Can Credo's $750 Million Optics Bet Turn Copper's Limits Into Your Portfolio Gains?

AI data centers are driving a structural shift in interconnects, with copper demand near 27 tons per megawatt and a projected 320,000-ton global copper deficit in 2026. For Credo Technology, fiscal Q3 2026 revenue hit $407 million, up 201.5% year over year, while the company also announced a $750 million cash-plus-stock acquisition of DustPhotonics to expand into silicon photonics and longer-reach optical solutions. Management said the deal could support fiscal 2027 revenue growth above 75% and make optical revenue exceed $500 million, though execution risk remains high.

Analysis

The market is starting to price the wrong bottleneck. Copper is still the near-term tollbooth for AI infrastructure, but the more durable monetization sits with vendors that can intermediate the transition from short-reach electrical connectivity to optical fabrics. That favors CRDO because it is one of the few names that can sell both the bridge solution and the end-state solution, which should deepen wallet share at hyperscalers and reduce the risk of being displaced by a pure-play optics incumbent.

The DustPhotonics deal is strategically important because it moves CRDO from component supplier to architecture enabler. That matters in procurement cycles: once a platform is designed around a vendor’s electrical stack, the same vendor has a credible path to supply the optical layer, lowering qualification friction and increasing attach rates. The second-order effect is negative for smaller niche optics vendors and potentially for pluggable transceiver suppliers whose margins may compress if hyperscalers accelerate vertical integration and dual-source only at the silicon level.

The bigger risk is timing mismatch. AI cluster operators can tolerate a copper bridge for a few quarters, but the optical ramp is likely to be lumpy and subject to customer validation cycles, so CRDO’s multiple is vulnerable if integration slips or if revenue mix shifts slower than the market expects. In that scenario, the stock could de-rate even if fundamentals remain strong, because the valuation already embeds a clean execution path into fiscal 2027.

Contrarianly, the market may be underestimating how long copper remains relevant. Optical is the destination, but rack-level and short-reach interconnects will not disappear quickly, meaning CRDO’s copper franchise may stay structurally valuable longer than the debate implies. That creates a favorable asymmetric setup: downside is capped by an already-proven AEC business, while upside comes from optionality on a larger optical TAM and customer consolidation around a single trusted supplier.

Key Decisions for Investors

• Long CRDO on pullbacks over the next 2-6 weeks; use weakness into post-deal digestion as the entry point. Risk/reward favors owning the bridge-to-optics story with a multi-quarter horizon, but size smaller than a normal growth long because the multiple already discounts execution.
• Pair trade: long CRDO / short a basket of pure-play optical beneficiaries with weaker balance sheets or less software/control-plane differentiation over 3-9 months. The thesis is that hyperscalers will prefer vertically integrated suppliers that can span electrical and optical layers, compressing the value of fragmented point solutions.
• Sell downside protection on CRDO via cash-secured puts or put spreads into earnings if implied vol remains elevated. The company’s cash position and copper franchise provide a fundamental floor, while upside from optical attach can re-rate the stock if integration milestones land.
• Maintain a tactical long FCX only as a slower-moving hedge against continued data-center copper demand, but fade strength versus CRDO. FCX benefits from the commodity cycle, yet it lacks the structural software/architecture optionality that drives multiple expansion.
• Avoid chasing the optics complex broadly; prefer waiting for evidence that CRDO’s 2027 guidance is translating into design wins. The cleanest risk/reward is company-specific rather than a sector basket, because execution dispersion should widen as the market separates platform winners from commodity suppliers.