Did the AI Bubble Just Pop? When I was working on Wall Street, I had a friend who managed a hedge fund and asked me to show him pretty much all tech companies – with one exception. Companies in China. He wanted nothing to do with companies there because, as he put it, “China = fraud.” That is an oversimplification, and at the time he was referring more to questionable accounting practices. But I did think about him today as news of a new AI model out of China sent investors into a tizzy when it comes to tech stocks and AI stocks. When you went to bed last night, you probably didn’t think DeepSeek would be the buzzword of the day. But it became that as word spread of this new AI model from a Chinese startup company that its creators say outperformed OpenAI’s ChatGPT and at a fraction of the cost to develop. To get back to my hedge fund friend, I think we do need to be skeptical about stories out of China. It doesn’t negate the fact that there are amazing companies in China, but it’s just true. We’ve seen misinformation from China before, and we’ll see it again. So, first things first, we don’t know yet if these performance and cost claims about DeepSeek are 100 percent accurate or not. I also have questions around the timing of the announcement. As you probably heard, last week’s unveiling of the Stargate Project threw $500 billion worth of gas on the already blazing AI space. The venture was announced at the White House last Wednesday, with OpenAI, Softbank, Oracle, and MGX joining forces to build and develop AI technology. The goal was to make America the AI leader, and DeepSeek quickly followed. In addition, President Trump has said he will impose trade tariffs on China, and this may have been a bit of a shot across the bow. If you want to hit America in the wallet, knocking down some of the world’s biggest-gaining stocks like Nvidia (NVDA) – and the whole “Magnificent Seven” crew – is not a bad way to do it. And speaking of knocking down stocks, here’s another twist to the story. The founder of DeepSeek, Liang Wenfeng, is also the founder of a quantitative hedge fund. (Gotta watch those quant guys!) He made a bunch of money, and assets under management for his hedge funds reportedly now exceed $8 billion. In China, it is illegal to short stocks. You short a stock when you expect it to decrease in value. In the U.S., it is not illegal to short stocks. So, what if you shorted all these big gainers and an announcement came out that hit them hard? You would make a lot of money. I am in no way saying that’s what happened here. I am just pointing out some of the connections. We see this here in the U.S. when often disreputable research firms short a stock, publish negative research on it, and profit from the resulting sell-off. I guarantee that some of today’s selling is high-frequency trades and algorithmic trading. They come to life in these big down days. Some of it is also forced selling for those investing in margin, which just accelerates the downdraft – and makes it even more likely that sellers are getting lower prices than they could if they waited. Potential Ramifications Maybe DeepSeek will be all it’s cracked up to be, but maybe it won’t. Even if true, this will not be an overnight change where every company that was once ordering Nvidia’s AI chips suddenly turns to cheaper – and inferior – alternatives. In other words, the entire landscape of AI and how it will be powered and by whose chips doesn’t change in the blink of an eye. Selling on this news is doing so based on what might happen, not what is happening. It’s not based on changing data. It’s based on the possibility that the data might change. Nvidia (NVDA) fell nearly 17% today, but it has a tremendous history of massive sales and earnings growth – and is forecasting even more sales and earnings growth. Maybe that will change over time, but I don't subscribe to the idea of a sudden massive sea change and seismic shift. Even if DeepSeek turns out to be all it’s cracked up to be, it will impact different stocks in different ways. The huge companies that have been big AI winners will face more pressure, but it will also be easier for smaller companies to join the AI race when the entry costs were previously prohibitive. This would be a gradual AI rotation, not the end of the AI as we know it or the popping of an AI bubble. The revolutionary technology is still there, and so is the growth. There are many legitimate questions surrounding DeepSeek, and we will get the answers over time. This early trading sure looks like an overreaction – which we’ve seen countless times before – and the future still looks bright… and profitable. At the end of the day, I follow the data. That’s the Quantum Edge that drives outperformance. And one thing I do know – the data will guide us to the most profitable decisions on current and future opportunities. Talk soon, 
Jason Bodner
Editor, Jason Bodner’s Power Trends | 
No comments:
Post a Comment