -- at least at current levels.
While that concern has been on the mind of shareholders for some time, it peaked this month after Alphabet's Google released its latest AI model, Gemini 3. The large language model (LLM) was widely received as an improvement on the latest models from both OpenAI and Anthropic, but critically, the model was trained with Google's own chips, not Nvidia's.
Google's Gemini 3 raises questions about Nvidia's AI chip dominance
These chips -- called TPUs -- are specifically designed and optimized for the kinds of operations that Google uses to train its LLMs, making them both cheaper to produce and cheaper to run. That directly challenges the primary narrative driving Nvidia's success, that its chips are so much more advanced than the competition that it is worth paying a hefty premium for them.
And while it is still true that Nvidia's GPUs are far more powerful and flexible than Google's TPUs, given Gemini 3 is so capable, it's reasonable to wonder why Google -- or any of the other massive hyperscalers with the means to develop their own chips -- wouldn't move toward relying primarily on their own chips.
Record Q3 earnings couldn't stop the November sell-off
Despite this fear, however, Nvidia's Q3 earnings showed no signs that its orders are slowing down -- quite the opposite. It once again delivered massive top- and bottom-line growth, both year-over-year and quarter over quarter, with gross margins that one would expect from a Software as a Service (SaaS) company, not a chipmaker.
Nvidia stock has regained its footing, rising roughly 4.3% so far in December. While peak fear seems to have subsided, I think investors should still be cautious. It is hard to argue with the eye-popping numbers Nvidia is delivering, but I do question how long its place at the center of the AI boom -- and the AI boom itself -- can continue.
There has been limited evidence of AI's impact so far in the real economy, and while there is certainly a lot of promise in the technology, there comes a point where that won't be enough for shareholders to stomach the hundreds of billions being invested.
Nvidia's valuation is entirely built on its growth trajectory continuing apace. In the near term, there are no real signs that will falter. Long term, however, I have my doubts.
This article was originally published on Fool.com. All figures quoted in US dollars unless otherwise stated.
