Pivoting to Defense: High Bond Yields and Skepticism Test the AI Trade

Stocks kicked off September on unstable footing as rising oil prices and Treasury yields tested investor appetite for major artificial intelligence winners. For the week, the S&P 500 eked out a 0.1% gain, while the Nasdaq rose 0.4%, recovering from a rough start to a historically difficult month. Both indexes secured their fifth weekly gain in six. A stronger-than-expected August jobs report sent bond yields higher Friday, with the 2-year Treasury yield reaching its highest level since January 2025. This report compounded inflation concerns already stoked by rising oil prices amid renewed hostilities between the U.S. and Iran. Consequently, market odds of a September Federal Reserve interest rate hike rose to 58%, up from 49.4% a day earlier, according to the CME FedWatch tool.

We made meaningful adjustments to the portfolio last week, reducing our exposure to some of the AI buildout’s biggest winners, which recently fell out of favor, while adding more defensive names. Even Nvidia’s blowout earnings the prior week failed to spark a broader AI rally, reinforcing our concern that investors are becoming less willing to reward strong results across the group.

On Monday, we trimmed Palo Alto Networks ahead of its Tuesday evening earnings report, locking in a roughly 148% gain on shares purchased in August 2024. The sale was primarily about securing profits after the stock’s more than 80% run this year, especially following its pullback from all-time highs in mid-August. This move did not change our conviction that AI will drive greater cybersecurity spending.

We exited Corning on Tuesday, locking in an average 52% gain. While the stock remains up 70% this year, shares have dropped more than 40% since record highs in late June. On Wednesday, we redeployed that capital into BNY and Kimberly-Clark, each starting at roughly 1% weightings. BNY offers a less cyclical, fee-driven financial profile, while Kimberly-Clark provides a cheaper valuation, a 4.75% dividend yield, and a potential catalyst from its pending Kenvue acquisition.

We also added to Micron on Thursday, bringing the position to about 1%. Our conviction remains high as AI-driven memory demand continues to outstrip supply, supported by bullish long-term outlooks from Nvidia.

Despite better-than-expected revenue and earnings alongside a bullish long-term AI outlook on Wednesday, the market remains reluctant to reward Broadcom’s growth. CEO Hock Tan raised his fiscal 2027 AI revenue forecast to $115 billion and expects it to double to $230 billion in fiscal 2028. Still, concerns around customer concentration, vendor financing, and data center pushback continue to weigh on the stock. We lowered our price target to $430 from $480 while maintaining our hold-equivalent 2 rating, and shares ended the week down 3%. Ahead of earnings the prior week, we sold half of our remaining Broadcom position to reduce exposure to the AI data center buildout. Jim Cramer noted during Thursday’s Morning Meeting, “The only regret I have is … I didn’t just get rid of it all.” However, he stated he is willing to hold the remaining position due to Broadcom’s deepening relationship with AI startup Anthropic, which is set to go public this fall.

Palo Alto Networks also delivered strong results after Tuesday’s close, reinforcing our view that AI is a major tailwind for cybersecurity. After Wednesday’s rout in the stock, we were glad we trimmed going into the print. Following Wednesday’s pullback, we upgraded Palo Alto to a hold-equivalent 2 from our sell-on-strength 3 rating and raised our price target to $400 from $380 on Tuesday night. CEO Nikesh Arora stated that roughly $1 trillion of global cybersecurity infrastructure needs to be modernized to defend against emerging threats, noting on CNBC’s “Mad Money” on Tuesday, “You cannot deploy AI successfully if you do not get cybersecurity right.” Nevertheless, shares sold off on profit-taking, leaving Palo Alto as the portfolio’s worst performer last week, losing more than 10%.

Nvidia’s $12.9 billion acquisition of Hugging Face strengthens the ecosystem around its chips by providing access to an open-source AI platform used by more than 18 million developers. We see parallels to Microsoft’s 2018 purchase of GitHub. Rather than directly monetizing the platform, Nvidia can deepen developer loyalty, make its hardware and software a natural choice for deploying AI models, and keep a strategic asset out of competitors’ hands. The deal also helps Nvidia hedge against hyperscalers’ custom chips. Whether developers use open-source or proprietary models, both require massive amounts of compute. As long as AI adoption keeps expanding, Nvidia stands to benefit. Despite gaining roughly 22% this year, Nvidia trades at just 14 times next year’s expected earnings. Cramer noted that remains too cheap given the company’s growth and expanding AI ecosystem, suggesting a larger share buyback could provide another catalyst. The stock rose 6% last week, edging ever closer to mid-May record highs.

Source link