Tech Giants' Surge in Capital Expenditure Benefits Cloud Networking Equipment Supply Chain

Overview

This report is Barclays' interpretation of the supply chain impact from the latest quarterly earnings of Microsoft (MSFT), Meta (META), and Google (GOOGL). The key conclusion is that the three tech giants, competing for AI leadership, have significantly raised their 2026 capital expenditure guidance, with a strong focus on GPUs, CPUs, and supporting network and data center infrastructure. This trend provides substantial benefits to cloud infrastructure supply chain companies covered by Barclays, particularly network equipment provider Arista, optical module/interconnect manufacturer Ciena, and contract manufacturer Celestica. However, traditional PC businesses still face demand weakness risks, and Apple's advertising revenue also carries uncertainties.

Core views

The capital expenditures of the three giants exhibit both volume and price increases. Microsoft's March quarter capital expenditure reached $30.9 billion (up 84% year-over-year), with about two-thirds allocated to short-lived assets like GPUs/CPUs. Management raised CY26 full-year capital expenditure guidance to approximately $190 billion (more than doubling year-over-year) and explicitly stated that FQ4 expenditures would increase sequentially to over $40 billion to accelerate capacity rollout. Meta raised FY26 capital expenditure guidance from $115-135 billion to $125-145 billion, exceeding market expectations of around $123 billion, with the increase primarily from AI infrastructure construction and higher component prices. Google's March quarter capital expenditure was $35.7 billion (up 107% year-over-year), and FY26 guidance was raised to $180-190 billion (up 102% year-over-year), with FY27 expenditures expected to increase significantly, reflecting an extremely optimistic outlook on AI computing demand. The supply chain benefits are clear and structurally differentiated. On the network equipment side, Arista (ANET) holds over 60% share in Microsoft, and as Microsoft continues to focus on AI spending, its switch business is likely to receive incremental orders. Meta is also a key customer for Arista (contributing about 16% of FY25 revenue). Celestica (CLS) also sees rising customer concentration, with Meta accounting for 15% of FY26 estimated revenue and Google's share expected to increase from last year's 32% to 35%, making it one of the core beneficiaries of this capital expenditure expansion. Ciena (CIEN) is particularly favored due to its deeper exposure to the data center interconnect (DCI) space. Corning (GLW), while having a relatively low share in data center business, recently signed two hyperscale customer agreements along with a previous $6 billion multi-year contract with Meta, supporting its hyperscale business throughout CY26. Not all sectors benefit equally. The PC market remains under pressure: Microsoft's Windows OEM and device revenue saw low single-digit year-over-year declines, and Barclays predicts CY26 PC shipments will see high single-digit declines, with risks like memory price drops further suppressing shipments. Additionally, while Google's traffic acquisition costs (TAC) grew 11% year-over-year to $15.23 billion, Barclays believes Apple's advertising revenue still faces potential risks due to its ISA agreement with Apple.

Analysis framework

The report adopts a typical "large customer capital expenditure → supply chain order mapping" analysis framework. It first extracts actual capital expenditure and forward guidance from the earnings reports of the three tech giants, breaking down their spending structure (e.g., short-lived vs. long-lived assets, servers vs. network equipment vs. data center construction). It then combines historical revenue share, share changes, and new contract information of supply chain companies with these large customers to quantitatively estimate the revenue elasticity of capital expenditure expansion for specific targets. This method essentially translates upstream giants' "investment willingness" into midstream hardware vendors' "revenue visibility," particularly suitable for B2B markets like AI infrastructure dominated by a few large buyers.

Methodology notes

• Industry/Sector Analysis FrameworkUpstream-Midstream-Downstream Transmission

  Tracking changes in capital expenditure guidance from downstream large customers to derive revenue growth opportunities for mid- and upstream hardware suppliers

  In markets like AI infrastructure dominated by a few tech giants, downstream buyers' capital expenditure plans are the most critical leading indicators for midstream equipment suppliers. The report quantitatively analyzes the magnitude of capital expenditure increases by Microsoft, Meta, and Google and their allocation across segments like servers, network equipment, and data centers, then overlays suppliers' revenue share with these customers to precisely identify which companies will most directly benefit from this wave of AI investment.

• Company Fundamentals & Financial Framework

  Capital expenditure structure breakdown (short-lived vs. long-lived assets)

  The report particularly notes that about two-thirds of Microsoft's capital expenditure is allocated to short-lived assets like GPUs/CPUs. Short-lived assets depreciate quickly and require frequent updates, implying stronger continuity and predictability for related supplier orders, whereas long-lived assets (e.g., data center buildings) see slowed procurement once completed. This breakdown helps investors assess the sustainability of supply chain benefits beyond just aggregate figures.

Asset mapping & comparison

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

• Arista Networks (ANET)

  Core Beneficiary: Over 60% share in Microsoft, Meta contributes ~16% of FY25 revenue, AI capital expenditure expansion directly drives switch demand

  Strengths

  Leading share among top cloud providers, products highly aligned with AI cluster architecture

  Comparison

  Compared to traditional players like CSCO, ANET's share growth in AI data center scenarios is faster

  Risks

  High customer concentration, potential volatility if giants adjust capital expenditure pace

• Celestica (CLS)

  Core Beneficiary: Meta accounts for ~15% of FY26E revenue, Google's share expected to rise from 32% to 35%, dual-driven

  Strengths

  Simultaneously tied to two giants raising capital expenditure, strong revenue visibility

  Comparison

  Compared to FLEX, CLS has superior customer structure in this AI infrastructure cycle

  Risks

  Contract manufacturing business has relatively low margins, component price increases may squeeze profitability

• Ciena (CIEN)

  Beneficiary: Deep exposure to data center interconnect (DCI) sector, benefiting from optical network upgrade demand as cloud providers expand data center scale

  Strengths

  High technical barriers in DCI, better competitive landscape than generic switches

  Comparison

  Compared to ANET, CIEN is more transmission-layer focused than switching-layer, benefit timing may slightly lag

  Risks

  DCI demand tied to data center construction pace, delays may affect orders

• Corning (GLW)

  Marginal Beneficiary: Newly signed two hyperscale customer agreements + previous $6 billion multi-year contract with Meta supports hyperscale business

  Strengths

  Fiber optic cable leader, long-term contracts provide revenue stability

  Weaknesses

  Data center/hyperscale business still accounts for low double-digit share of total revenue

  Comparison

  Compared to pure AI infrastructure plays, GLW's business elasticity is relatively limited

  Risks

  Traditional communications business may drag overall growth

• Apple (AAPL)

  Potential Negative Impact: Google TAC growth but ISA agreement uncertainties pose risks to advertising revenue

  Weaknesses

  Advertising revenue relies on third-party platform partnerships, bargaining power subject to regulatory environment

  Risks

  ISA agreement terms changes or regulatory intervention may lead to advertising revenue decline

Key data

• Microsoft CY26 Capital Expenditure GuidanceApproximately $190 billionUp over 100% year-over-year, including about $25 billion impact from component price increases
• Meta FY26 Capital Expenditure Guidance$125-145 billionRaised from previous $115-135 billion range, above market expectations of around $123 billion
• Google FY26 Capital Expenditure Guidance$180-190 billionRaised from previous $175-185 billion, up 102% year-over-year
• Microsoft March Quarter Capital Expenditure$30.9 billionUp 84% year-over-year, above market expectations of $27.2 billion
• Arista's Share in MicrosoftOver 60%AI spending focus likely to drive more switch business
• Celestica's Revenue Share from GoogleFY26E ~35%Up from FY25's 32%

Impact & implications

For the cloud infrastructure supply chain, the simultaneous and significant upward revisions in capital expenditure guidance by the three giants confirm high demand visibility for AI-related hardware in 2026, particularly as network equipment and optical interconnect segments shift from "GPU dominance" to "system-level support." Arista, Celestica, and Ciena are seen as core beneficiaries due to high customer concentration or critical product positioning. However, investors should note that the PC market's weakness has not reversed amid the AI boom, requiring caution for related targets (e.g., HPE, DELL's PC businesses). Additionally, the advertising cooperation agreement between Apple and Google may introduce regulatory or commercial terms uncertainties for AAPL's advertising revenue.

Risks

• Long-term value proposition of AI investments remains uncertain; if returns fall short, giants may adjust capital expenditure pace in the future
• PC market demand continues to weaken, with risks like memory price drops further suppressing shipments
• Apple and Google's ISA advertising cooperation agreement faces regulatory or commercial terms change risks

What to watch

• Subsequent quarters' actual capital expenditure execution and guidance adjustments by tech giants
• Revenue growth validation from hyperscale customers for Arista, Celestica, etc., in next earnings season
• Impact of component price trends on capital expenditure structure and supplier margins