Meta’s multi-GW partnership strengthens visibility of Broadcom’s AI revenue growth

Overview

This report focuses on the investment implications of Broadcom Inc and Meta expanding cooperation on AI custom chips. J.P. Morgan notes that Meta will deploy MTIA ASICs and related networking through 2029 under an agreement with multi-year, multi-generation, and multi-GW characteristics, further confirming strong hyperscaler demand for custom chips and joint design of data center networking. The report reiterates AVGO’s Overweight rating and says it remains a leading pick in semiconductors.

Core views

The core view is that Broadcom is benefiting from a trend of hyperscalers and OEMs shifting to custom ASICs. The Meta partnership spans multiple MTIA generations, including Athena, Iris, Arke, Astrid, and Apollo, with expected ramp for five projects over the next three years; the first 1GW deployment is expected in 2027, with roughly $12–$15B in revenue opportunity for Broadcom. This transaction together with partnerships with Google, Anthropic, and OpenAI supports confidence that Broadcom’s FY27 AI revenue will be materially above $120B.

Analysis framework

The report uses an event-driven company research framework, comparing Meta’s new deal with Broadcom’s existing collaborations with Google, Anthropic, and OpenAI to assess custom ASIC design cycles, deployment pacing, network attach revenue, degree of customer lock-in, and multi-year revenue visibility. It also incorporates J.P. Morgan’s estimates for EPS, revenue, valuation multiples, stock performance, and style factors to form the rating and target price view.

Methodology notes

• Valuation methodsforward_eps_multiple

  Forward EPS multiple valuation

  The Dec-26 target price assumes AVGO trades around 24.5x, broadly consistent with peers at 23–25x, applied to CY27 EPS of about $20.45.

• industry_analysiscustom_silicon_design_in

  Custom ASIC multi-generation design-in

  The report treats Meta’s multi-generation MTIA roadmap as a source of revenue visibility because complex ASIC programs typically require years of joint design, validation, and deployment, with high customer switching costs.

• growth_driverai_compute_cluster_scaling

  AI compute cluster scaling

  The report says that to scale AI compute clusters efficiently, chips, networking, and software frameworks must be closely coordinated, and Broadcom’s combined strength in ASICs and network silicon therefore benefits.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Broadcom Inc (AVGO.US)

  Primary beneficiary

  Strengths

  Has scale and technical capability across Wireless, data center networking, AI/deep learning ASIC, storage, infrastructure chips, hardware, and software; maintains multi-GW collaborations with clients such as Meta, Google, Anthropic, and OpenAI.

  Weaknesses

  Valuation already reflects elevated growth expectations, and revenue realization depends on customer AI capex, project ramping, and execution across multiple ASIC generations.

  Comparison

  The report says the 24.5x valuation multiple is broadly consistent with the peer range of 23–25x; 12-month absolute performance was 113.5%, and relative performance was 84.6%.

  Risks

  Macroeconomic or regional demand shocks, customer deployment delays, complexity in AI custom-chip projects, competition, and valuation compression.

• Meta Platforms Inc (META.US)

  Key customer and partner

  Strengths

  Accelerates self-built AI accelerator deployment through the MTIA ASIC roadmap, with plans to expand multi-generation chips and networking collaboration.

  Weaknesses

  The report does not provide a standalone rating or target price for META; its role is mainly a driver of Broadcom’s revenue visibility.

  Comparison

  The report compares the Meta partnership with Google, arguing that as design cadence accelerates and project complexity rises, deep co-design should support partnership durability.

  Risks

  If Meta adjusts AI infrastructure investment pace or MTIA deployment plans, Broadcom’s related revenue realization could be affected.

Key data

• RatingOverweightJ.P. Morgan reiterates an Overweight rating on AVGO.
• Target Price$500.00Dec-26 target price.
• Current Price$380.78AVGO price on April 14, 2026.
• Implied Upsideabout 31.3%Calculated from target price $500.00 and stock price $380.78.
• Initial Meta DeploymentMore than 1GWThe report says the initial 2027 commitment exceeds 1GW, representing the first phase of subsequent multi-GW deployments.
• 2027 Revenue Opportunity$12–$15BRelated to the first 1GW MTIA deployment, including networking revenue.
• AI Revenue ViewFY27 above $120B+The report says the Meta transaction further strengthens confidence in materially exceeding this level.
• CY27 EPS Estimateabout $20.45Used for target price valuation.
• Valuation Multiple24.5xGenerally aligned with a peer range of 23–25x.
• Projected FY27 Adj. EPS$19.61Summed quarterly forecasts shown in the chart.
• FY27E Revenue$166,491 mnCompany estimated revenue shown in the chart.
• Market Cap$1,892,096 mnAs disclosed in company data table.

Impact & implications

The investment implication is constructive: Meta collaboration expands Broadcom’s AI ASIC revenue from a single-customer or single-generation path toward a multi-customer, multi-generation, multi-GW growth trajectory. If deployment progresses as expected, Broadcom could gain not only ASIC revenue but also associated upside from networking chips such as Tomahawk and Jericho, improving revenue elasticity and visibility over the AI data center capex cycle.

Risks

• Unexpected negative shocks to global or regional end-market demand, altering Broadcom product demand.
• AI ASIC project development is complex and has long cycle times; if Meta MTIA or other customer projects are delayed, revenue timing could be affected.
• High growth expectations make valuation sensitive; if FY27 AI revenue or EPS falls short of expectations, support for the target price may weaken.
• Customer concentration in hyperscaler-scale projects means changes in capex cycles can amplify revenue volatility.
• Competition in custom chips, networking silicon, and AI infrastructure may pressure share, pricing, or margins.

What to watch

• The actual ramp pace of the five-generation Meta MTIA program: Athena, Iris, Arke, Astrid, and Apollo.
• Whether the first 1GW MTIA deployment in 2027 occurs on schedule and realizes the $12–$15B revenue opportunity.
• Incremental orders, deployment scale, and revenue recognition from other GW-scale ASIC collaborations with Google, Anthropic, OpenAI, and others.
• Whether Broadcom’s FY27 AI revenue comes in meaningfully above $120B+.
• Subsequent revisions to CY27 EPS of about $20.45 and FY27E revenue and margin forecasts.
• Changes in partnership governance and related-conflict disclosures after Hock Tan shifts from Meta board member to advisor.