由 jpichiban Tue, 28 Jan 2025 19:23:47 GMT
From Wall Street to Silicon Valley, investors are still trying to get their arms around the emergence of DeepSeek, the Chinese startup whose more efficient but competitive artificial intelligence model sent shockwaves through the U.S. stock market this week. There are clear signs of stabilization Tuesday, though. The tech-heavy Nasdaq Composite added about 1%, cutting into some of Monday’s 3.5% rout. Artificial intelligence chip king Nvidia , one of the hardest-hit Club names a day ago, climbed jumped 6% Tuesday, to nearly $126 a share. Fellow chipmaker Broadcom , another severely beaten-up AI stock, added just over 1%. Club holding Eaton , on the other hand, fell another 3% after tumbling 15.6% on Monday. However, we took advantage of that second day of weakness to buy back Eaton shares we sold in late October, when the stock traded more than 10% higher than Tuesday’s price. Eaton makes electrical components and power management systems needed to help energy-intensive AI data centers run. Still, it’s safe to say that within our portfolio, these three stocks may encounter the most short-term headwinds tied to the DeepSeek fallout. At the very least, the elevated uncertainty may mean investors are not willing to pay as much of a premium for their future earnings as they did last week, leading to what’s known as “multiple” compression. Indeed, Morgan Stanley’s lowered price targets for both Nvidia and Broadcom are a function of that very dynamic. A similar rationale motivated Melius Research’s downgrade of Eaton. In changing their rating to hold from buy, analysts flat-out said they aren’t concerned about the electrical equipment supplier’s earnings this year or in 2026 — just how much investors will pay for them. That could lead to less upside than expected than previously expected, but it’s not the same thing as an erosion of the fundamental business. The more we learn about DeepSeek, however, the more it’s clear that plenty of stocks in the portfolio can benefit from its advancements, particularly lower day-to-day costs of using the model, a process known as “inference.” Here is our latest thinking on parts of the portfolio that stand to gain. Potential winners Before we get into each stock, it’s worth remembering: Being a winner from DeepSeek does not mean a company will realize the benefits immediately, or even next week or next quarter. It’s going to be a long-term trend, and, as the seemingly overnight popularity of DeepSeek shows, things could always change given this type of AI technology is still nascent. Meta Platforms is an easy winner if AI becomes cheaper to develop and implement because already, we’ve seen evidence that AI is enhancing its business through more engagement and better return on investment for advertisers. With DeepSeek making its models open source, companies such as Meta are in a position to learn from its efficiency gains, analysts at Citigroup told clients. This could mean that Meta, which also has embraced open-source AI, could reduce its future capital expenditure needs, as new models are developed at a lower cost. The timing of DeepSeek’s ascent is convenient, at least as far as curious investors are concerned. Meta is set to report earnings Wednesday evening, and DeepSeek figures to be a big topic of conversation with CEO Mark Zuckerberg on the conference call. “They want AI cheap, they get AI cheap,” Jim Cramer said Tuesday of Meta. Shares of Meta added another 3% on Tuesday after posting a record close of $659.88 on Monday. The big question going into earnings: Did Zuckerburg have DeepSeek on his radar before Friday’s announcement of plans to spend up to $65 billion on capital expenditures this year? That figure was a lot higher than the consensus Wall Street analysts’ estimate of roughly $51 billion in 2025 capex spending. Amazon is another potential winner, in large part because the cloud-computing giant has adopted a more “open source” approach to AI. While the company developed its own family of AI models called Nova , it has positioned its Amazon Web Services (AWS) cloud unit as a friendly place for all the leading AI models. The company’s Amazon Bedrock service on AWS provides access to models from startups such as Anthropic and Cohere, as well as Meta’s open-source Llama models, among others. “Amazon has anticipated AI is moving to commoditized models and hardware and has designed its whole cloud/AI platform around this,” Oppenheimer analysts wrote to clients Monday. And, as with Meta, Amazon being able to save money on AI development by leveraging DeepSeek’s technology is a good thing for its bottom line. Amazon is a financial backer of Anthropic. “They can crib” off the success of others, Jim said. “This is another [stock] that a lot of people felt was overbought, too high, as recently as Friday. Then you have [DeepSeek], and all of a sudden people start thinking, ‘Gross margin expansion,'” he said. Amazon shares added 1.5% on Tuesday after squeaking out a modest gain Monday. Apple also lands in the potential winners’ circle. Jim has long touted the financial merits of Apple’s partnership-heavy AI strategy, calling it a “free rider” because it hasn’t had to spend as aggressively on capital expenditures compared with tech giant peers. Incorporating ChatGPT from Microsoft-backed OpenAI into the Apple Intelligence software is the No. 1 example. Now comes along DeepSeek, which has fueled hopes that even cheaper AI models can be incorporated into Apple’s iPhone and other hardware down the road. “If the cost for AI inferencing truly comes down and better models are developed faster and cheaper, in our opinion, the edge AI applications would see the most benefit. This should benefit smartphone and PC makers in theory,” Bank of America wrote to clients Monday. During Tuesday’s Morning Meeting , Jim reiterated his view that Apple is an “own it, don’t trade it” stock. However, he cautioned once again that Apple’s quarterly results on Thursday evening could be noisy. “OK quarter, bad guide,” Jim said, regarding his expectations. Shares of Apple jumped nearly 4% on Tuesday, building off Monday’s 3.2% gain. The stock is still down nearly 5% year to date, despite the strong week. The S & P 500 is up 3% in 2025. Even the Nasdaq, which had its worst session of 2025 on Monday, is up nearly nearly 2% year to date And finally, Salesforce is a winner of cheaper AI. Even before the DeepSeek news, investors — including the Club’s portfolio managers — were upbeat about the potential for the company’s new suite of AI agent-building tools, dubbed Agentforce, to spark an acceleration of topline growth in the coming quarters. The prospect of cheaper AI could not only help with adoption among customers but lower what Salesforce needs to spend going forward. That would lead to an expanding bottom line, too. “Essentially, we believe this [DeepSeek] technology can benefit those in our space who are the furthest along in their AI journeys,” which includes Salesforce, software analysts at Wolfe Research wrote Monday. “This is a parabolic move, which you know worries me,” Jim said, rereferring to Salesforce’s 4% advance Monday and additional 4.5% rise Tuesday. Still, Salesforce is a “mass adoption play,” he said. “If there’s mass adoption of AI, then you’re going to use Salesforce.” Jim is set to interview Marc Benioff, the CEO and chair of Salesforce, on Tuesday night on “Mad Money.” Unclear (for now) At this point, the impact on Google parent Alphabet is less certain. The company definitely benefits from lower computing costs, which would reduce its capex spending. However, the concern would be that cheaper AI models lead to greater adoption among consumers, putting pressure on its cash cow Google Search business. The threat of AI disrupting search is not new. It’s been lingering ever since the launch of ChatGPT in late 2022 sparked this current AI boom. DeepSeek is just another reminder that we still don’t have a resolution. In a not e to clients Monday, Wells Fargo analysts said they maintain a preference for tech giants with “open platforms,” such as Meta and Amazon, over the more “closed” ecosystem that Alphabet has. “Very hard to tell” what DeepSeek could mean for Alphabet, Jim said. He noted, however, that he’s been on the fence about the stock even before DeepSeek’s emergence. “This is one I remain worried about. I don’t know if they can blow out the quarter.” Its fourth-quarter report is set to be released after the close next Tuesday, Feb. 4. Shares of Alphabet fell 4.2% on Monday during the initial DeepSeek fallout, but were higher by about 1.5% in afternoon trading Tuesday. To be sure, Google has its own generative AI tools baked into Search, which provides AI results in addition to returning the regular search engine results. Google’s Gemni AI is also a standalone like ChatGPT. Microsoft also finds itself in the to-be-determined camp. The potential benefits on the capex side are present. However, analysts at Oppenheimer noted that, unlike Amazon, Microsoft’s Azure unit has leaned heavily on the proprietary models of OpenAI, calling that a negative in light of the DeepSeek news. Analysts acknowledged Microsoft has recently started to move away from such an OpenAI-heavy approach and “downshift” into a strategy that looks more like Amazon’s — but they argued the company needs to speed up that process. For its part, UBS said Microsoft’s pivot toward building data centers to run AI models for customers on a day-to-day basis — and focusing less on making Azure the place for OpenAI to train its models — is validated by DeepSeek. We’ll be listening closely to Microsoft CEO Satya Nadella on the company’s earnings call Wednesday evening. Another bit of convenient timing. Jim said he believes there is a growing gulf between Microsoft and OpenAI, which has been cozying up with Oracle , most notably on the Stargate Project to build lots of data centers. It was announced at the White House last week alongside President Donald Trump . “We’ve got to see how that’s resolved,” Jim said. “If Microsoft loses OpenAI, they’ve got to spend and then spend and then spend to make their own [AI].” (Jim Cramer’s Charitable Trust is long NVDA, ETN, AVGO, META, AMZN, AAPL, CRM, GOOGL and MSFT. See here for a full list of the stocks.) As a subscriber 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 trade alert before buying or selling a stock in his charitable trust’s portfolio. If Jim has talked about a stock on CNBC TV, he waits 72 hours after issuing the trade alert before executing the trade. THE ABOVE INVESTING CLUB INFORMATION IS SUBJECT TO OUR TERMS AND CONDITIONS AND PRIVACY POLICY , TOGETHER WITH OUR DISCLAIMER . NO FIDUCIARY OBLIGATION OR DUTY EXISTS, OR IS CREATED, BY VIRTUE OF YOUR RECEIPT OF ANY INFORMATION PROVIDED IN CONNECTION WITH THE INVESTING CLUB. NO SPECIFIC OUTCOME OR PROFIT IS GUARANTEED.A DeepSeek artificial intelligence logo on a laptop, arranged in Riga, Latvia, on Monday, Jan. 27, 2025. Chinese artificial intelligence startup DeepSeek rocked global technology stocks Monday, raising questions over America’s technological dominance. Andrey Rudakov | Bloomberg | Getty ImagesFrom Wall Street to Silicon Valley, investors are still trying to get their arms around the emergence of DeepSeek, the Chinese startup whose more efficient but competitive artificial intelligence model sent shockwaves through the U.S. stock market this week.
原文链接：https://www.cnbc.com/2025/01/28/4-portfolio-stock-winners-from-deepseeks-emergence-and-2-were-less-sure-about.html