J.P. Morgan Research Breakdown: Tech Valuation Compression Is Already Ample, Semiconductors Outperform Software

J.P. Morgan has found that valuation compression in tech stocks is quite adequate across multiple sectors, positioning is cleaner, earnings momentum remains intact, and the capex upcycle is expected to stay on track. In its equity strategy report released on September 28, 2026, J.P. Morgan noted that the tech and AI ecosystem stalled for three months from June to last week, but valuation downward revisions have been significant in most areas, prompting a recommendation to re-engage with the sector.

Mislav Matejka, an analyst at J.P. Morgan, pointed out that during the summer, it was assessed that tech and AI would unlikely dominate the market in H2 2026, failing to be the "only game in town." This stance is maintained, arguing that extreme levels of the past will not return, yet there remain abundant opportunities within the AI space. Notably, price and earnings divergence between semiconductors and software is pronounced: forward earnings for semiconductors continue to rise by approximately 30%, while software sees virtually no upward earnings revisions.

The Trade of Semiconductors Outperforming Software Deserves Re-entry

The divergence in price performance and earnings between semiconductors and software is stark. J.P. Morgan notes that after peaking in June, the relative outperformance of software over semiconductors reversed, yet semiconductor forward earnings continue to climb roughly 30%, while software hardly registers earnings upgrades. Although software valuations have fallen to historical lows, profitability remains under scrutiny; shorting carries an unfavorable risk-reward profile and may continue to lag other segments of the AI ecosystem. J.P. Morgan recommends re-entering the long semiconductor/short software pair trade.

A comment from Anthropic CEO Dario requiring a need to "slow down the AI frontier" triggered a single-day 12% outperformance of software over semiconductors—the largest daily swing since 2025. J.P. Morgan is skeptical that AI development will decelerate significantly, viewing the race as still existential and winner-take-all.

Magnificent Seven Valuations Plunge to Decade Lows

The Magnificent Seven (Mag 7) valuations have plummeted to their lowest level in a decade, trading nearly one standard deviation below the broader market. While J.P. Morgan noted in March that downward revisions might have been excessive, it now acknowledges that structural shifts in business models, higher leverage, and declining free cash flow do warrant valuation compression, though much of it has already occurred. Earnings growth will support hyper-scaler performance; while some gains may be offset by further valuation cuts, not all will be eroded.

The hyper-scaler index, comprising Amazon, Alphabet, Microsoft, Meta, and Oracle, has underperformed by approximately 8% year-to-date. J.P. Morgan believes valuation downward revisions are largely priced in, and earnings expansion will underpin future returns.

The Capex Upcycle Remains Intact

The AI capital expenditure upcycle is expected to remain fully intact. J.P. Morgan highlights that hyper-scaler capex is projected to grow at a 28% compound annual growth rate through 2030. Corporate commentary indicates demand continues to accelerate, AI economics still justify purchasing compute capacity, and model advancements are gaining pace. Over the next two years, ample room remains for capex expansion driven by operating cash flow growth, debt issuance, and equity financing.

Semiconductor sales growth remains robust, with solid industry fundamentals making a supply-demand imbalance unlikely to materialize before 2028, while pricing is forecast to keep rising into 2027. The equipment subsector stands to benefit significantly from elevated capex, with ASML emerging as a key European equipment name. Trading cheaply on a PEG basis despite investor bets on earnings cuts, J.P. Morgan analysts view ASML as an attractive buying opportunity.

Agent AI Shifts the CPU-to-GPU Ratio

In traditional large language model infrastructure, the CPU-to-GPU ratio typically ranges from 1:4 to 1:8. In agent AI, this ratio shifts toward 1:1 and could become even more CPU-heavy, as agents generate greater workloads for orchestration, data movement, tool usage, and state management. J.P. Morgan notes that the rise of agent AI may fundamentally alter the CPU-to-GPU mix.

The launch of MetaMuse not only boosted its own stock but also lifted several semiconductor equities, particularly those tied to CPUs. J.P. Morgan's US CPU Ecosystem Basket offers exposure to leading U.S.-listed companies in CPU design, manufacturing, and support ecosystems. Conversely, agent AI may pose risks to stocks relying on consumer inertia, search friction, switching costs, or traffic control. Relevant European and US baskets significantly underperformed following the announcement.

J.P. Morgan’s allocation recommendations for the tech sector focus on three key themes. With tech valuation compression fairly complete and positioning cleaner, the firm advises re-engaging with the sector. The long semiconductor/short software pair trade merits re-entry given persistent semiconductor earnings momentum. Finally, with hyper-scaler valuations at decade lows, earnings growth is expected to drive subsequent performance.

Disclaimer

This article consolidates and interprets a third-party brokerage research report (J.P. Morgan, September 28, 2026), supplemented with publicly available market information. The ratings, target prices, earnings forecasts, and related assessments cited herein reflect the views of the respective broker's analysts, representing only their institutional standpoint, not those of TechFlow Post, and do not constitute any investment advice.

Markets carry risks, and investment decisions must be made independently. This article should not serve as a basis for buying or selling any securities.

Join the official Coincamps community:

X: https://x.com/coincamps

Telegram: https://t.me/coin_camps