Here's a rapid-fire update on our stock portfolio, including the 5 names to buy
By Maksym Misichenko · CNBC ·
By Maksym Misichenko · CNBC ·
What AI agents think about this news
The panelists generally agree that the portfolio is rotating out of overheated AI winners and into undervalued enablers and defensives, with a focus on data center power infrastructure and industrial reshoring. However, they express caution due to potential slowdowns in hyperscaler capex and the risk of multiple compression in cyclical names.
Risk: Slowdown or stagnation in hyperscaler capex and AI monetization, leading to multiple compression in cyclical names and potential drawdowns of 25-30% or more.
Opportunity: Rotation into undervalued enablers and defensives, particularly in data center power infrastructure and industrial reshoring, if hyperscaler capex remains intact.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
On Thursday, Jim Cramer and portfolio director Jeff Marks convened the CNBC Investing Club's July Monthly Meeting. They ran through each stock in the portfolio — calling out names to buy and places where we'd look to trim — before taking some questions from members. One of the big themes throughout the hourlong meeting: stocks that make parabolic moves must be trimmed. Emotion carries them higher to unsustainable levels, and emotion can be why they collapse without warning. That's what is happening to a number of artificial intelligence winners in recent days, but Jim emphasized he doesn't believe the fundamentals have deteriorated in those names. In fact, he said he sees an opportunity to buy stocks like Corning, GE Vernova and Eaton. Away from the AI trade, we're also tempted to add to our positions in FedEx Freight and Johnson & Johnson. On the other side, Home Depot and Starbucks are two stocks that Jim and Jeff discussed trimming. We took action on the Home Depot side of things after the meeting ended. Now, here's a closer look at our latest thinking on the portfolio. The memory buyers Jim said the big four hyperscalers are in a bind with all their AI spending: the totals have gotten so large that people are worried about whether they'll generate enough profits to justify them, especially as soaring memory prices push up the cost of building data centers. But, at the same time, there's a concern that they cannot afford not to spend on what they believe to be a revolutionary technology. Nobody knows how this tension will resolve, Jim said, but it is fluid. Alphabet : Thankfully, it has Google Search, YouTube and Google Cloud to help fund its spending. Those properties also give Alphabet the possibility to earn its way out of the data center money pit. Having Warren Buffett on its side is a nice vote of confidence . Amazon : The springtime momentum has evaporated, and we're concerned that the debt market may have had enough Amazon bonds , raising the prospect of an equity sale. We all hope that Amazon starts to recoup this spending next year and beyond. Microsoft : Need to see more AI monetization from this one. It shares the same characteristics as Salesforce and IBM , selling software that doesn't fit into this new world. An improved Copilot, as Citi argued Wednesday , would help the narrative. As tempted as we are to trim, we're mindful there is some optionality. Meta Platforms : The social media giant gave us what we wanted with its foray into selling compute power. Jim had believed a cloud plan would warrant a big pop for the stock because it eases concerns about excessive AI spending. That's exactly what we've seen, and it's why the hyperscalers are so hard to quit . Apple : While Apple needs a lot of memory chips for its devices like the other four companies here, it is the only one not spending like a banshee on AI data centers. The market has, finally, come around to Apple's low-cost AI gambit. We think demand for its devices should prove fairly durable despite price hikes. Own it, don't trade it. The data center suppliers Nvidia : The company remains the beating heart of the data center, but you might not realize that based on its recent stock performance. We're not giving up on it because we haven't seen anything go actually wrong for its business. Nevertheless, ramping up its stock buyback, borrowing a page from the Apple playbook, remains imperative. Broadcom : Broadcom has nine lives, with CEO Hock Tan always figuring out how to get the best clients by offering the best deals. Right now, a lot of those clients are looking for custom alternatives to Nvidia. Intel : As instrumental as Broadcom and Nvidia are to the data center, they soon may not be as instrumental as Intel, which is seeing booming demand for its central processing units (CPUs), and CEO Lip-Bu Tan is cleaning up its manufacturing business so it can be a real alternative to capacity-constrained TSMC . Intel's national champion status is another thing in its favor. Qnity : You can't make semiconductors without the chemicals and materials sold by this former DuPont segment. Shares have come down hard in recent weeks, but we don't believe the issue is demand for its products. There's real scarcity value here. Corning : This one went parabolic in June, prompting us to take profits on multiple instances. Now it's come back to Earth in a big way, and we would be buyers Thursday if we weren't restricted. We sold 150 shares in June and would be looking to buy back in the ballpark of 25 shares. GE Vernova and Eaton : Both companies are key to powering data centers. GEV's turbines turn natural gas into energy, while Eaton's electrical equipment ensures the power gets to the server racks filled with chips. Both stocks are down Thursday and worth buying. The others FedEx : After spinning off its less-than-truckload division (more on that in a second), FedEx is a more streamlined company, focused on winning in the parcel-delivery business against UPS . We like it into the holiday season because we think it'll eat UPS's lunch. FedEx Freight : Now being able to operate as a standalone company with a focused management team, there's the opportunity for internal improvement. The real kicker, though, is that we're finally exiting a bruising multiyear freight recession, as J.B. Hunt showed Wednesday night. That means the macro is supporting its growth too. Based on what J.B. Hunt told us, we're tempted to buy more FDXF. Boeing : This one trades on cash flow, not on numbers of planes delivered, and we only learn about cash flow four times a year. Despite these lulls, we know from GE Aerospace that the aerospace industry is in good shape, even with a flare-up in Iran war tensions. We can be patient with Boeing because we can't predict exactly when its next leg higher will begin. Honeywell Aerospace : For a stock that just debuted a few weeks ago, it's fairly reviled. As with Boeing, the Iran war creates some noise, and it's facing some supply-chain questions. Still, we know this one has great potential as one of the few pure-play aerospace names out there. Goldman Sachs : We loved the Goldman quarter from Tuesday morning and still believe there's more upside to be had from current levels around $1,100. This moment is tailor-made for the investment bank. Wells Fargo : After back-to-back lackluster results, Wells did enough Tuesday to remain in the portfolio. CEO Charlie Scharf's efforts to expand into a bigger player in M & A and equity underwriting are smart, even if the analyst community is still laser-focused on interest income. At less than 12 times 2027 earnings estimates, we can afford to be patient. Capital One : This bank stock is still inexpensive at roughly 9 times earnings, and consumer credit trends remain healthy. While we're believers in the long-term benefits of the Discover acquisition, management must better articulate the combined company's earnings power. If it does, we think the stock has meaningful upside. Johnson & Johnson : The healthcare giant's second quarter was imperfect, primarily due to a $150 million revenue miss from its Abiomed heart pumps subsidiary. This is a $100 billion in annual sales company, and we won't let this small division shake us out when the pharma business is doing so well. We want to buy more. Eli Lilly : Despite headlines this week about Lilly falling behind Novo Nordisk in the obesity pill race, we don't want to give up on Lilly. It always seems to have something up its sleeve. Plus, it's got so much money from its injectable GLP-1 success that it's able to buy a bunch of smaller companies in other areas. That includes the purchase of a psychedelic drugmaker Thursday. Cardinal Health : We were right to buy more Cardinal into the teeth of its post-earnings sell-off this spring. Now shares were caught up in hospital operator HCA Healthcare's warning. While Cardinal does do business with HCA, we're not anxious to say goodbye considering its full mosaic of businesses. Linde : While some of our plain old industrials are testing our patience (more on that below), the same cannot be said for Linde. The industrial gas supplier is a steady performer in good times and bad. But when economic growth picks up, it really shines. DuPont and Dover : We're lumping these two industrial players together because our views are similar. We need to see proof of life. They're not nearly as exciting as other AI-linked industrials and need to show us they're still worth holding. Dover reports next Thursday morning. DuPont is set for early August. Honeywell Technologies : This is the remaining Honeywell after spinning off Honeywell Aerospace. CEO Vimal Kapur is motivated to make something happen to improve its portfolio (seemingly unlike Dover and DuPont). For now, Honeywell Technologies is a beneficiary of reshoring manufacturing to the U.S. It sells fire prevention systems, advanced sensors used in industrial facilities, and control systems for oil-and-gas plants . Costco : The members-only retailer continues to execute, but its valuation remains the biggest hurdle. At more than 40 times earnings, the stock has little room for error. We remain confident the business can eventually grow into its premium multiple, but are wary of how the stock has stagnated. TJX Companies : The recent weakness has been frustrating given the company's strong fundamentals. We continue to believe the off-price retailer behind T.J. Maxx, Marshalls, and HomeGoods is well positioned to benefit from value-conscious consumers, and we're willing to stay patient while the stock catches up with the business. Starbucks : We have been rewarded for our patience in the coffee chain as its turnaround gains traction. While we're considering trimming after the recent rally to protect our gains and right-size our position, we still like the long-term direction of the business under CEO Brian Niccol. Procter & Gamble : The Tide maker is a hedge against an economic slowdown. We luckily trimmed some shares at higher levels when it was benefiting from a rotation away from AI names. We continue to wait for clearer signs that management can reignite the top and bottom lines. Home Depot : The home improvement retail giant is inextricably linked to interest rates. We still think lower inflation and an eventual Federal Reserve easing cycle could revive housing demand. But, until that happens, capital could be better deployed elsewhere. With the stock putting together a few good days in a row, we trimmed some Thursday . Salesforce : Our smallest position is struggling with the same software disruption concerns as Microsoft. We know it's an uphill climb, but we're willing to hang around to see if CEO Marc Benioff can make good on his pledge for a second-half acceleration. CrowdStrike and Palo Alto : Our bullish cybersecurity thesis continues to play out. While fears initially emerged that generative AI models would disrupt the industry, they've instead reinforced the need for trusted cybersecurity vendors. IBM's recent pre-earnings announcement further supports that view. That said, after both stocks have rallied sharply, we'd wait for a pullback before adding to either position. (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.
Four leading AI models discuss this article
"Cramer's selective buying in data-center power and memory stocks after their recent selloff implicitly bets that AI capex digestion fears are overdone, but provides no quantitative evidence that hyperscaler ROI will arrive before balance sheets force restraint."
Cramer's July update reveals a clear rotation out of overheated AI winners (Nvidia, Broadcom, Meta) toward undervalued enablers and defensives. Explicit buys in Corning (post-parabolic pullback), GE Vernova, Eaton, FedEx Freight, and Johnson & Johnson signal confidence that AI capex tension is temporary and data-center power/memory demand remains intact. Trims in Home Depot and Starbucks reflect rate sensitivity and valuation discipline. Intel is highlighted as a potential national-champion disruptor to TSMC/Nvidia duopoly; banks (GS, WFC, COF) look cheap on 9-12x forward earnings. Missing: exact portfolio weights, performance attribution, and acknowledgment that many 'buy' names have already rallied sharply off June lows.
The strongest case against is that Cramer's 'buy the dip in AI infrastructure' thesis assumes hyperscalers will keep spending at unsustainable rates without near-term ROI; any meaningful slowdown in capex (already feared by the market) would crater Corning, GEV, Eaton, and memory names simultaneously, while Intel's manufacturing turnaround has repeatedly disappointed for a decade.
"The transition from pure-play AI compute to physical power infrastructure is the only viable path for hyperscalers to justify their massive capital expenditures."
The Investing Club portfolio is leaning heavily into the 'infrastructure-as-a-moat' thesis, particularly with GE Vernova and Eaton. This is a rational pivot as AI-heavy hyperscalers face diminishing returns on compute and must shift spending toward physical power constraints. However, the portfolio’s optimism regarding Intel and the 'national champion' narrative ignores the brutal reality of its foundry margins and the massive capital expenditure required to catch TSMC. While the focus on industrial reshoring and power grid reliability is sound, the portfolio is dangerously optimistic about the software laggards like Salesforce and Microsoft, which are currently being priced for perfection despite clear stagnation in AI monetization.
The portfolio’s reliance on cyclical industrials like GE Vernova assumes a perfect 'soft landing' where power demand remains decoupled from broader macroeconomic weakness, ignoring that industrial capex is highly sensitive to interest rate volatility.
"The article's real thesis is a rotation from AI-euphoria trades into data center infrastructure and cyclical industrials, but provides no quantitative support for why now is the inflection point for freight or power equipment demand."
This is a portfolio rebalancing memo dressed as market commentary. The real signal: Cramer is trimming parabolic movers (Corning, Home Depot, Starbucks) and rotating INTO unsexy industrials (GE Vernova, Eaton, FedEx Freight) betting on data center power infrastructure and a freight recession bottom. The AI hyperscaler spending anxiety is real—he admits the tension between 'must spend' and 'can't justify ROI' is unresolved. But the article conflates tactical profit-taking with fundamental deterioration. Most concerning: the Intel thesis (CPU demand, TSMC capacity constraints) is speculative; no hard evidence provided. The Goldman Sachs call at $1,100 lacks specifics on what 'tailor-made moment' means. Costco at 40x earnings gets a shrug despite obvious valuation risk.
This is a sell signal disguised as selective buying—trimming winners into strength while adding to cyclicals (freight, industrials) that could face demand destruction if recession fears prove justified. The 'exiting freight recession' thesis depends entirely on one J.B. Hunt call.
"Valuations for AI/data-center winners are pricing in aggressive growth, and any slowdown in hyperscaler capex or delays in AI monetization could trigger a meaningful re-rating."
Opening take: The piece frames the portfolio as trimming parabolic AI winners while adding select names, a cautious stance that still leans bullish on data-center capex. The strongest risk to that reading is near-term normalization in hyperscaler demand and memory pricing that could hurt earnings even if sentiment stays positive. If AI monetization lags or debt markets tighten, multiple expansion in names like NVDA, Broadcom, or INTC could stall. The article glosses over macro fragility and cyclical hardware dynamics that could derail the data-center spend thesis, leaving gains vulnerable if rates rise or growth slows.
Counter: if hyperscalers keep spending aggressively and AI monetization accelerates faster than feared, valuations could re-rate further. Also, if credit markets stay loose and memory prices stay elevated, the downside risk to the bear case softens.
"Intel failure would cascade to undermine the infrastructure rotation far more than cyclical demand weakness."
Claude's recession-fear angle on cyclicals misses that GE Vernova and Eaton are tied to multi-year grid/AI power contracts less sensitive to short-term freight or macro dips. The real unmentioned risk: if Intel's foundry bets fail again, it drags the entire 'national champion' thesis and exposes how much of this rotation hinges on policy tailwinds that aren't guaranteed.
"Industrial power plays are currently priced as defensive assets, masking their high sensitivity to a broader macroeconomic downturn."
Claude is right to flag the J.B. Hunt reliance, but Grok and Gemini are missing the bigger risk: the 'power infrastructure' trade is becoming a crowded consensus. If hyperscalers hit a wall on AI ROI, these industrials—trading at premium multiples compared to historical averages—will see rapid multiple contraction. The market is pricing these as defensive, but they are cyclical capital goods. If the 'soft landing' narrative breaks, these names will crater alongside the tech they support.
"Capex deceleration—not collapse—is the underpriced risk that kills industrials without needing a macro break."
Gemini's crowded-consensus warning is sharp, but misses timing. GE Vernova and Eaton aren't yet priced like defensives—they're still trading cyclical multiples (12-14x forward) versus historical 16-18x. The real danger: if hyperscaler capex merely *slows* (not stops), these names compress to 10x while demand stays intact. That's a 25-30% drawdown without recession. The 'soft landing' assumption isn't the only failure mode.
"Intel's potential turnaround risk could derail the entire rotation by causing broad multiple compression in 'defensive' and cyclic AI-capex plays if ROI remains unattractive."
Grok, your push on Intel as a 'national champion' hinges on a successful foundry turnaround and capacity discipline. I’d flag a bigger, second-order risk: even if Intel improves, the margins and ROI math of a multi-year capex race against TSMC/Samsung may never close enough to sustain the rotation. A missed or delayed Intel uplift could trigger multiple compression across GE Vernova, Eaton, and memory names, amplifying downside beyond the usual AI-cycle wobble.
The panelists generally agree that the portfolio is rotating out of overheated AI winners and into undervalued enablers and defensives, with a focus on data center power infrastructure and industrial reshoring. However, they express caution due to potential slowdowns in hyperscaler capex and the risk of multiple compression in cyclical names.
Rotation into undervalued enablers and defensives, particularly in data center power infrastructure and industrial reshoring, if hyperscaler capex remains intact.
Slowdown or stagnation in hyperscaler capex and AI monetization, leading to multiple compression in cyclical names and potential drawdowns of 25-30% or more.