Goldman Sachs believes that Broadcom's better-than-expected performance guidance, along with the company's latest comments on margin improvement, securing key production components, and adding new custom chip clients, is leading the market to reassess its valuation.
On March 4, analysts including James Schneider, Ph.D. at Goldman Sachs, stated in a report that the company's FY27 AI semiconductor revenue outlook significantly exceeds $100 billion. Combined with the raised quarterly guidance, the stock price is expected to rise substantially.
Broadcom's management indicated it has secured the supply of key components (memory, lasers, packaging) necessary to support its revenue forecasts through FY28. Furthermore, due to economies of scale and yield improvements, the company no longer anticipates that rack-level system shipments will dilute margins in the coming quarters.
Based on this, Goldman Sachs maintains its Buy rating and places the stock on its Conviction List. The firm raised its 12-month price target from $450 to $480 and significantly increased its AI semiconductor revenue forecasts for FY26, FY27, and FY28 to $60 billion, $130 billion, and $171 billion, respectively.
Guidance Exceeds Expectations Across the Board, Goldman Sachs Raises Revenue Forecasts
Goldman Sachs views the current quarter's results as largely in line with market expectations, but the Q2 guidance is notably above consensus. The company forecasts Q2 revenue of $22 billion, with AI semiconductor revenue at $10.7 billion, which Goldman Sachs says reflects accelerating demand and improved visibility into the order backlog.
Management has set higher medium-term targets, stating that AI semiconductor revenue could significantly exceed $100 billion in FY27, based on data center deployments of up to 10GW, with growth expected to continue into FY28. Goldman Sachs describes this as a significant upward revision of expectations.
Under these assumptions, Goldman Sachs has raised its FY26, FY27, and FY28 AI semiconductor revenue forecasts to approximately $60 billion, $130 billion, and $170 billion, respectively, emphasizing that this represents a material upgrade to investor models.
Margin Concerns Alleviated, Secured Supply Chain is a Key Signal
Concerns regarding margins stemmed from potential component shortages and rising costs. Goldman Sachs notes that Broadcom has proactively secured the supply of key components like memory, lasers, and packaging, sufficient to support its revenue forecasts through FY28, thereby easing downward pressure on margins.
The company has also walked back its earlier suggestion that rack-level system shipments would dilute profits, believing that larger-scale production and yield improvements can absorb the impact. Goldman Sachs expects this will help Broadcom maintain its current AI product margin structure in the medium term.
Improved Customer Transparency, New Projects Ramping Up
Goldman Sachs identifies another key positive development as increased transparency regarding the customer list. Management newly disclosed collaborations with clients such as Meta, Anthropic, and OpenAI, with six custom chip projects now expected to enter volume production phases.
The company stated that its collaboration on Google's seventh-generation TPU remains strong, with limited medium-term impact from COT (Customer-Owned Tooling). Meta's custom XPU is projected to ship multiple GWs in FY27, Anthropic plans to deploy 3GW in FY27, and OpenAI targets 1GW deployment in FY27.
Major Forecast Upgrade Translates to Valuation, Goldman Sachs Raises Price Target and Earnings Assumptions
Consequently, Goldman Sachs has "rewritten" the earnings trajectory, increasing its normalized EPS estimate from $12 to $16. Applying a 30x P/E multiple yields the new $480 price target. The firm also notes that Broadcom's cost reduction pace in AI networking and custom silicon could keep pace with Nvidia.
At the product level, Goldman Sachs is optimistic about the continued ramp of Ethernet and switching chips, with Tomahawk 6 leading and Tomahawk 7 expected to see a strong ramp-up. The firm forecasts that networking revenue will account for 40% of AI semiconductor revenue in Q2, stabilizing between 33% and 40% thereafter.
Goldman Sachs also highlights risks, including a potential slowdown in AI infrastructure spending, loss of market share in custom compute, prolonged inventory digestion in non-AI businesses, and intensifying competition in the VMware segment.
Comments