Every month, JPMorgan quietly updates its highest-conviction stock picks. Most changes are incremental. The April 2026 update is different. The bank added two names — Palo Alto Networks (PANW) and JFrog (FROG) — and kept Constellation Energy (CEG) anchored at the center of its energy thesis. Together, these three moves encode a clear message: the AI trade is no longer about who makes the chips. It is now about who secures the infrastructure, who manages the software pipeline, and who keeps the lights on.

PANW: Cybersecurity as Non-Negotiable Infrastructure

JPMorgan’s decision to add Palo Alto Networks formalizes a conviction that has been building. The bank is now described as “incrementally more positive” as the cybersecurity company continues to gain market share and repositions itself as a direct AI beneficiary. The thesis is straightforward: as generative AI lowers the cost and complexity of launching sophisticated cyberattacks, enterprise security budgets are not being cut — they are being reinforced.

The financials support the conviction. PANW delivered fiscal Q1 2026 revenue of $2.47 billion, representing 16% year-over-year growth and exceeding the high end of guidance. Operating margins exceeded 30% for the second consecutive quarter. Next-Generation Security ARR reached $5.85 billion, up 29% year-over-year. Looking further out, management has reaffirmed a 40%+ adjusted free cash flow margin target by FY2028, inclusive of the CyberArk and Chronosphere acquisitions. The company raised its long-term NGS ARR target to $20 billion by FY2030.

Analyst consensus backs the position. Of the 56 analysts covering PANW, 44 rate it a Strong Buy or Buy (LSEG), with a consensus price target implying approximately 31.5% upside from current levels. This is not a speculative AI play. It is defensive infrastructure with expanding margins and durable secular demand — precisely the profile JPMorgan favors in a volatile macro environment.

FROG: The Contrarian Bet on the AI Software Supply Chain

JFrog is the less obvious pick — and likely the more revealing one. The DevOps platform is down roughly 23% year-to-date, with underperformance attributed to what analysts describe as the “fear of long-term AI disruption.” JPMorgan is betting that fear is misplaced, noting that JFrog is “well positioned to benefit from meaningful AI-related tailwinds.”

The core thesis centers on the DevOps-to-MLOps transition. Every enterprise AI application requires a software supply chain — a system to build, manage, secure, and deliver software artifacts from development to production. JFrog’s platform is designed to be exactly that system of record. The company recently announced an integration with NVIDIA’s Agent Toolkit to develop a new Agent Skills Registry, further cementing its positioning in AI governance. Full-year 2026 revenue guidance of $623–$628 million implies approximately 17.5% year-over-year growth. The company also authorized a $300 million share repurchase program in February 2026, underscoring balance sheet confidence.

All but one analyst covering FROG rates it a Strong Buy or Buy (LSEG), with an average consensus target implying approximately 42% upside. Adding a stock down 23% on the year in a market nervous about AI software disruption is a high-conviction move. The market is pricing in disruption risk. JPMorgan is pricing in adoption acceleration.

CEG: Nuclear as the AI Economy’s Structural Power Source

Constellation Energy is not a new JPMorgan name — but its continued presence on the list carries its own weight. CEG functions as a nuclear safe haven: a power producer whose output is contracted, carbon-free, and structurally disconnected from commodity price volatility.

The anchor of the thesis is CEG’s landmark deal with Meta. Constellation and Meta signed a 20-year power purchase agreement for the full 1,121 megawatts of output from the Clinton Clean Energy Center in Illinois, commencing June 2027. The agreement replaces the state’s expiring zero-emission credit program with a direct market-based contract — ensuring the plant’s long-term operations without ratepayer support. Meta’s rationale was direct: “Securing clean, reliable energy is necessary to continue advancing our AI ambitions.” Beyond Clinton, Constellation is evaluating the deployment of advanced reactors and small modular reactors at the site.

The structural case is durable. Nuclear output does not fluctuate with natural gas prices. Contracted cash flows extend two decades. And with AI hyperscalers increasingly anchoring to nuclear as a power source — Meta following Microsoft’s earlier 20-year agreement with Constellation at Three Mile Island — the demand side of the equation is only growing.

The Rotation in Plain View

Step back from the individual names and a coherent rotation is visible. The first phase of the AI trade rewarded semiconductor and hardware names. That phase is maturing. JPMorgan’s April reshuffle — removing EQT in favor of names insulated from natural gas volatility, and adding cybersecurity and DevOps software — reflects a deliberate shift toward the operational layer of the AI economy: who secures it (PANW), who builds and manages it (FROG), and who powers it (CEG). These are not speculative bets. They are investments in the infrastructure requirements of an AI economy moving from proof-of-concept into deployment at scale. Each name carries a distinct risk profile, but together they represent a rotation worth tracking with discipline.

For informational and educational purposes only. Not investment advice. Trading involves risk, including loss of principal.