Before taping a Friday “Mad Money,” I always look at the week’s market performance just in case the percentages are noteworthy. They usually aren’t. But this week things were different, very different. The S & P 500 and the Nasdaq had their best weekly gains of 2024, gains that made up for the prior week, which was the worst of the year. This kind of reversal has to be mentioned, even if it only shows the trivial nature of these big increases and declines. These are moves you might as well chalk up to meaningless volatility, largely based on random articles about the Federal Reserve, punctuated by the continued despair and moaning about the loss of the data center. The week before, Broadcom had allegedly disappointed, causing Nvidia , the inevitable key to this market, to continue its descent. But last week started with a bang. Larry Ellison, co-founder and chairman of Oracle and one of the most revered people in Silicon Valley, said that over time his company could operate another 1,000 or 2,000 data centers, up from current 162. There was a delayed reaction to the news, as analysts debated whether Ellison was hallucinating or not. When they realized he wasn’t, the stock jumped to new highs, a move we have been waiting for since the enterprise software giant swung and missed twice, both times brutally. It was enough to make you think that Oracle’s huge buy of medical records company Cerner was so bad that management had taken its eye off the main chance. Maybe the $28 billion deal was a disappointment, a victim of an entrenched and better system named Epic, a private company. Could it have been stillborn and the data center initiative a cover-up? When you lose money twice in a stock you tend to doubt everything. In the end, you couldn’t dismiss Ellison’s comments on Sept. 10 after the company reported quarterly results that topped expectations. He was too convincing. He easily could have said he needs another 200 data centers. That would have done the trick. But he went all in and had the numbers to back it up. The good news helped reverse the gloom because so much has been riding on data center growth. That includes everything from the physical (think Eaton and Vertiv ) to the digital ( Marvell and Broadcom). Oh, yeah and the real winner Nvidia , because Ellison verified Jensen Huang’s view that not only are Nvidia platforms necessary, they pay for themselves almost instantly. That’s quite a difference from the week before when Oracle, Meta , Microsoft and Google played the fool, with each spending $50-to-$60 billion on Nvidia’s semiconductors only to find that they had radically overspent. The only real winner was the freeloading Apple , something quite evident at the Apple event on Sept. 10, something apropos of its giant installed base. As I said in our Monthly Meeting, nobody wants to be left behind in the AI buildout and become what Bing was to search. At one time Microsoft’s Bing was a chief rival of Google in search, but Google (now the premier part of Alphabet ) chose to spend and and locked up the market. Nobody wants to be “Bing-ed” in this AI moment by underinvesting. Oracle understands that. By the time Nvidia’s Huang spoke on Wednesday , we realized three things: Blackwell will be relatively on time; there were tense and emotional battles about allocations; here would be no end to the spending as the use cases were all about staying in the race. We needed to hear these comforting words because of a suboptimal presentation by a high-ranking JPMorgan executive last week who told analysts that expectations for net interest income and expenses in 2025 were too optimistic. Anyway, Jensen’s narrative held right into Friday’s session and stocks like Broadcom benefitted mightily. Plus, an article in The Wall Street Journal suddenly put a 50-basis-point cut by the Fed into the mix. That alone would have corrected all of the prior week’s sell-off. I had been preaching sitting on your hands through this period, but we have recently put some money to work because the market had finally become oversold. When the market goes down, you have never sold enough. When it goes up without you, you almost feel it’s the same. Almost. Which brings me back to the initial concept of the big declining week followed by the big advancing week. We have been treated to a hefty dose of September being a bad month. What kind of month gives you the benefit of the doubt and a subsequent rally of huge proportion? How about a good one? So what do we do now? There might be a stair-step pattern going and if we only get a 25-point cut this week you have to pounce into the weakness. Do not worry about dip buying — it pays off. The homebuilders and the retailers soared. They will have to shed a lot of their gains on a 25-point cut. This week I will be in San Francisco for Dreamforce, Salesforce’s annual AI event, trying to get the measure of useful artificial intelligence. Most of us are sick of window dressing AI and commodity AI, we are now in the show-me phase. And not many are as we saw from the week’s only real downer, the forecast by Adobe . Without sufficient explanation or a cost reduction, one tied to headcount or earnings per share, the group is once again vulnerable. One thing is for certain: the financials, part of the broadening out, were lost as JPMorgan did them in with its haphazard estimate cut. Healthcare had its good sides though as devices, big pharma and hospitals showed up. The real action was with Microsoft and Amazon , two stocks that had not really gotten any traction post earnings. That change hardly got notices, but it was extremely meaningful because it showed that some of the Magnificent Seven still have legs. As I look at the landscape, I keep coming back and urging the buy of Dupont , a special three-part breakup that makes it a super exciting stock. A good week for Kamala Harris turns out to be a big week for Nextracker — and that’s really it for now. (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.
Oracle Chairman of the Board and Chief Technology Officer Larry Ellison delivers a keynote address during the 2019 Oracle OpenWorld in San Francisco on Sept. 16, 2019.
Justin Sullivan | Getty Images
Before taping a Friday “Mad Money,” I always look at the week’s market performance just in case the percentages are noteworthy. They usually aren’t. But this week things were different, very different. The S&P 500 and the Nasdaq had their best weekly gains of 2024, gains that made up for the prior week, which was the worst of the year.