Cloud Capacity Ramp Validates Growth Potential; Raises Google’s Price Target to $460

Overview

This report focuses on data center capital expenditures and future capacity conversion efficiency among North American internet giants. Morgan Stanley employs a bottom-up cost and chip-generation framework to estimate incremental compute capacity (measured in GW) for AWS and Google Cloud over the next two years (2026–2027), linking this physical capacity to revenue forecasts to derive a key metric: 'incremental revenue per GW.' The core conclusion is that current market expectations for cloud growth may be overly conservative—particularly for Google Cloud—whose supply-demand dynamics suggest substantial upside potential. Accordingly, the firm raises Alphabet’s bull-case price target while maintaining positive views on Amazon and Meta.

Core views

Capacity Expansion and Revenue Conversion: The report builds a bottom-up cost/GW framework incorporating nine different chip generations (including NVIDIA and custom silicon). The model indicates hyperscale data centers will add approximately 14GW of incremental capacity in 2026 and ~20GW in 2027. After excluding internal workloads, roughly 6GW/8GW of this capacity in 2026/2027 is expected to serve public cloud offerings from AWS and Google Cloud. Implied Reasonableness of Revenue per GW: Based on current revenue forecasts, AWS and Google Cloud are projected to generate ~$14 billion and ~$11 billion in incremental revenue per GW respectively in 2027. The report considers these levels reasonable or even conservative, citing: 1) Neo-clouds (e.g., CoreWeave) currently generate ~$10 per watt, whereas hyperscalers’ software services should command a premium; 2) Industry discussions indicate negotiated deals are achieving revenue/GW levels exceeding $20 billion. Google Cloud’s Upside Bias: Among major cloud providers, Google Cloud’s 2027 outlook appears most conservative. Its lower implied revenue/GW (~$11 billion) may partly reflect TPU accounting treatment, but supply-demand alignment suggests the greatest room for upward revisions. Faster-than-base-case growth for Google Cloud would drive multiple expansion and earnings upgrades, supporting stronger stock performance. Stock-Specific Views: For Alphabet (GOOGL), given Google Cloud’s growing contribution and leadership in generative AI, the bull-case price target is raised from $425 to $460—implying a 28x P/E on 2027 bull-case EPS of ~$16.50. For Amazon (AMZN), high-margin businesses continue to enhance profitability while sustaining investment; the $330 target implies a 29x P/E on 2027 EPS of $11.30. For Meta (META), structural shifts toward efficiency and productivity, combined with improved Reels monetization and AI-driven ad measurement advances, support a $775 target price.

Analysis framework

The report employs a unique 'capacity-to-revenue' linkage methodology. First, it estimates physical compute capacity additions (in GW) for hyperscale data centers over the next two years using a bottom-up approach based on chip-generation costs and rack allocations. Second, it compares these physical capacity increments against revenue forecasts in financial models to calculate 'incremental revenue per GW' as an efficiency metric. Finally, it benchmarks this metric against industry standards (e.g., bare-metal GPU leasing revenue from Neo-clouds and negotiated deal prices) to assess the reasonableness and conservatism of current revenue growth assumptions. This approach directly ties infrastructure investment to financial output, providing a hard constraint and sanity check on cloud growth sustainability.

Methodology notes

• Industry/Sector Analysis FrameworkVolume-price decomposition

  Decomposes cloud revenue growth into two dimensions: 'capacity increment (GW)' and 'revenue per unit capacity (Revenue/GW)'

  By breaking total revenue growth into physical compute capacity growth (volume) and revenue generated per unit of compute (price/efficiency), it becomes clearer whether growth stems from pure scale expansion or improvements in operational efficiency/pricing power.

• Valuation MethodDCF (Discounted Cash Flow)

  Uses DCF model and long-term EBITDA multiples to determine price targets

  The report sets price targets using a combination of discounted cash flow (DCF) modeling and long-term EBITDA multiples, assuming a weighted average cost of capital (WACC) of ~8% and a terminal growth rate of ~3%.

Asset mapping & comparison

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

• Alphabet Inc. (GOOGL.O)

  Beneficiary; Google Cloud growth expectations are conservative with the largest upside potential

  Strengths

  Google Cloud acceleration, generative AI leadership, expense discipline driving operating leverage

  Comparison

  Compared to AWS, Google Cloud’s revenue/GW expectation is lower, implying a larger expectation gap

  Risks

  Global advertising growth slowdown, margin pressure from new AI products, regulatory constraints on ad targeting

• Amazon.com Inc (AMZN.O)

  Beneficiary; AWS growth remains solid, high-margin businesses drive profitability

  Strengths

  AWS at an adoption inflection point, retail share gains, advertising business growth

  Risks

  AWS revenue deceleration or margin compression, longer-than-expected investment cycle, merchandise margins below expectations

• Meta Platforms Inc (META.O)

  Beneficiary; efficiency gains and AI-driven ad monetization

  Strengths

  Structural efficiency improvements, better Reels monetization, AI-driven engagement gains

  Risks

  Declining user engagement, Reality Labs losses widening significantly, execution missteps in data center build-out leading to higher-than-expected capital intensity

Key data

• 2026 Hyperscale Data Center Incremental Capacity~14 GWBottom-up model forecast
• 2027 Hyperscale Data Center Incremental Capacity~20 GWBottom-up model forecast
• 2027 AWS Incremental Revenue per GW~$14 bnImplied value, considered conservative
• 2027 Google Cloud Incremental Revenue per GW~$11 bnImplied value, excluding external TPU sales, considered most conservative
• GOOGL Bull-Case Price Target$460Raised from $425, implying 28x P/E on 2027 EPS of $16.50
• AMZN Price Target$330Implies 29x P/E on 2027 EPS of $11.30
• META Price Target$775Implies ~23x P/E for 2027

Impact & implications

The report argues that as data center capacity comes online, pent-up demand will translate into revenue on financial statements, supporting—or even exceeding—current cloud growth expectations. Google Cloud, in particular, has conservative assumptions that could lead to sustained earnings upgrades and valuation re-rating over the coming quarters. For investors, the focus should shift from concerns about capital spending alone toward capacity utilization rates and improvements in revenue efficiency per unit of capacity.

Risks

• Further global advertising growth slowdown pressuring margins
• New AI products exerting greater-than-expected margin pressure due to low monetization rates and high compute intensity
• Regulatory restrictions limiting ad targeting capabilities
• Reality Labs losses expanding significantly
• Execution failures in data center construction leading to persistently high capital intensity
• Macroeconomic uncertainty or weakening consumer spending

What to watch

• Actual pace of hyperscale data center capacity ramp-up
• Whether Google Cloud revenue growth exceeds base-case assumptions
• AWS revenue growth trajectory and margin expansion
• Meta’s Reels monetization progress and ROI on AI investments
• Changes in negotiated industry revenue/GW levels for capacity deals