My week in San Francisco, where I attended Salesforce’s annual AI conference and visited with the CEOs of Starbucks, Broadcom, and many others, was filled with fresh insights about AI, stocks, and the market at large. Below are my top 10: Consider these takeaways to inform your investing decisions in the days and months ahead. 1. Salesforce’s AI platform is working — just not the way some think it is. The skepticism surrounding Salesforce’s Agentforce, its platform with agentics and virtual robots to augment your business, is growing thinner after CEO Marc Benioff’s keynote to kick off the company’s annual Dreamforce conference. There’s a two-pronged rap that has caused Salesforce’s stock to be one of the worst performers in the market. One is that its agentics initiative is so powerful that it created geniuses who can write code that is as good as the kind that Salesforce’s legacy software unit creates; therefore, you don’t need legacy Salesforce anymore. And second, because your agentics codewriters are so effective, you can lay off people and hence not have to pay as much as you currently do to Salesforce for its help. Remember, businesses pay for legacy Salesforce by “the seat,” or user, so if you have fewer people doing sales, you don’t need as much Salesforce product. Every year, Benioff gives an impressive 90-minute keynote about all of his new products aimed at increasing his clients’ sales. When I was helping to run thestreet.com, I found his addresses very helpful for gaining new subscriptions. I learned a lot about what to do from the keynotes. This time, Marc had a very different kind of presentation. He brought in leaders of four companies — PepsiCo , Dell , FedEx , and Williams-Sonoma — and showed how they were using agentics to help streamline their businesses, save money, and grow revenues. Actual real use cases where nobody cut back on legacy Salesforce and instead used it more efficiently. All the clients had the option to cut back legacy — it is now included as an option when you take Agentforce — but it doesn’t seem like many are taking it. In fact, at an investors meeting during the conference, Salesforce raised its organic revenue growth rate to above 10% for the 2026 through 2030 fiscal years — it had fallen to 9% — and said it expected sales of more than $60 billion in 2030, much higher than analysts’ estimate of $58.37 billion. This projection is important because Marc has been beating the company’s internal projections but not those of the Street. It had become a perennially disappointing quarterly progression. That’s over. Now he will show you that Agentforce has become meaningful, hopefully in the multibillions of dollars in the next 18 months, and the stock will start its upward climb. I pulled up with three of the four CEOs who were part of the presentation to confirm everything I have written here, and I feel much better about our position than I have in a very long time. 2. OpenAI backers are getting worried … The skepticism surrounding OpenAI’s spending is growing thicker as it moves into consumer products like Sora, which generates realistic and animated videos from text prompts, and “erotica” for adults — both of which signal that it’s not succeeding in the enterprise. Analysts hate consumer software because consumers are fickle and won’t pay for it. Enterprises pay a lot and don’t change vendors. The fact that OpenAI is already going deep into the consumer is making a lot of its backers extremely nervous. The company needs a gigantic revenue ramp next year to justify all of its spending, and the crowd I hung with thinks it won’t be possible. 3. … but Broadcom’s Hock Tang isn’t. Against that skepticism is the talk I had with Hock Tan, the very tough Broadcom CEO, who is a total believer in the work he intends to do with OpenAI and says they will be tremendous partners. Hock is way too stringent and hard-nosed to work with pie-in-the-sky companies. It was hard for me to reconcile my own skepticism about OpenAI with Hock’s strong belief in the company. Let’s just say it made me more confident of OpenAI’s staying power. 4. The AI buildout is not a zero-sum game. Advanced Micro Devices CEO Lisa Su was on everyone’s lips as the dragon slayer, or the David vs. Goliath, for her battles with Jensen Huang and Nvidia . It’s a false construct. These hyperscalers need everything they can get, I mean everything, and that includes custom chips from Broadcom, light-duty chips from AMD, and heavy-duty software-stuffed stacks from Nvidia. It’s not zero-sum. Yet people seem to think it is. AMD’s chips won’t be ready until the middle of next year; by then, there will be a whole new iteration from Jensen, the Vera Rubin, which has dramatically more power than AMD’s offering. If you sell Nvidia off of OpenAI’s deals with Broadcom or AMD, you are making a mistake. But you can buy AMD and Broadcom to go with your Nvidia. 5. It’s time to sell your nuclear stocks … We keep hearing that the data centers use so much power that it is going to change the grid. I come back thinking that the grid is up to it, but the energy it produces has to be distributed better. Go back and listen to my interview with PG & E CEO Patti Poppe. There is no power shortage, she said, just a lot of power that can be produced when no one uses it. She said if you can spread the power use around and perhaps store it, that will be a better solution than what we are doing now. I felt better about the power portion of the equation after speaking to her and to Hamid Moghadam , the CEO of Prologis , which has several large data centers, and they are powered with rooftop solar. Many tech execs don’t know much about power and should spend some time with the big utility CEOs to get a better handle on the problems on both sides of the equation. Further, the execs I talked to say you would be nuts not to sell every uranium and nuclear stock. There’s no real push to build new nukes, even of the small size, or GE Vernova , which is the most established maker, would be seeing orders. Sell every one of these. Check out my new book, “How To Make Money In Any Market,” where I outline my system for building long-term wealth through your investments. 6. … your quantum stocks, too. Quantum is so not ready for prime time that I would sell all of these stocks. Nobody is seriously talking about quantum as a way to get things done, and if they are, it is about seeing how IBM does with it commercially. The quantum stocks, like the nuclear stocks, are totally overblown and are living by press releases, typically from a cash-strapped government. While not pie in the sky, the quantum computing breakthroughs won’t be producing commercial compute for a very long time. All of these companies need to do huge equity offerings. Insider selling will soon overwhelm them, and they have really weak-kneed shareholders. 7. Don’t sleep on Meta’s smart glasses. Mark Zuckerberg’s view of the glasses and a small handheld pocket computer as a way to convey AI to humans may be taking shape faster than I thought. Meta’s glasses, which are a little Elvis-Costello-like, send a great deal of info back to the cloud. But they also have a lot of computer power on premises, so to speak, because they are loaded with Arm chips. These glasses can call, take pictures, and speak lots of languages, among many other uses. They seem to be a must for any traveler and any business person trying to get things done overseas. 8. You want to buy Dell stock if it comes down. Dell is pulling away from all of the other coordinators of Nvidia installation and the racks that the chips are a part of. I thought that Hewlett Packard Enterprise was catching up, but it had a miserable quarter. There are real dreams in Nebius , and I would sell that one tomorrow; that’s how far ahead Dell is. I regret that I didn’t pull the trigger on buying Dell. It should be bought if the stock price comes down because it is doing so much better than everyone else. I am not buying into IREN , the bitcoin and data mining company. Sell that one and buy Coreweave . These sales may not immediately be right. But the insider selling that is to come, and the corporate financing that must come, will spell the end of the upward spiral. 9. Levi Strauss is back. Away from tech, I was very impressed by the growth of Levi’s women’s line, and Levi Strauss is on a new growth path that is not reflected in the stock. My whole team was blown away by the offerings that CEO Michelle Gass has put together, and the stock is way too cheap versus the innovation she’s brought to the party. 10. San Francisco is coming back, too. The city of San Francisco is getting safer, thanks to the tireless work of Mayor Daniel Lurie. The density, which had vanished, is coming back because of start-ups that are looking for cheap real estate. There are way too many open storefronts still and far too few police — hence Marc Benioff’s wayward remarks about the need for National Guard troops. There are fewer restaurants because of so many closures. The nighttime is still a little quiet. But I felt safe. The menacing tent cities and the over-the-top drug use on city streets are gone. It is nascent, but it is real. (See here for a full list of the stocks in Jim Cramer’s Charitable Trust.) 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.