European Semiconductors: Orbital Computing Opens a New Frontier, Memory Momentum May Have Peaked

Overview

This report compiles Morgan Stanley’s global technology webcast views on space technology, orbital computing opportunities, and the memory cycle. The core takeaway is that low Earth orbit satellites, optical interconnect, AI accelerators, memory, power, and thermal management could jointly form a new orbital compute infrastructure, but commercialization remains in an early stage in the short term; at the same time, the memory cycle remains strong, while the pace of marginal improvement may slow.

Core views

The report sees orbital computing as a potential new frontier for AI infrastructure: in the near term, satellite-side AI handling of image, sensor, and inference tasks dominates; in the medium term, distributed orbital clouds are formed through optical inter-satellite links; in the long-term bull case, satellites equipped with AI accelerators, memory, solar arrays, and radiators form a space-based virtual data center. Within the European semiconductor supply chain, the report clearly favors STMicroelectronics and Infineon as critical enablers. For memory, prices are still rising but the rate of change appears to be peaking, and the market is likely to focus more on company earnings and hyperscaler pressure rather than the companies' own commentary from memory vendors.

Analysis framework

The analytical framework combines thematic research, supply-chain mapping, staged technical roadmap decomposition, and equity valuation. The orbital computing section evaluates feasibility through power cost, launch cost, satellite hardware, optical communications, thermal management, radiation tolerance, and scheduling software. The memory section focuses on year-over-year price changes, the breadth of upward earnings revisions, and relative performance within the AI capex cycle. STMicroelectronics valuation is handled through sum-of-parts valuation.

Methodology notes

• Thematic supply chain analysisOrbital Computing Three-Stage Framework

  From orbital edge computing to orbital cloud, and then to space AI infrastructure

  The report breaks orbital computing into near-term, medium-term, and long-term commercialization stages to assess technical readiness, supply-chain beneficiaries, and potential market size.

• Valuation methodSum-of-the-parts valuation

  STMicroelectronics target price €78

  The report values STM’s core business at FY28e EPS of €2.11 with an 18x P/E, values optical and LEO businesses at FY28e EPS of €1.34 with a 36x P/E, and discounts them at a 10% WACC to arrive at a €78 target price.

• Cycle trackingEarnings revision breadth and price change rate

  Memory momentum is peaking at the margin

  The report uses DRAM earnings-revision breadth near about 89% and a slowing price increase rate entering Q4 2026 as a signal that memory relative outperformance may weaken.

Asset mapping & comparison

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

• STMicroelectronics NV (STMPA.PA)

  Beneficiary of European semiconductor and LEO/optical opportunities

  Strengths

  The report assigns an Overweight rating and a €78 target price, and sees its core business, optical segment, and LEO-related business as having different growth trajectories. Upside factors include broader SiC adoption, design wins in automotive and industrial, an optics ramp in 2H26/FY27, and FY26-FY28 cumulative LEO sales above $3bn.

  Weaknesses

  The core business remains exposed to automotive and industrial cycles, and there is pressure on MCU pricing and smartphone sensor content.

  Comparison

  In the report’s stock preference, STM and Infineon are listed as critical enablers in the European orbital computing supply chain.

  Risks

  Global macro weakness, weaker automotive and industrial OEM sales, downward MCU pricing pressure, DRAM shortage affecting automotive sales, loss of key smartphone design slots, or a decline in sensor content.

• Infineon Technologies AG (IFXGn.DE)

  European semiconductor orbital computing supply-chain enabler

  Strengths

  The report places it alongside STM as a European critical enabler, potentially benefiting from higher power, power-management, and semiconductor content.

  Weaknesses

  The report body does not provide separate valuation details or a target-price derivation for this company.

  Comparison

  It is another preferred European semiconductor supply-chain name in the report alongside STM.

  Risks

  Semiconductor cycle, automotive and industrial demand, customer capex pace, and orbital computing commercialization below expectations.

• Memory sector

  Cyclical asset in the AI capex cycle

  Strengths

  DRAM and NAND year-over-year prices remain strong; the report prefers DRAM and traditional memory.

  Weaknesses

  The rate of price change is expected to peak, and DRAM earnings-revision breadth is near historical highs, so relative performance may come under pressure.

  Comparison

  Compared with NAND, the report prefers DRAM and legacy memory, and is least constructive on memory-module makers.

  Risks

  Hyperscaler pressure, stock performance being driven by company earnings rather than tape, narrowing earnings-revision upside, and cyclical pullbacks.

Key data

• Report date2026-07-07The cover shows July 7, 2026 10:05 AM GMT.
• Orbital compute capex per wattAbout US$60/W in 2030 to US$9/W by 2040Driven by reusable launches, lower satellite hardware costs, and improved payload economics.
• DRAM earnings-revision breadthAbout 89%Near historical highs, and the report sees this as a signal that memory relative performance may weaken.
• LEO sales growthFY26-FY28 CAGR 48%The report says this is mainly driven by end-user terminal demand.
• STMicroelectronics target price€78Represents about 23% upside; the current price in the report is €63.36.
• STM valuation assumptionsCore business 18x FY28e P/E; optical and LEO business 36x FY28e P/E; WACC 10%Used for sum-of-the-parts valuation.

Impact & implications

If the orbital computing cost curve continues to decline, European semiconductor companies may gain structural opportunities in power management, sensors, optical communications, radiation-tolerant chips, thermal management, and edge AI payloads. For investors, the report suggests separating long-term structural bull opportunities in AI capex from the short-term cooling momentum in the memory cycle: the former supports selection of key enablers, while the latter calls for avoiding excessive pursuit of memory beta when earnings-revision breadth is at elevated levels.

Risks

• Orbital computing depends on reusable launch costs continuing to decline; if reusable launch cost, satellite hardware cost, or payload economics improvement underperforms, commercialization timing could be delayed.
• LEO satellites move fast, have limited resources, and experience continuously changing links, imposing high requirements on scheduling, orchestration, and dynamic resource management.
• Orbital computing faces challenges including radiation, thermal management, on-orbit maintenance, hardware refresh cycles, optical network bandwidth, orbital debris, security, and collision avoidance.
• Although memory prices are still rising, the rate of change is peaking and earnings-revision breadth is at elevated levels, which could lead to underperformance relative to the rest of the market.
• STMicroelectronics faces downside risks from weaker macro conditions, soft automotive and industrial demand, MCU pricing pressure, DRAM shortages affecting automotive sales, and potential loss of smartphone design slots.

What to watch

• Whether orbital compute capex/W declines from about US$60/W in 2030 to US$9/W by 2040 along the trajectory in the report.
• The progress of SpaceX reusable launch costs, satellite hardware cost declines, and improvements in payload economics.
• Adoption speed of optical inter-satellite links, laser terminals, coherent optics, optical DSP, and DWDM in satellite networks.
• Whether FY26-FY28 LEO sales reach or exceed the cumulative $3bn scenario mentioned in the report.
• How year-over-year growth rates of DRAM and NAND prices evolve after Q4 2026, and whether DRAM earnings-revision breadth declines from the roughly 89% peak.
• The impact of hyperscaler capex and earnings pressure on memory and AI supply-chain equities.