Report Interpretation
Covering the latest research from top Wall Street investment banks
Report InterpretationHilo Research

China data centers: Alibaba's 20GW-plus 2032 data-center target creates a positive read-across for China data-center and neocloud providers

Morgan Stanley argues that Alibaba's new global capacity ambition implies 16-17GW-plus of net additions over seven years, extending the demand runway for data-center vendors. The report highlights VNET and Kingsoft Cloud among potential beneficiaries.

InstitutionMorgan Stanley
Date20260922
IndustryChina data centers

Summary

Morgan Stanley argues that Alibaba's new global capacity ambition implies 16-17GW-plus of net additions over seven years, extending the demand runway for data-center vendors. The report highlights VNET and Kingsoft Cloud among potential beneficiaries.

Industry view: In-Line
China data centersAlibaba Apsara ConferenceAI computeCloud capexVNETKingsoft Cloud
  • Alibaba targets more than 20GW of global data-center capacity by 2032.
  • Morgan Stanley estimates Alibaba will have 3-4GW of capacity at end-2025, implying 16-17GW-plus of net additions thereafter.
  • The implied average addition is 2-3GW annually over the next seven years.
  • Strong demand for Alibaba's in-house T-head chipset is seen as supporting order delivery.
  • The Asia Pacific industry view remains In-Line.

Report Interpretation

Overview

This event update interprets announcements from Alibaba's Apsara Conference as supportive for China data-center operators and neocloud providers. Morgan Stanley sees the capacity target as evidence of a larger and longer-lasting AI-driven infrastructure buildout.

Core views

Alibaba's CEO announced plans to train a 5-10 trillion-parameter model, expand global data-center capacity to more than 20GW by 2032, and materially increase shipments of its new T-head chipset. The stated capacity target was linked to strong customer demand. Morgan Stanley estimates that Alibaba should reach 3-4GW of capacity by end-2025. On that basis, the 2032 target implies 16-17GW-plus of global net capacity additions over the following seven years, or roughly 2-3GW per year on average. The institution views this as reaffirming or modestly exceeding Alibaba's prior 2025 Apsara Conference guidance for a tenfold capacity expansion over ten years. The report argues that this pace of capacity buildout creates volume upside and lengthens the demand cycle for data-center vendors. It views the demand signal as broadly positive for China data-center players, including VNET and OW, and for neocloud providers, including Kingsoft Cloud and OW. Strong demand for Alibaba's internally developed chipset is also viewed as helping secure order delivery. For the covered-company valuation materials, Morgan Stanley uses a 5.5x 2027e EV/EBITDA base-case multiple for one case, below the US neocloud peer average of 13x and median of 10x. The discount reflects lower asset yields and greater uncertainty over supply-chain availability; EBITDA is preferred because depreciation policies differ across companies and EBIT is limited. A separate base case uses a 10-year DCF with an 8.7% WACC, comprising a 4.0% cost of debt, 14.0% cost of equity and 50% debt weighting, plus a 3% terminal growth rate. Potential upside factors cited include stronger-than-expected AI GPU demand, lower interest rates, better-than-expected development of Xiaomi's AI model, new wholesale contracts, faster customer move-ins, further China or US rate cuts, and REIT-based asset monetization at accretive valuations. Downside factors include supply-side restrictions that create procurement shortfalls, higher funding costs from higher rates, slower Chinese AI-model progress, reductions in hyperscaler cloud and AI capex, delayed capacity delivery, and weak sales execution.

Analysis framework

Morgan Stanley starts with Alibaba's announced AI-model, capacity and chipset plans, compares the 2032 target with its estimate of end-2025 capacity, and converts the gap into implied annual additions. It then links the resulting infrastructure demand to data-center and neocloud providers, while presenting company valuation assumptions and explicitly listed upside and downside drivers.

Methodology notes

  • Industry AnalysisSupply-demand framework

    Capacity-demand analysis

    The report uses Alibaba's stated capacity target and estimated starting capacity to infer the volume and duration of prospective data-center demand.

  • Valuation methodsEV/EBITDA valuation

    2027e EV/EBITDA multiple valuation

    A 5.5x 2027e EV/EBITDA base-case multiple is compared with US neocloud peers, with the discount attributed to lower asset yields and supply-chain uncertainty.

  • Valuation methodsDCF (Discounted Cash Flow)

    10-year discounted cash flow model

    The DCF base case discounts projected cash flows using an 8.7% WACC and assumes 3% terminal growth.

Asset mapping & comparison

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

  • Kingsoft Cloud Holdings (KC.O)
    Identified as a neocloud player that could benefit from the positive industry read-across.
    Strengths
    Potential exposure to stronger AI GPU demand and improved AI-model development.
    Weaknesses
    The cited valuation discount reflects lower asset yields and supply-chain availability uncertainty.
    Comparison
    A 5.5x 2027e EV/EBITDA base case is below US neocloud peers' 13x average and 10x median.
    Risks
    Supply restrictions, higher interest rates and slower-than-expected AI-model development in China.
  • VNET Group Inc (VNET.O)
    Identified as a China data-center player that could benefit from Alibaba's capacity expansion.
    Strengths
    Potential upside from new wholesale contracts, faster move-ins, rate cuts and accretive REIT asset monetization.
    Risks
    Hyperscaler AI-capex reductions, delayed capacity delivery and weak sales execution.

Key data

  • Alibaba global data-center capacity target20GW+ by 2032Announced at the Apsara Conference, citing strong customer demand.
  • Estimated Alibaba capacity at end-20253-4GWMorgan Stanley estimate used as the starting point for the capacity-gap calculation.
  • Implied net capacity additions16-17GW+ over seven yearsEquivalent to an average of 2-3GW per year globally.
  • EV/EBITDA base-case multiple5.5x 2027eBelow the US neocloud peer average of 13x and median of 10x.
  • DCF WACC and terminal growth8.7% WACC; 3% terminal growthWACC assumes 4.0% cost of debt, 14.0% cost of equity and 50% debt weighting.

Impact & implications

Morgan Stanley believes Alibaba's announced expansion points to greater data-center equipment and capacity demand, with a longer demand cycle for China data-center vendors and neocloud providers. The report identifies execution, supply availability, AI-demand intensity and funding conditions as key determinants of whether that positive read-across is realized.

Risks

  • Supply-side restrictions could result in procurement shortfalls.
  • Higher-than-expected interest rates could raise funding costs.
  • Chinese AI-model development could progress more slowly than expected.
  • Hyperscalers could reduce cloud capex, particularly AI-related investment.
  • Capacity delivery could be delayed and sales execution could weaken.

What to watch

  • Progress toward Alibaba's more than 20GW global capacity target by 2032.
  • AI GPU demand and shipment growth for Alibaba's T-head chipset.
  • New wholesale contracts and customer move-in timing for data-center operators.
  • China and US interest-rate developments and potential REIT asset monetization.
  • Hyperscaler cloud and AI-capex trends and supply-chain availability.
Zhejiang ICP No. 2022035445-5
Disclaimer: Market data, charts, indicators, research views, and other information provided on this website are intended solely for information display, research communication, and educational reference. They should not be regarded as personalized investment advice, securities recommendations, trading instructions, solicitations, or guarantees of return. While we strive to improve the reliability of our data and content, such information may still be subject to delays, errors, incompleteness, or untimely updates due to source differences, methodological limitations, system processing, or market volatility. Users should exercise independent judgment based on their own circumstances and bear all risks and responsibilities arising from the use of this website.

Settings

Sign in to view recent logins