Earnings Call Highlights
Broadcom's Q2 FY2026 call was a triumphant, demand-driven narrative with Hock Tan projecting extraordinary confidence in AI semiconductor momentum that bordered on euphoric, with virtually no hedging on the trajectory through 2028. Management's posture was assertive and expansive, anchoring a $100B+ AI semiconductor revenue target for FY2027 and explicitly extending visibility into FY2028 for the first time, a notable shift from prior quarters. Analysts were broadly constructive, probing for supply constraints, gross margin durability, and customer concentration rather than questioning the demand thesis itself. The dominant theme was the transition from chip supplier to compute capacity platform architect, with Broadcom positioning itself as the infrastructure backbone for frontier AI labs at gigawatt scale.
- AI semiconductor revenue hit a record $10.8B in Q2, up 143% YoY, with Q2 bookings of over $30B against that shipment figure — Hock Tan noted that customer visibility has extended from 2027 to 2028 in just the past quarter, driven by hyperscalers and frontier labs placing long-lead orders to secure power, HBM, and chip supply simultaneously.
- Broadcom guided Q3 AI semiconductor revenue to $16B (up 200%+ YoY) and full-year FY2026 AI revenue to $56B, with FY2027 reiterated at 'in excess of $100 billion' — and Tan explicitly stated FY2028 will be 'substantial growth' from FY2027, the first time 2028 has been framed as a growth year rather than a placeholder.
- Broadcom announced a new AI XPU platform with Apollo, Blackstone, and other investors to deploy 20+ gigawatts of compute capacity through 2029, with a first tranche valued at $35B being launched by Apollo — this positions Broadcom not just as a chip designer but as a financier and capacity orchestrator for frontier labs like Anthropic and OpenAI who need gigawatt-scale compute they cannot self-fund.
- On networking, Tan confirmed the 100-terabit Tomahawk 6 has been shipping for over a year and a 200-terabit next-gen switch is taping out this quarter, while CPO with 1.6T DSPs and co-packaged lasers is described as 'the de facto standard' — networking represented ~40% of Q2 AI revenue, which Tan characterized as near the ceiling, with a normalized mix closer to 30% going forward.
- Gross margin is compressing toward ~74% in Q3 as AI semiconductor mix grows within the overall revenue base, but Tan and Spears were emphatic this is a mix effect rather than structural semiconductor margin deterioration, with operating margin holding flat at 67% — a key reassurance for investors modeling the long-term profitability of the AI ramp.