Goldman Sachs Research Report Interpretation: Broadcom Reiterates AI Revenue Target of 230 Billion, 1GW Cluster to Generate 30 Billion Annually

Written by: Rita

Broadcom reiterated its FY27 and FY28 AI revenue targets—$115 billion and $230 billion respectively—at the Communacopia Conference, identifying supply as the main constraint. On September 8, Goldman Sachs released a key takeaways report from Broadcom’s participation, highlighting that Broadcom’s management believes that data center land, power, and facility construction are the real bottlenecks for AI deployment, while supply chain constraints are relatively manageable. Broadcom also disclosed a $35 billion XPV financing collaboration with Blackstone and Apollo, which can support over 20 gigawatts of AI computing capacity.

Broadcom expects that most of the economic value from AI will flow to leading frontier model owners. Based on current economics, a 1-gigawatt AI cluster brings in around $30 billion in annual revenue, with operating costs of about $10 billion. This value is distributed among hardware manufacturers, semiconductor suppliers, and data center operators.

AI revenue targets are clear; supply constraints are key

Broadcom reaffirmed its FY27 AI revenue target of $115 billion and FY28 target of $230 billion. Goldman Sachs noted that these targets are based on a clear assessment of customer visibility and deployment readiness. Broadcom’s differentiated capabilities in custom chip design, advanced packaging, and AI networking have enabled it to maintain and expand its share of strategic clients’ budgets.

Management specifically emphasized its long-term partnership with Google, while also disclosing the Jalapeno custom chip project with OpenAI and a strategic relationship with Anthropic. On supply constraints, Broadcom believes that data center land, power, and facility construction are the major bottlenecks for AI deployment, while supply chain constraints are relatively manageable and the company has more direct influence over them.

XPV Platform: $35 billion to accelerate AI infrastructure construction

Broadcom highlighted the XPV financing platform. The recent $35 billion collaboration announced with Blackstone and Apollo can support over 20 gigawatts of AI computing capacity. The main advantage of this structure is that it offers attractive financing solutions to emerging AI labs while limiting Broadcom’s direct financial liabilities.

XPV helps Broadcom reach a broader customer base while removing key bottlenecks in AI infrastructure expansion. For Broadcom, this is a low-risk way to expand its business, growing its client base without bearing the financial pressure of large-scale capital expenditures.

Open source vs. closed source: The watershed in value distribution

Management has a clear view on open source vs. closed source models. Over the long term, AI value will primarily flow to frontier model developers rather than open source weight models.

Broadcom supported this view with data. Industry training expenditure is around $200 billion a year, with frontier models and the open source ecosystem each accounting for about half. Frontier models contribute roughly 75% of industry revenue, with approximately $120 billion in revenue aligned with about $100 billion in costs. Open source models generate only about $30 billion in revenue despite similar investment. This significant difference in input-output ratio explains why Broadcom believes the owners of leading frontier models are the main beneficiaries of the AI economy.

Valuation and risks

Goldman Sachs maintains a Buy rating on Broadcom, with a 12-month target price of $540, based on a normalized 30x earnings multiple of $18 per share. The current share price is about $357, implying roughly 51% upside potential.

Downside risks include four factors: a slowdown in AI infrastructure spending, loss of share in the custom computing business, ongoing inventory digestion in non-AI business segments, and intensifying competition in the VMware space. Goldman Sachs believes a slowdown in AI spending is the greatest potential risk, but Broadcom’s current order book and customer visibility provide a high degree of near-term performance certainty.