Arm stock pops 6% as CEO Haas issues $15 billion revenue expectation for new chip

Arm forecasted 2031 revenue of $25B with roughly $15B coming from its newly announced in-house Arm AGI CPU and projected EPS of $9, versus just over $4B in annual revenue in 2025. The chip, unveiled with Meta as the initial customer, reportedly carries ~50% gross margin; shares rose ~6% in after-hours trading (having closed ~1.5% lower earlier). This represents a major strategic shift from licensing/royalties to direct chip sales, creating a large high-margin revenue opportunity and potential competitive disruption in the data-center CPU market.

Analysis

A move from pure-IP to also offering finished silicon changes Arm’s revenue durability and multiple regime: recurring royalty economics would partially convert into product revenues with higher unit gross margins but material cyclicality, inventory risk and working-capital needs. That shift implies a valuation re-rate trade-off — less of a subscription-like multiple and more of a mixed hardware-software multiple — which will take 12–36 months of cadence and transparent margin disclosure to arbitrate in the market.

Competitive dynamics will bifurcate by customer type. Large hyperscalers with internal SoC roadmaps retain option value to double down on bespoke designs, forcing Arm’s addressable hardware customers to skew toward mid-tier cloud, telco, and enterprise vendors that lack in-house ASIC teams; that increases TAM but simultaneously creates channel conflict with existing licensees and raises the probability of churn or contract renegotiations. Secondary beneficiaries include foundries, EDA/IP vendors and specific memory suppliers if the new workloads favor high-bandwidth memory stacks; incumbents with integrated fabs or differentiated software stacks can blunt share loss by competing on system-level throughput rather than pure CPU IPC.

Key tail risks are customer retaliation or accelerated in‑house designs by big cloud buyers, regulatory scrutiny around selling a product to your own licensees, and supply constraints at high-end fabs that could cap adoption. Short-term re-rates will be driven by partner adoption announcements and benchmark transparency (days–weeks), while true royalty and share-shift effects play out over multiple years (12–48 months).

Key Decisions for Investors

• Long ARM equity or 24–36 month LEAP calls on a meaningful pullback (buy into weakness; target 2x upside if adoption signals materialize within 18–24 months). Hedge: sell a 12–18 month covered call to finance 30–40% of premium. Downside: regulatory/customer pushback could compress value by 30–50%.
• Pair trade — long ARM / short AMD (or short INTC) over 6–18 months: size the short ~50–75% of the long notional to capture asymmetric upside from TAM expansion while hedging market-wide semiconductor cyclicality. R/R: if ARM wins share among mid-tier buyers, expect outsized alpha; if hyperscalers in‑source, the pair reduces market beta.
• Long MSFT or GOOGL 9–18 month call spreads to capture incremental demand for inference infrastructure without paying full premium for outright calls (buy 12–18 month call, sell higher strike). R/R: capped upside but lower theta burn; risk is in‑sourcing by large cloud customers which would mute incremental purchases.
• Event hedge: buy 6–12 month put protection on incumbents (INTC/AMD) sized to 25–40% of semiconductor exposure if customer announcements or independent benchmarks show superior price/perf of new entrants. This protects against rapid share reallocation in data‑center AI inference.