Workspaces
Log in Sign up

Broadcom's Q2 FY2026: Reaffirming a $100 Billion AI Runway Amid a Sharp Market Correction

Updated

Sources

Broadcom's Q2 FY2026: Reaffirming a $100 Billion AI Runway Amid a Sharp Market Correction

On June 3, 2026, Broadcom (AVGO) reported its fiscal Q2 2026 earnings, delivering blowout results that highlighted the "insatiable" demand for AI infrastructure. The company's performance provides a critical secondary anchor for the broader AI capex cycle, showing that spending is not just concentrated in Nvidia GPUs but is flowing heavily into custom silicon (ASICs) and AI networking.

Blowout Financials & "Insatiable" Demand

Broadcom posted record fiscal Q2 revenue of $22.19 billion (up 47.9% YoY), with an operating margin of 49.0% (record adjusted operating margin of 67%) and free cash flow of $10.26 billion.

CEO Hock Tan described the demand environment in stark terms:

"demand for the company's artificial intelligence semiconductors remains 'simply insatiable.'"

AI semiconductor revenue surged 143% year-over-year to a record $10.8 billion, representing 49% of Broadcom's total revenue. Crucially, forward-looking demand signals were exceptionally strong, with quarterly AI semiconductor bookings exceeding $30 billion—representing a nearly 3x book-to-bill ratio and providing massive backlog growth.

Upgraded Guidance & Multi-Year Outlook

Following the strong quarter, Broadcom sharply raised its near-term and long-term outlook:

  • Q3 FY2026 AI Semiconductor Guidance: Raised to $16 billion, representing over 200% YoY growth.
  • Full-Year FY2026 AI Semiconductor Guidance: Raised to $56 billion, up approximately 180% from fiscal 2025.
  • FY2027 AI Semiconductor Outlook: Reaffirmed to exceed $100 billion, with Tan stating the trajectory is "very much on track, if not stronger."
  • FY2028 and Beyond: Engagements are already underway with six core custom silicon customers, ensuring multi-year visibility.

Contractual Gigawatt Commitments

In an unprecedented disclosure, Hock Tan detailed the multi-year contractual agreements and power capacity allocations (expressed in gigawatts) with Broadcom's primary custom silicon and TPU customers1:

  1. Google: A long-term agreement signed in April 2026 to develop and supply multiple generations of TPUs and AI networking.
  2. Anthropic: Agreements for TPU-based compute access, including more than 1 gigawatt of capacity in 2026 and another 5 gigawatts beginning in 2027. This directly connects to Anthropic's landmark $35 billion private-credit financing package Broadcom, Apollo, and Blackstone Establish Landmark $35 Billion AI XPV Platform for Anthropic's 1GW Buildout.
  3. OpenAI: Broadcom has delivered initial silicon and remains on track for production in late 2026. The contract includes a commitment to deploy 1.3 gigawatts in 2027 as part of a larger 10-gigawatt agreement by 2029.
  4. Meta: A partnership to deploy 3 gigawatts through 2028 involving Meta's MTIA XPUs, with an initial 1-gigawatt order scheduled to begin delivery in the second half of 2027.

These gigawatt-scale commitments underscore the physical reality of the AI capex boom, where hardware deployment is directly gated by power availability, and major AI labs are locking in capacity years in advance.2345

Networking and Software Moats

Networking represented nearly 40% of Broadcom's Q2 AI revenue, driven by Ethernet switch and optical interconnect demand. While Tan expects this to normalize closer to 30% over time as custom XPUs ramp, Broadcom maintains a clear generation of leadership in rack-scale connectivity. Meanwhile, Infrastructure Software revenue rose 9% to $7.2 billion, supported by the successful deployment of VMware Cloud Foundation (VCF) 9.1 for on-premises enterprise AI workloads.

Broadcom's results confirm that the AI capex story KKR, Nvidia, and Vistra Launch $10 Billion Helix Venture to Bypass Grid Bottlenecks remains robustly supported by contracted backlogs and physical power allocations, mitigating fears of a near-term slowdown.

  1. An instance of Direct integration of power and compute bypasses grid interconnection lines. — Technology buyers are locking in multi-gigawatt power capacity allocations years in advance to anchor their physical hardware deployments. ↩︎

  2. An instance of Hardware deployment is no longer gated by silicon fabrication but by gigawatt-scale power allocations. — Major chip design agreements are increasingly expressed in terms of gigawatt power allocations rather than chip volume, proving that electricity access has replaced chip fabrication as the ultimate limiting factor. ↩︎

  3. An instance of Software-style financing collapses when computing demands the capital and power of a physical utility. — Broadcom's multi-year, gigawatt-scale client agreements underscore how the AI scaling race has transformed into a utility-style fight for physical power capacity. ↩︎

  4. An instance of Generational compute capacity is fully monetized and locked in before the silicon even exists. — Broadcom’s customers are entering multi-gigawatt supply agreements to secure hardware reservations and power allocations long before the chips are delivered. ↩︎

  5. An instance of Generational compute capacity is fully monetized and locked in before the silicon even exists. — Broadcom's massive pre-sold capacity deals with Google, Anthropic, and OpenAI demonstrate that generational compute capability is locked in prior to silicon delivery. ↩︎

Part of

This finding is an example of a pattern recurring across your work:

Backlinks

Revision history

  • Update Broadcom's Q2 FY2026 earnings note with the historic details from the June 3, 2026 report, including the $10.8B AI revenue, $30B bookings, and specific gigawatt-scale contracts with Google, Anthropic, OpenAI, and Meta.
    · by the agent
  • Update Broadcom's Q2 FY2026 note to capture the actual reported figures, the Q3 guidance details, the reasons behind the 13.8% weekly crash, gross margin compression, Google insourcing concerns, and the wider semiconductor selloff.
    · by the agent
  • Add explicit wikilinks to nvidia-q1-2027-record-financials-agentic-ai and hyperscaler-capex-surge-2026 to follow the cross-linking guidelines.
    · by the agent
  • Update Broadcom's Q2 results, the 14% post-earnings stock slide, the reasons for the guidance disappointment, Goldman Sachs' $525 price target hike, and Broadcom's key role in the $36B Anthropic TPU deal and $35B custom chip financing platform.
    · by the agent
  • Update Broadcom's earnings note with the actual June 4, 2026 market reaction (the 14% drop / $315B wipeout), the details of Hock Tan's 10GW AI chip shipping target and secured HBM supply, and the connection to Nvidia's competitive risk disclosures in its latest 10-Q filing.
    · by the agent
  • Refining Broadcom's Q2 FY2026 financial metrics, Q3 forecasts, long-term 2027 guidance, and the market's reaction.
    · by the agent
  • Add cross-reference to hyperscaler-capex-surge-2026
    · by the agent
  • Updated without a stated reason.
    · by the agent