Broadcom shares jump before the bell as chipmaker agrees Google and Anthropic deals

Broadcom will produce future AI chips for Google and signed an expanded deal with Anthropic giving the startup access to about 3.5 gigawatts of computing capacity; shares rose 3.7% in premarket trading. The stock is down ~10% YTD despite a blowout March earnings print and CEO Hock Tan's projection of AI chip revenue in 2027 "significantly in excess of $100 billion"; Citi maintains a Buy and models revenue rising above $130 billion from the Google deal.

Analysis

This is less a pure product win and more a demand-visibility event that de-risks a multi-year revenue cadence for a company that sells bespoke silicon and high-margin networking/firmware stacks. If deployments follow a typical hyperscaler phasing, expect 12–36 month lumpy revenue recognition (NRE up-front, volume later) and a step-function margin expansion as ASPs and IP licensing shift from engineering spend to scale production; modelers should assume a 2–3 year lag between contract signing and steady-state contribution to gross margin.

Second-order winners include HBM suppliers, advanced packaging providers and TSMC capacity holders — constrained HBM/advanced-node supply can bottleneck full-box throughput long before system software becomes the limiter, creating outsized pricing power for wafer and substrate vendors in 2025–2027. Conversely, smaller custom-AI ASIC vendors lose optionality: being squeezed on price vs. large-scale, co-designed solutions and on access to hyperscaler deployments.

Key risks that can reverse the positive momentum are execution (chip/system qualification slippage), software lock-in (TPU-like stacks that limit third-party ISV adoption), and macro/geopolitical shocks that disrupt advanced-node supply chains; any two of these occurring within 12 months materially compresses the consensus multi-year upside. Finally, antitrust optics of deep hyperscaler-design partnerships and the timeline for customer migration away from incumbent accelerators are underestimated by the street — the market is pricing broad victory but not the multi-year cadence or regulatory noise that could compress multiples.

Key Decisions for Investors

• Buy AVGO stock on weakness with a 12–24 month horizon: target +40–60% if revenue ramps toward the upper-end 2027 scenarios; initial position size 3–6% of liquid portfolio, stop-loss -15% to control execution risk tied to deployment delays.
• Buy a long-dated AVGO call spread (expiry 12–24 months): pay-for 25–45% OTM call spread sized to cost ~3% portfolio notional to capture asymmetric upside if hyperscaler deployments scale; max loss = premium, potential payoff >3x if consensus revenue growth materializes.
• Pair trade: long AVGO / short NVDA (0.5–0.7x notional) over 6–18 months to isolate exposure to hyperscaler-custom silicon vs GPU-driven SaaS AI demand; hedges broad multiple compression while keeping upside if Broadcom wins share of custom silicon designs.
• Sell short-dated (1–3 month) put protection rather than outright short for small pure-play AI accelerator vendors lacking hyperscaler contracts: collect premium to benefit from temporary nervousness, but limit exposure because execution or a surprise hyperscaler pivot can gap those names higher.