This chip stock is tumbling. Bullish Evercore ISI thinks it will be worth $1 trillion one day

Arm shares are falling despite a better-than-expected fiscal Q4, as investors focus on supply constraints that could limit its ability to meet surging AI-driven chip demand. Evercore ISI reiterated outperform and lifted its 12-month price target to $326 from $227, arguing Arm could ultimately reach a $1 trillion valuation; the stock would be worth nearly 5x its current ~$209 billion market cap at that level. Morgan Stanley, Deutsche Bank and Barclays highlighted advanced-node wafer availability at TSMC as the key near-term bottleneck, with Arm still guiding to about $1 billion of revenue in FY2027-2028 tied to a roughly $2 billion data-center chip opportunity.

Analysis

The market is treating this as a supply-constrained story, but the more important implication is that Arm’s near-term revenue is becoming a function of who controls advanced-node capacity rather than who has the strongest product architecture. That shifts bargaining power toward TSMC in the near term and makes every incremental capacity allocation decision at 3nm/2nm disproportionately important for ARM’s upside path, while also indirectly tightening the competitive gap for smaller CPU designers that lack ecosystem pull.

The bigger second-order winner is actually the foundry and equipment complex over a multi-quarter horizon: if ARM’s design wins are real but shippable volume lags, capital spending in the supply chain can still rise before Arm monetizes fully. That creates a gap where order growth accrues to TSM/ASML-style beneficiaries first, while ARM’s reported revenue may lag the demand curve for 2-4 quarters, which is exactly the sort of setup that can frustrate momentum holders and create sharp post-earnings volatility.

Consensus appears to be underappreciating how much of the current debate is timing rather than thesis. If supply is the bottleneck, then any incremental improvement in wafer availability or a clearer allocation commitment from TSMC can re-rate ARM quickly; but if capacity remains tight, the stock can de-rate even as the long-term TAM story stays intact. In other words, this is a classic quality-vs-execution trade: long-duration upside remains high, but the next 1-2 quarters are likely to be driven by how much of the backlog can be physically manufactured, not by headline demand.

The contrarian read is that the market is discounting too much of the long-term embedded option value and too little of the near-term choke point. That usually argues for buying weakness only when paired with a defined catalyst window, not chasing strength into a bottleneck narrative. The setup also leaves room for relative-value trades: ARM can outperform on any supply clarity, while TSMC can underperform if investors conclude it is the gating variable and not fully capturing pricing power from that position.

Key Decisions for Investors

• Buy ARM on further weakness only if it holds above the post-earnings gap for 2-5 trading days; use a 3-6 month horizon and target a 20-30% rebound if supply commentary improves, with a stop if management again refuses to tighten FY27/28 guidance.
• Initiate a pair trade: long ARM / short TSM at roughly 1:1 notional for a 1-3 month window, betting that demand-backed re-rating at ARM outpaces the foundry bottleneck discount at TSM if capacity remains the gating factor.
• If looking for cleaner expression of the bottleneck, buy TSM put spreads or short-dated downside hedges into the next capacity/allocation update; payoff is best if investors start pricing in delayed ARM monetization over the next 4-8 weeks.
• Avoid chasing the megacap-quality narrative in EVR alone; the better risk/reward is to own the analyst-confirmation trade in ARM only after a supply data point, since the current move is more about delivery timing than thesis failure.
• For longer-duration accounts, accumulate ARM on a 6-12 month view only if valuation resets into a 15-20% drawdown, because the optionality to a $2B+ opportunity remains intact but near-term execution risk is high.