Morgan Stanley reiterates Overweight on Broadcom and raises its target price to $505 after a solid quarter and higher FY28 estimates.

Overview

This earnings review assesses Broadcom's quarterly results, AI growth outlook and valuation. Morgan Stanley reiterates Overweight and raises its price target to $505, arguing that strong AI, networking and software execution more than offset concerns about elevated near-term expectations and limited visibility into 2028.

Core views

Broadcom's reported quarter was modestly ahead of expectations. Revenue of $29.591bn rose 33.4% quarter-on-quarter and 64.3% year-on-year, exceeding the Street's $29.241bn and Morgan Stanley's $29.404bn estimate. Semiconductor Solutions generated $20.839bn, up 38.8% quarter-on-quarter and 127.4% year-on-year, while Infrastructure Software produced $8.752bn, up 21.9% sequentially and 29.0% year-on-year. Gross margin was 75.0%, above the Street's 73.5% and Morgan Stanley's 74.0%, and non-GAAP EPS of $3.32 exceeded the Street's $3.22 and Morgan Stanley's $3.24 forecast. Next-quarter revenue guidance of $34.8bn was also above the Street's $34.657bn and close to Morgan Stanley's $34.844bn estimate. The central earnings driver remains AI. Management indicated 2027 AI revenue of $115bn, slightly below Morgan Stanley's prior $120bn expectation but consistent with prior guidance of "well above $100bn." Management also said AI revenue would double again in 2028. Morgan Stanley notes that second-half AI revenue will more than triple from the prior year and considers the scale of growth remarkable. It nevertheless continues to haircut its CY28 assumptions because it does not believe semiconductor visibility can reliably extend that far; even with that conservatism, its FY28 revenue estimate rises materially to $248bn from $214bn. The report distinguishes between strong underlying execution and a potentially difficult near-term expectations setup. Some bulls had anticipated more than $150bn of 2027 revenue based on a 10 GW forecast and higher dollar-per-GW assumptions, whereas management's more specific guidance is $115bn. These higher external expectations have also created concern about Broadcom's TPU-related market share. Morgan Stanley argues that such optimistic assumptions are unnecessary for a positive investment case: customer breadth is improving, with two AI labs expected to be the two largest customers in CY28, and the company has six XPU customers, four of them very large. Two new architectures are ramping, which introduces initial-production uncertainty, but the firm views expansion beyond TPUs as a positive for concerns about TPU share shifts. Networking remains strong, while the margin outlook is framed as a mix effect rather than a deterioration in economics. Morgan Stanley expects gross-margin percentages to face pressure from higher high-bandwidth-memory pass-through content, but notes that higher memory prices can reduce the percentage while marginally increasing gross-profit dollars. It expects robust growth in both ASICs and GPUs, although it cautions against drawing broad conclusions from near-term supply-constrained results. Over the past two years, NVIDIA data center, AMD GPU and Broadcom ASIC businesses have each grown by roughly three times; over the next 12 months, the report expects supply-chain constraints to shape growth more than demand. Morgan Stanley expects NVIDIA's Vera Rubin to be the most desired product in the near-term window and believes that its ramp will not satisfy all end demand, leaving Blackwell and comparable products positioned to perform well. For Broadcom, the report remains positive on continued remarkable ASIC growth but is more skeptical of the narrative that cloud vendors can resell custom ASICs to other hyperscale customers at a markup, because the report does not see a cost-per-token advantage versus Rubin. It nonetheless believes Broadcom can maintain more than 80% market share of its serviceable addressable markets over time, supported by its industry contacts, despite expected competition. The longer-term thesis also includes a cyclical recovery in non-AI semiconductors and VMware execution. Morgan Stanley expects a late-next-year rebound in core semiconductor businesses after excess inventory in networking and storage, and expects VMware to integrate successfully, focus operations, cut costs and generate stable cash flows. The report identifies Broadcom's large absolute AI exposure, networking and ASIC growth, and potential VMware upside as reasons it can sustain a premium multiple. Morgan Stanley increases FY26 revenue/non-GAAP gross margin/EPS estimates to $106bn/75.1%/$11.66 from $106bn/74.8%/$11.59. For FY27, it models $167.9bn/70.7%/$18.27, versus $167.4bn/70.8%/$18.28 previously, leaving the top line broadly unchanged until the FY28 increase. The price target rises from $502 to $505 because the modest estimate revision is applied to the unchanged 28x ModelWare EPS multiple, roughly 26x non-GAAP EPS. The report values Broadcom at 28x CY2027e ModelWare EPS of $18.04, or about 26x non-GAAP EPS of $19.63, and considers this in line with or below AI peers. It argues that Broadcom and NVIDIA remain among the most attractively valued names in its coverage. The report's scenario framework shows a $640 bull case based on upside revenue surprises, new XPU-customer ramps, networking strength, VMware synergies and multiple expansion as confidence in AI and M&A strategy increases. Its $505 base case assumes strong CY27 revenue growth from XPU, networking and software. Its $300 bear case assumes revenue disappointment, underwhelming synergies, failure of new customer engagements to reach production, and a valuation multiple falling to about 19x 2027e ModelWare EPS of $15.50.

Analysis framework

Morgan Stanley compares reported revenue, segment results, gross margin and non-GAAP EPS with Street and internal estimates, then incorporates management guidance into its fiscal forecasts. It evaluates AI revenue, customer breadth, product ramps, supply constraints, market-share expectations, networking, margins and VMware integration before applying a forward earnings multiple to derive the price target and presenting bull, base and bear scenarios.

Methodology notes

• Valuation methodsP/E and PEG Valuation

  Forward price-to-earnings valuation using ModelWare and non-GAAP EPS.

  Morgan Stanley values Broadcom at 28x CY2027e ModelWare EPS of $18.04, roughly 26x non-GAAP EPS of $19.63, to support its $505 price target.

• Industry AnalysisSupply-demand framework

  AI semiconductor growth assessed through end demand, product ramps and supply-chain constraints.

  The report compares ASIC and GPU growth, discusses constrained supply and expected demand for Vera Rubin, and treats supply-chain factors as a key determinant of growth over the next 12 months.

• Event-Driven and Behavioral FinanceExpectation Gap and Expectation Management

  Comparison of market expectations with management guidance and internal forecasts.

  The report highlights that some market participants expected over $150bn of 2027 revenue while management guided to $115bn, creating near-term uncertainty despite strong operating results.

Asset mapping & comparison

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

• Broadcom Inc. (AVGO)

  Primary covered company; the report sees AI, networking, ASIC growth and VMware integration as key drivers.

  Strengths

  Revenue and EPS beat expectations; broadening XPU customer base; strong networking; expected AI and software growth.

  Weaknesses

  Gross-margin percentage pressure from memory pass-through content and limited visibility into CY28.

  Comparison

  The report considers Broadcom's valuation broadly in line with or below AI peers and views AVGO and NVIDIA as among the most attractively valued names.

  Risks

  Networking share loss to NVIDIA/Mellanox, uncompetitive ASICs, customer losses, and weaker VMware acquisition execution.

• NVIDIA Corp. (NVDA)

  Competitor and valuation comparison within AI semiconductors.

  Strengths

  Morgan Stanley expects strong demand for Vera Rubin and believes the ramp cannot meet all end demand.

  Comparison

  The report expects NVIDIA, AMD GPU and Broadcom ASIC businesses to have each grown about three times over two years; it also compares AVGO favorably with NVIDIA on valuation.

• Advanced Micro Devices (AMD)

  Competitor and comparison in AI GPUs.

  Comparison

  AMD GPU growth is compared with NVIDIA data center and Broadcom ASIC growth over the past two years.

Key data

• Quarterly revenue$29.591bnUp 33.4% quarter-on-quarter and 64.3% year-on-year; above Street $29.241bn and Morgan Stanley $29.404bn.
• Semiconductor Solutions revenue$20.839bnUp 38.8% quarter-on-quarter and 127.4% year-on-year.
• Infrastructure Software revenue$8.752bnUp 21.9% quarter-on-quarter and 29.0% year-on-year.
• Gross margin75.0%Above Street 73.5% and Morgan Stanley estimate of 74.0%.
• Non-GAAP EPS$3.32Above Street $3.22 and Morgan Stanley estimate of $3.24.
• Next-quarter revenue guidance$34.8bnAbove Street $34.657bn and close to Morgan Stanley's $34.844bn estimate.
• 2027 AI revenue commentary$115bnSlightly below Morgan Stanley's prior $120bn forecast; management says revenue will double again in 2028.
• FY28 revenue estimate$248bnRaised from $214bn previously.
• Price target$505Raised from $502 using the same 28x ModelWare EPS multiple.

Impact & implications

Morgan Stanley believes Broadcom's earnings, AI growth runway, networking resilience and potential VMware benefits support its Overweight rating and higher target price. It also flags that the shares may face nearer-term uncertainty if elevated market expectations for 2027 AI revenue reset toward management's $115bn outlook.

Risks

• Near-term upside may be limited if market expectations for 2027 AI revenue remain above management's $115bn guidance.
• New XPU architectures carry initial-production uncertainty.
• Broadcom could lose networking share to NVIDIA/Mellanox.
• ASIC products could prove uncompetitive or customers could shift to competitors.
• VMware acquisition integration and expected synergies could underwhelm.
• Revenue growth could disappoint, new customer engagements could fail to reach production, and the valuation multiple could decline.

What to watch

• Progress of two new architecture ramps and customer engagements reaching production.
• AI revenue trajectory against management's $115bn 2027 commentary and its statement that 2028 revenue will double again.
• Broadening of the XPU customer base and the role of AI labs among CY28 customers.
• Networking momentum, high-bandwidth-memory effects on gross margins and gross-profit dollars.
• Evidence of non-AI semiconductor recovery in networking and storage markets.
• Realization of VMware integration, cost-cutting and cash-flow synergies.