Ideas
Tech is on AI, not rate, cycle.
The Fed's rate-hike cycle is bad for most stocks near term, but tech is different because it is tied to the AI/industrial revolution and the huge data-center buildout rather than the rate cycle; demand is strong enough that buyers should return to tech.
Data-center component makers are buys.
The data-center buildout is the economy's main source of strength, isn't slackening, and is not very rate-sensitive. After the Fed fallout, companies that make components for data centers, including Nvidia, Intel and Micron, are buys.
High oil risks embedding inflation.
Oil prices are on a relentless run because of the Iran war. If they stay higher for longer, inflation becomes embedded and forces more Fed rate hikes; the Fed cannot offset this without crushing demand.
Pharma works regardless of rate hikes.
In a Fed tightening cycle he wants stocks that can work regardless of rates, and pharma fits that because demand is less rate-sensitive.
CoreWeave buy despite rate speculation.
CoreWeave borrows heavily, which is harder and more speculative with the Fed hiking, but it sits in the middle of the great AI industrial buildout, so he thinks it's okay to buy while recognizing the added speculation.
RTX multiple compresses under rate hikes.
RTX trades at about 27 times earnings, a high multiple that must compress in a Fed rate-hiking cycle; that's why the stock is going lower.
AI safety drives cyber security demand.
Cyber security is uniquely positioned because AI safety concerns make security more essential, and cyber stocks are not as rate-sensitive as the rest of the market; he owns Palo Alto and CrowdStrike for the club and likes Okta.
Own Palo Alto; AI needs security.
Palo Alto Networks is a cyber security leader he owns for the travel trust; Wall Street is finally recognizing you can't have powerful AI without equally powerful cyber security, and the stock has caught fire.
AI adoption requires Palo Alto security.
AI is a once-in-a-lifetime technology that will not slow down; the right conversation is how to deliver it safely and securely. Every technology historically requires cyber security, enterprises need governance and guardrails for agents, and Palo Alto is building hooks into AI labs' products for real-time cyber defense.
ARM buy on multi-year chip shortage.
ARM Holdings fell from $452 in June to just under $244, but he sees a multi-year chip shortage regardless of AI-slowdown fears; the trust had a great position, took profits, and that was a big mistake.
ARM demand strong; AI compute platform.
Demand for ARM technology has never been stronger across edge, automotive, robotics and data center because AI is compute-intensive and requires CPUs, GPUs and memory. Confidence in the $2 billion business has increased from May to July to September, data center will become ARM's largest business, and ARM is the compute platform for AI.
AI chip supply constrained for years.
The big constraint for AI chips will be supply: wafers for logic chips, substrates, testers, memory and the entire supply chain. He sees complex supply-chain issues for the next number of years.
Okta essential as AI agents roam.
Okta reported a phenomenal quarter, is no longer held back by AI-displacement worries, and its identity-verification software is more essential than ever as AI agents roam the internet; the stock is up 140% over six months.
Okta controls AI agent identities.
AI agents are a powerful new identity type. To make them more powerful, companies give them access to data, tools and actions, so they must have identities to be controlled, monitored and made visible. The foundation of the fix is visibility and control, with identity as the key, and Okta provides the blueprint for the secure agentic enterprise.
Nokia is a buy right now.
He likes Nokia very much, calls it a terrific situation and would be a buyer right now; the caller cites expansion in New Jersey, Pennsylvania and Texas, Nvidia and Google partnerships, networking and data-center buildouts, and AI defense systems.
SHIP is cheap with safe yield.
Bulk transportation is on fire; SHIP is not an expensive stock and has a safe yield, so he would be a buyer.
Hub Group rough until oil calms.
Hub Group is tough after JB Hunt's bad numbers; his travel trust has a small trucker position and he expects rough going until oil calms down, then it should be okay.
BWXT too expensive despite nuclear.
BWX Technologies is a great company with nuclear exposure, but at 30 times earnings it is too expensive in a tightening cycle, so he says hold off.
Take money and avoid NextEra.
If Dominion Energy shares convert into NextEra Energy in the merger, take the money and run; NextEra is not a good stock to own versus Dominion.
Redwire loses money in tightening cycle.
In a rate-tightening cycle he cannot buy companies losing money left and right; Redwire loses money, so it won't work.
Own CrowdStrike for AI cyber defense.
He relies on companies like Palo Alto, Okta and CrowdStrike to stay one step ahead of enemy AI swarms; cyber security stocks make sense, and he owns CrowdStrike and Palo Alto for the club.
This CNBC video, published September 16, 2026,
features Jim Cramer, Nikesh Arora, Rene Haas, Todd McKinnon
discussing XLK, AI-SECTOR, NVDA, MU, INTC, WTI, XLV, CoreWeave, RTX, CIBR, PANW, ARM, SMH, OKTA, NOK, SHIP, HUBG, BWXT, NEE, RDW, CRWD.
21 trade ideas extracted by AI with direction and confidence scoring.
Speakers:
Jim Cramer,
Nikesh Arora,
Rene Haas,
Todd McKinnon
· Tickers:
XLK,
AI-SECTOR,
NVDA,
MU,
INTC,
WTI,
XLV,
CoreWeave,
RTX,
CIBR,
PANW,
ARM,
SMH,
OKTA,
NOK,
SHIP,
HUBG,
BWXT,
NEE,
RDW,
CRWD
