BofA raises Arm Holdings stock price target on AI CPU potential

BofA raised its price target on Arm Holdings to $245 from $180 while keeping a Neutral rating, citing solid quarterly results and improving AI/data center share gains. The stock trades at $237.30, up 117% year-to-date and near its 52-week high of $239.50, with BofA modestly lifting fiscal 2026-2027 estimates on stronger CPU content assumptions. Management still sees AGI chiplet CPU demand roughly double current supply, but BofA remains cautious on Arm’s ability to capture the full 15% market share target by FY2031.

Analysis

The market is starting to price Arm less as a pure IP royalty compounder and more as a platform with optionality on AI server CPUs and chiplets. That transition matters because it changes the earnings mix from highly visible licensing/royalty annuities to a more competitive, capital- and execution-sensitive model, which should cap how far multiples can extend even if headline growth stays strong. In other words, the stock can keep working on estimate revisions, but the valuation ceiling rises only if Arm proves it can convert design wins into durable socket share without margin dilution.

The second-order read-through is more important for the AI supply chain than for Arm itself: every incremental Arm server win potentially pressures x86 incumbents and forces hyperscalers to keep multi-sourcing, which lowers the odds of a winner-take-all CPU standard. That dynamic favors the buyers of compute, not just the supplier, because diversified CPU sourcing improves negotiating leverage and should keep AI infrastructure capex from becoming vertically locked to a single architecture. It also means the biggest surprise could be on content per server rather than unit share — small gains in attach rates can still drive meaningful royalty expansion if AI server volumes compound fast enough.

Near term, the stock is vulnerable to a classic expectation-overhang setup: the move has outrun the proof point, and any evidence that AGI chiplets remain a niche within a constrained customer set will hit the multiple before it hits the top line. The next catalyst window is the next two earnings prints, when investors will focus less on demand rhetoric and more on whether supply constraints are truly demand-constrained or just under-shipments. If commentary shifts toward slower design-win conversion, the recent rerating could unwind quickly despite decent reported numbers.

The contrarian case is that consensus may be underestimating how much of the upside is already embedded in the “AI CPU share” narrative. A 69x outer-year multiple assumes not just growth but sustained strategic relevance, yet the customer base is still concentrated and buyers are sophisticated enough to dual-source around any single vendor’s economics. That makes the current setup more attractive for volatility monetization than outright chase buying.

Key Decisions for Investors

• Sell upside in ARM via 1-2 month call spreads into strength; use strikes above the current market where implied growth assumptions look most stretched. Risk/reward favors premium capture over directional chase after the multi-day gap.
• If long AI infrastructure, prefer a basket tilted to compute buyers over the CPU supplier: pair long META and MS against a partial short/underweight in ARM over the next 3-6 months. Thesis: hyperscalers benefit from architectural optionality and bargaining power while ARM faces execution risk on share conversion.
• For traders wanting ARM exposure, wait for a post-earnings pullback or a failed breakout rather than buying the open. Entry on a 5-8% retrace offers better asymmetry because the stock is already pricing near-perfect execution.
• Monitor the next two quarterly calls for language around multi-sourcing, customer concentration, and ship-ship timing; if management softens on conversion rates, rotate out immediately. The stock likely trades on guidance tone more than reported EPS from here.