They say that a week is a long time in politics. But with Nvidia (NASDAQ: NVDA) stock shedding the largest daily amount of market cap in US stock market history, one day must have felt like a lifetime for its shareholders.
A total loss of $1trn in market cap across the Magnificent 7 highlights the dangers of extreme market concentration in just a handful of stocks. The question I am now asking myself is, are we are at the beginning of the unravelling of the biggest bubble in stock market history?
Large language models
I don’t claim to be an expert in large language models but I have for some time become deeply concerned about the path the tech industry finds itself on.
Although there is no way of knowing for sure how much the hyperscalers have invested in Nvidia chips, the accounts of Alphabet, Meta, and Microsoft point to it being in the tens of billions of dollars. And what have they got to show for it?
Without trying to sound too cynical, my answer is not a lot.
Every time I use ChatGPT, I am less than impressed with its results. I have yet to talk to anyone who would rely on the answers any generative AI model provides without first fact checking. It’s little wonder that all these models come with a disclaimer.
What about the thorny issue of hallucinations? This happens when a model presents a falsehood as categorically true. What business will want to rely on such flawed technology, particularly when it comes to mission-critical apps?
Lessons from history
It’s easy to discount my criticisms. After all, generative AI is a new technology. Surely, a solution to these problems will come with time. Maybe it will, but that fundamentally misses the point.
The entire industry is selling this technology as the future of AI. Whether DeepSeek succeeds is neither here nor there — such a collective mindset across Silicon Valley is inherently dangerous. I would even go as far as to say that such groupthink poses an existential threat.
History is littered with examples of companies that were at the forefront of the advancement of a new technology, and yet when the race was run were pipped at the post.
One of the most instructive was the 20-year long pioneering innovation battle to produce a commercialised home videotape. The winner there didn’t emerge until JVC produced the VHS standard.
Buy the dip?
My point in using the JVC example, is an attempt to highlight that the battle for AI supremacy will not be run over 100 metres but over a marathon. But too many investors today believe that the race has run, and the riches and the spoils are there to be collected.
As I write this article, on 28 January, Nvidia has recovered some of its loses. This is to be expected. Many would-be investors will view its share price weakness as a buying opportunity. I am not so sure.
The stock is almost certainly heading for a period of heightened volatility. But I still think that far too many investors are totally oblivious to the risks here. ‘Buy the dip’ is so ingrained in people’s psyche, that only an all-out crash will change such a mindset. I certainly won’t be touching the stock.
