Morgan Stanley maintains Overweight on Apple; AI server chip shortcomings are the core short-term debate

Overview

This report focuses on Apple's capabilities and timeline in AI server chips, particularly the potential impact of M2 Ultra, Baltra ASIC, Private Cloud Compute, and Siri AI. Morgan Stanley believes Apple has strong vertical integration capabilities in low-power mobile and PC chips, but high-power, high-performance AI inference chips are a new challenge; if its in-house chips underperform in the near term, Siri AI responsiveness may weaken or external cloud compute costs may increase. Even so, the report maintains a positive rating on Apple, believing that the iPhone replacement cycle, resilience in services revenue, and long-term AI and other investments support medium- to long-term value.

Core views

The core views include: first, Apple's AI server silicon may face performance challenges, but supply chain checks do not show M-series capacity plans being canceled; second, the Baltra ASIC is expected to enter small-volume production in the first half of 2027, and a second-generation AI ASIC may reach volume production in 2028 and approach the power envelope of commercial AI GPUs; third, near-term attention should focus on memory costs, progress in AI features, and the performance of the iPhone 17 cycle; fourth, under the base case FY27 EPS is $10.30, with a price target of $360.00, a bull-case target of $453.00, and a bear-case target of $194.00.

Analysis framework

The report analyzes the issue using supply chain checks, TSMC packaging capacity indicators, AI ASIC production timelines, risk-reward scenarios, and relative valuation methods. On valuation, the price target is based on 9.1x FY27 EV/Sales, a multiple derived from regression against technology and consumer platform peers, corresponding to about 35x FY27 P/E.

Methodology notes

• Valuation methodsEV/Sales and P/E scenario valuation

  Derive the price target using FY27 EV/Sales and FY27 EPS/P/E, and set bull, base, and bear scenarios.

  The base case uses 9.1x FY27 EV/Sales for a price target of $360.00; the bull case uses 11x FY27 EV/Sales and about 41.1x FY27 P/E for a price target of $453.00; the bear case uses 6.3x FY27 EV/Sales and 24.6x FY27 P/E for a price target of $194.00.

• Supply chain verificationSupply chain checks and advanced packaging tracking

  Track Apple's AI chip progress through signals such as TSMC SolC, CoWoS, ASIC tape-out, and capital expenditure.

  The report treats SolC bookings, CoWoS bookings, second-generation AI ASIC tape-out progress, and Apple's capex trajectory as key indicators for judging whether Apple's AI server chips are advancing as planned.

• Risk-rewardRisk Reward scenario analysis

  Present the price target range and key triggering factors through bull, base, and bear scenarios.

  Scenario differences mainly come from the iPhone 17 cycle, services revenue growth, AI feature adoption, gross margin, memory costs, and pressure on consumer spending.

Asset mapping & comparison

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

• Apple, Inc. equity / AAPL.O / AAPL US

  Primary covered asset

  Strengths

  Strong installed base of iPhones, double-digit services revenue growth, chip and hardware-software vertical integration capabilities, and long-term growth potential in AI/payments/cloud/health/home.

  Weaknesses

  Relatively limited experience in AI server chips, questioned M2 Ultra AI server performance, and memory costs plus product mix changes may pressure gross margin.

  Comparison

  The report believes Apple leads in low-power SoC design, but unlike commercial AI GPU suppliers such as NVIDIA, it still needs to prove its capabilities in high-power AI inference chips through Baltra and the second-generation ASIC.

  Risks

  Weak Siri AI experience, greater reliance on external cloud and NVIDIA chips, an iPhone 17 cycle below expectations, slower services growth, regulatory pressure, and geopolitical risks.

• NVIDIA GPUs

  Potential substitute or external compute dependency

  Strengths

  Commercial AI GPUs have proven capabilities in high-performance AI inference workloads.

  Weaknesses

  If Apple relies more on external GPUs and cloud platforms, costs may be higher and control weaker.

  Comparison

  Apple's second-generation AI ASIC is described as potentially approaching the power envelope of merchant AI GPUs.

  Risks

  Greater dependence on external compute may affect Apple's AI service costs and differentiation.

• TSMC advanced packaging / SolC / CoWoS

  Key supply chain observation indicators

  Strengths

  SolC and CoWoS bookings can serve as important signals of demand for Apple's M-series and AI ASICs.

  Weaknesses

  Apple may use different terminology in the supply chain, making external tracking harder to interpret.

  Comparison

  SolC may be used across multiple product categories including M5 Max, M5 Ultra, workstations, and AI servers, so usage needs to be distinguished.

  Risks

  If packaging capacity, HBM adoption, and tape-out progress fall short of expectations, AI ASIC ramp-up may be affected.

Key data

• RatingOverweightMorgan Stanley's stock rating on Apple.
• Price target$360.00Base-case price target.
• Closing price$333.74The closing price disclosed in the report for Jul 17, 2026; the original text shows S333.74, interpreted in the context of a U.S. dollar price.
• Bull-case price target$453.00Corresponds to 11x FY27 EV/Sales and about 41.1x FY27 P/E.
• Bear-case price target$194.00Corresponds to 6.3x FY27 EV/Sales and 24.6x FY27 Bear P/E.
• FY27 EPS$10.30FY27 EPS used by the report for base-case valuation.
• FY27 revenue forecast$557,504mmFY27 fiscal-year revenue forecast disclosed in the table.
• FY27 net profit forecast$149,643mmFY27 net profit forecast disclosed in the table.
• FY27 total revenue YoY growth14.9%FY27e total revenue growth rate in the key earnings inputs table.
• FY27 iPhone revenue YoY growth19.1%FY27e iPhone revenue growth rate in the key earnings inputs table.
• FY27 Services revenue YoY growth12.1%FY27e Services revenue growth rate in the key earnings inputs table.
• FY27 gross margin47.5%FY27e gross margin in the key earnings inputs table.

Impact & implications

For Apple, AI server chip capability directly affects Siri AI performance, latency, and cloud cost structure. If progress on in-house chips falls short of expectations, Apple may use underpowered chips in Private Cloud Compute or rely more on Google Cloud/NVIDIA GPUs, thereby reducing AI service experience or profit margins; if Baltra and the second-generation AI ASIC progress as planned, they should help improve AI workload costs and long-term platform control. From an investment perspective, the report still believes the medium- to long-term iPhone replacement cycle, services revenue growth, and ecosystem monetization support the Overweight rating, though near-term AI execution and cost uncertainties need to be digested.

Risks

• Insufficient performance of Apple's AI server chips, causing complex Siri AI query performance or latency to deteriorate.
• More Siri AI workloads shifting to Google Cloud, increasing costs.
• Baltra ASIC or second-generation AI ASIC progress falling short of expectations.
• Rising memory input costs pressuring gross margin.
• The iPhone 17 cycle falling short of expectations, with consumer spending weaker than expected.
• Apple Intelligence adoption and feature expansion coming in below expectations.
• Slower Services growth.
• Geopolitical tensions.
• Increasing regulatory pressure in areas such as the App Store.

What to watch

• TSMC's SolC capacity bookings.
• TSMC's CoWoS capacity bookings, especially whether they relate to packaging demand similar to HBM and high-power AI GPUs.
• Tape-out progress of the second-generation AI ASIC.
• Whether Apple's capital expenditure trajectory continues to accelerate.
• Performance, latency, and user feedback after Siri AI launches this fall.
• Whether iPhone 17 upgrade willingness and the replacement cycle accelerate.
• Whether Services revenue growth reaccelerates.
• Whether gross margin can withstand memory costs and product mix pressure.