Here are 5 bear cases for Nvidia — and why each one is wrong

Remember Nvidia? I want to do the unthinkable: I will tell you about Nvidia coming from the mouth of countless short sellers who want to break into this company's shares and intellectual property capabilities. It's a daunting and somewhat siege situation, even for me, his chief supporter in the media. It's a brutal parade of fear, so steel yourself and — spoiler alert — I don't believe any of this will derail the story of the company now worth $4.4 trillion. The list is negative, malicious, anti-Jensen Huang, and contains a narrative that is truly opposed to the foundation's portfolio for your selfish purposes. Please wear your seat belts. Here are five criticisms beloved by Nvidia bears and short sellers, and why I believe they are flawed. Bear scenario No. 1: Nvidia is no longer the king of AI chips. There are now better competitors to Nvidia's chips. Advanced Micro Devices claims its most powerful chip, the MI450, is considerably better and cheaper than Nvidia's next-generation Vera Rubin. That's why OpenAI announced its massive deal with AMD two weeks after Nvidia's $100 billion deal, which was widely considered a lazy Susan deal; Nvidia gives OpenAI the money, and OpenAI gets the chips it needs. If Nvidia's deal was this hot, there would be no need for the AMD connection. Additionally, the AMD deal was handled at the highest level by CFO Sarah Friar, and she was the golden, grown-up manager of the process until she made "reversible" comments about the federal government's interest in the deals. The press was very brutal about this AMD deal and it created a lot of fear among the faithful. Although Amazon bills itself as a good friend and customer, it has developed its own chip, which it continues to develop, to distance itself from Nvidia. Google has a chip so good that others are clamoring for it, including its arch-nemesis Meta, which is tired of paying Nvidia so much for its chips. Google's chips are so good they will force Nvidia to cut prices and destroy its margins, which is why its shares are trading at about 25 times next year's earnings. Google's chips are a death star for Nvidia and Google, and can do no wrong in anyone's eyes, so much so that we'll have to deal with it too (and maybe Johnson & Johnson, but that's a topic for another post). Why it's wrong: Analysts estimate that Nvidia controls 70% to 95% of the AI chip market for training and deploying models like OpenAI's GPT. That's a pretty competitive moat. And for good reason: CEO Jensen Huang is the best there is; He is a true genius who can help you write software that can make everything faster and more accurate, creating platforms full of dazzling chips and software and taking artificial intelligence and accelerated computing to the next level. Nvidia's next-generation AI superchip platform, Vera Rubin, will help you "reason" much better than anything else out there, including AMD chips. None of these competitive threats compare to Vera Rubin, and while they're no Tinkertoys, they're merely substitutes for sillier tasks. Bear Case No. 2: Jensen is a good-brained guy who is not deeply involved with the US government. While Huang advocated for his company to be allowed to sell good-but-not-great chips to China to prevent our competitor from developing its own architecture, China went further and created a better product and no longer needed Nvidia. Why it's wrong: Huang has been telling the same story since early last year, when the Trump Administration tightened Biden-era rules: Keeping Nvidia out of the Chinese AI market on national security grounds would actually undermine U.S. tech leadership. That logic resonated with Trump, who began allowing sales of less advanced chips to China over the summer and said this month that Nvidia could ship its more advanced H200 AI chips to “approved customers” in China and elsewhere as long as the U.S. gets a 25% cut. As I have said from the beginning, Jensen is not only an engineer and architect, but also an incredible salesman and statesman. He knows how to tell the story. Bear case no. 3: Vera Rubin won't be ready on time. Despite stories to the contrary, this streak refuses to go away. The idea is that if Vera Rubin doesn't arrive on time (the company said it will in the second half of 2026) customers will have to go elsewhere, he says. The Vera Rubin platform, named after the famous astronomer who worked on dark matter, will replace Blackwell. Why it's wrong: Jensen has a 20-year game plan where he's thinking backwards, not two years ahead, so we should stop doubting and take the long view. We have to stay long because while there are many threats to Nvidia's dominance, they have never been quantified: Nvidia designs and TSMC makes better chips. Better chips equal better sales and better margins. The company's dominance was constantly under threat, and thus it became the largest company in the world. Bear case number 4. Nvidia's chips burn too hot and are unaffordable. The chips require a lot of power; This power cannot be achieved without multi-year installations because the actual power source, GE Vernova turbines, will not be available until 2030. Quantum computing is just around the corner, which will allow other chips to sneak in and still be faster and cheaper. Why it's wrong: The Information reported last year that Nvidia's AI Blackwell chips were overheating when connected to dedicated servers, prompting suppliers to change their server rack designs several times. At the time, I said the subsequent decline in the stock price was a buying opportunity. CEO Jensen Huang previously stated on CNBC that demand for chips is "insane." Despite its many bugs, The Information is now in Nvidia's registry of ills. There is no outcome adjustment in this new world, so a simple short selling assumption is too good to check. But what if it's wrong? Bear scenario No. 5: OpenAI won't get the money it needs. ChatGPT's maker urgently needs at least $100 billion, otherwise Nvidia's chips will sit in a warehouse, hurting the company's earnings. Why it's wrong: If OpenAI gets the trillion-dollar valuation and the $100 billion it needs to give Oracle to complete its data centers, they'll be filled with Nvidia chips, some from AMD. It can be done. On Friday, I described how the data center industry could regain its place in the market when this happens. There's a mad scramble to get Vera Rubin, while an even more advanced version is planned for release in 2028, all while getting excited about Richard Feynman. Oh yeah, no one has anything close to the next-gen chip. NVDA YTD mountain Nvidia Year to Date The one thing I didn't miss is China's threat to exert control over Taiwan and deny Nvidia access to the TSMC foundry it needs to produce chips. These stories are taken to the streets and swallowed by the financial media. Are you beleaguered like me? Why not sell the shares and buy Alphabet? Accept that you are on the wrong horse, and that it is emotionality and ignorance that led you to this horse. If you knew what was really going on you would sell, and your faith in Jensen comes from believing he's a great guy, but that's in the past. Either you manage to stay loyal to the guy or you don't want to anger him and you disconnect. I know my logic seems delusional. As I like to say, Jensen and Nvidia have the long knives in their hands; It's been that way ever since the stock broke the $2 trillion mark and outperformed other so-called better chips. When I write "How to Make Money in Any Market" I talk about finding Nvidia when others didn't. I now write about Nvidia, even though others choose to believe reports from competitors claiming that Nvidia has been defeated. When I first chose to believe in Nvidia, everyone laughed at me, including my colleagues. I now read Nvidia obituaries the same way I hear prophecies of destruction when I talk about Nvidia being the best in the semifinals. The best of technology. The best of all companies in the world. I say I'm prepared to be wrong again. But I'm not selling because I might have been wrong before. I chose to do nothing because I know Jensen and Nvidia and I'm not blinded by love; I am aware of empirical data and excellent information. I guess you could say I'm blind to the facts. I happily accept this form of blindness. And happily. (See here for a complete list of stocks in Jim Cramer's Charitable Trust.) When you subscribe to the CNBC Investing Club with Jim Cramer, you will receive a trade alert before Jim makes a trade. Jim waits 45 minutes after sending a trading alert before buying or selling a stock in his charitable foundation's portfolio. If Jim talked about a stock on CNBC TV, he waits 72 hours after issuing the trading alert before executing the trade. THE ABOVE INVESTMENT CLUB INFORMATION IS SUBJECT TO OUR TERMS AND CONDITIONS AND PRIVACY POLICY, TOGETHER WITH THE DISCLAIMERS. NO CIVIL OBLIGATIONS OR DUTIES EXIST OR SHALL BE RESULTING FROM YOUR RECEIVING ANY INFORMATION PROVIDED IN CONNECTION WITH THE INVESTMENT CLUB. NO SPECIFIC RESULT OR PROFIT CAN BE GUARANTEED.