CY27 Hyperscale Capex Expectations Raised Again, AI Compute Demand Still Faces Upside Pressure
AI summary card
CY27 Hyperscale Capex Expectations Raised Again, AI Compute Demand Still Faces Upside Pressure
Morgan Stanley expects the combined CY27 capex of Microsoft, Alphabet, Amazon, Meta and SpaceX to be about $1.15 trillion, up about 47% year over year, while also being positive on the structural boost from AI diffusion to compute, storage and selected European software companies.
- Consensus CY27 capex expectations for the four major hyperscale cloud providers excluding SpaceX have been raised by 20% since July, indicating that investment appetite remains strong.
- CY27 capex for the five companies is expected to be about $1.15 trillion, up about 47% year over year; CY28 growth is expected to slow to about 11%, making subsequent forecast revisions a key point to monitor.
- More than 60% of enterprises are already using open-weight models in their technology stacks, and more models and pricing tiers may continue to support compute and storage demand by expanding AI adoption.
- An SAP reseller survey shows that 33% of resellers have observed customers adopting paid Business AI or Joule, while 53% say customers are still mainly in the pilot stage.
- IONOS was upgraded to Overweight, with FY27 constant-currency revenue growth expected to accelerate to about 8.6%, while the cloud business and AI products provide additional upside optionality.
Report interpretation
Overview
This report reviews important events in the European Software and Services industry over the week and identifies hyperscale capex, open-weight models, SAP enterprise AI adoption, legal-tech competition and IONOS growth prospects as core topics. The report argues that cloud revenue, data center commitments and component demand continue to strengthen, potentially driving further upward revisions to hyperscale capex expectations. Meanwhile, AI commercialization in enterprise software remains at an early stage, with near-term value more reflected in migration efficiency, cost reductions and product penetration than in standalone AI revenue.
Core views
First, the capex curves for Microsoft, Alphabet, Amazon, Meta and SpaceX have risen significantly, with a combined CY27 forecast of about $1.15 trillion, and compute and storage demand remains supported. Second, open-weight models do not necessarily weaken total compute demand; lower costs and richer model choices may expand AI usage through Jevons paradox. Third, SAP’s near-term AI opportunity mainly comes from accelerating S/4HANA migration, with 80% of resellers expecting AI migration tools to increase customers’ willingness to migrate over the next 12 to 24 months. Fourth, legal research and workflow products are expected to continue converging, but replicating TR and RELX’s content coverage, value-added interpretation and citation networks still presents high barriers. Fifth, IONOS’s improving growth, pricing and cloud business outlook support its upgrade to Overweight.
Analysis framework
The report cross-validates using Visible Alpha capex consensus expectations, a three-scenario analysis of open and closed models, an AlphaWise reseller survey, company calls and product updates, channel and system integrator checks, as well as NTM P/E, PEG and company guidance.
Methodology notes
Aggregate annual capex forecasts for major hyperscale companies and compare year-over-year growth and recent revision magnitude.
The report builds a combined capex curve using forecasts for Microsoft, Alphabet, Amazon, Meta and SpaceX, and focuses on tracking whether CY27 and CY28 expectations continue to be revised upward.
Assess three possible states: closed models winning, a mixed ecosystem, and open-weight models reaching frontier performance.
The framework is used to assess changes in model pricing, enterprise adoption, on-device deployment and compute demand under different competitive landscapes, and uses Jevons paradox to explain how cost declines may expand total usage.
Use reseller feedback to measure paid customer adoption, pilot proportions, spending willingness and the impact on S/4HANA migration.
The survey shows enterprise AI adoption remains early, but resellers generally expect AI to increase SAP customer spending and accelerate S/4HANA conversion.
Combine expected growth, margins, valuation multiples and management guidance to identify rating changes and catalysts.
IONOS’s upgrade reflects factors including FY27 growth acceleration, a PEG of about 0.93x, cloud business potential and AI products not yet fully incorporated into forecasts.
Asset mapping & comparison
Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).
- ALPHABET INC（GOOGL.US）A core hyperscaler in the report’s combined capex forecast and an important driver of cloud revenue and data center investment expansion.
- Strengths
- GCP usage and revenue growth, AI model demand and expanded data center commitments can support long-term cloud business scaling.
- Weaknesses
- High capex intensity may pressure short-term free cash flow and return on investment.
- Comparison
- Together with Microsoft, Amazon and Meta, it forms the four-company capex watch group excluding SpaceX.
- Risks
- AI demand falling short of expectations, data center construction delays, rapid declines in unit compute costs or regulatory restrictions.
- META PLATFORMS INC（META.US）A core hyperscaler in the report’s combined capex forecast, with AI training, inference and recommendation systems supporting its infrastructure demand.
- Strengths
- Its large user base and advertising scenarios help convert AI infrastructure into product engagement and commercial efficiency.
- Weaknesses
- High infrastructure investment increases depreciation and cash flow pressure, and returns depend on continued monetization of advertising and AI products.
- Comparison
- Together with Alphabet, Microsoft and Amazon, it drove the upward revision to CY27 capex expectations versus July.
- Risks
- Longer capex payback period, open models reducing differentiation, and regulatory and data privacy constraints.
- SAP SEThe report uses the AlphaWise reseller survey to assess SAP’s enterprise AI positioning and S/4HANA migration opportunity.
- Strengths
- A large enterprise customer base, business data and S/4HANA migration demand provide distribution advantages for Business AI and Joule.
- Weaknesses
- Paid AI adoption remains early, most customers are still in the pilot stage, and standalone AI revenue contribution is limited.
- Comparison
- Compared with pure AI products, SAP’s clearer near-term opportunity is to use AI tools to accelerate core ERP migration.
- Risks
- Customer pilots failing to convert into paid deployments, prolonged migration cycles and open models weakening AI premiums.
- IONOS Group SE（IOSn.DE）The report upgrades its rating to Overweight and expects FY27 group revenue growth to accelerate.
- Strengths
- Customer momentum, pricing improvement, Cloud Solutions growth and AI product innovation jointly provide upside potential.
- Weaknesses
- Second-quarter adjusted EBITDA was below expectations, and part of the growth depends on the timing of cost investments and execution quality.
- Comparison
- A PEG of about 0.93x is considered still attractive relative to its prospects for growth acceleration.
- Risks
- Failure to achieve the FY26 adjusted EBITDA target of about €530m, slower cloud business growth or weaker-than-expected AI adoption.
- RELXThe report discusses the impact of Legora’s entry into the U.S. foundational legal content market on the competitive landscape for legal research and workflows.
- Strengths
- It has more than 200 billion legal, news and public records from over 50,000 sources, as well as value-added interpretation and citation networks.
- Weaknesses
- A closed data ecosystem may encourage new entrants to build their own content libraries and prompt customers to seek integrated alternatives.
- Comparison
- RELX and Thomson Reuters remain leaders in the U.S. legal research market, and it is difficult for Legora to replicate the same product capabilities.
- Risks
- AI-generated legal research products accelerating substitution, foundational legal content becoming more open, and intensified competition in legal workflows.
Key data
- CY27 Combined Capex ForecastAbout $1.15 trillionCovers Microsoft, Alphabet, Amazon, Meta and SpaceX.
- CY27 Capex Year-over-Year GrowthAbout 47%Expected to be about half of CY26 growth.
- CY27 Forecast Revision for the Four Major HyperscalersRaised by 20% since JulyExcludes SpaceX, reflecting continued strengthening in second-quarter capex intentions.
- CY28 Expected Capex GrowthAbout 11%Growth slows significantly, making the direction of subsequent revisions a core variable.
- Enterprise Open-Weight Model Usage RateOver 60%Mainly used in specific scenarios emphasizing speed, security or high-frequency usage.
- Paid AI Adoption Among SAP Customers33%The proportion of customers reported by resellers as having adopted paid Business AI or Joule; about 10 percentage points of this represents broad adoption.
- SAP Customers Still in AI Pilots53%Indicates that enterprise AI commercialization remains at an early stage.
- Resellers Expecting AI to Increase SAP Spending87%Expected timeframe is the next 12 months.
- AI Tools Promoting S/4HANA Migration80%Resellers expect customer willingness to migrate to increase over the next 12 to 24 months.
- IONOS FY27 Revenue Growth ForecastAbout 8.6%Calculated at constant exchange rates, with growth acceleration coming from customer momentum and pricing contribution.
- IONOS ValuationAbout 0.93x PEGCalculated based on the Thursday closing price cited in the report.
- IONOS Rating and Price TargetOverweight, €38.75The table shows a price of €30.20 on August 6, 2026, corresponding to mechanically calculated potential upside of about 28.3%.
Impact & implications
Upward revisions to capex expectations are a positive signal for demand for data center components, cloud infrastructure, computing power and storage, and also indicate that hyperscalers such as Alphabet and Meta remain in a high-intensity investment cycle. However, massive investment may also pressure near-term free cash flow and raise return thresholds. For European software companies, the near-term value of AI is more likely to be realized through improved migration efficiency, lower inference costs, better product penetration and enhanced pricing power. SAP may benefit from faster S/4HANA migration, IONOS may benefit from cloud growth and AI product innovation, while legal information platforms such as RELX still have barriers in content depth, expert interpretation and citation networks.
Risks
- The expectation of about 47% year-over-year CY27 capex growth is already high, and forecasts may be revised downward if cloud revenue, token processing volume or data center demand slows.
- CY28 capex growth is expected to fall to about 11%, indicating continued uncertainty around the sustainability of high growth.
- Open-weight models may lower model pricing and software differentiation, making it difficult for some enterprises to convert usage growth into profits.
- AI adoption among SAP customers is still mainly in the pilot stage, and commercialization and standalone AI revenue realization may be slower than market expectations.
- IONOS’s growth acceleration depends on pricing, customer momentum, cloud business execution and returns on cost investments.
- Morgan Stanley has investment banking or other business relationships with multiple covered companies, and investors should prudently assess research independence in conjunction with the disclosures.
What to watch
- Whether CY27 capex growth forecasts of about 47% and CY28 forecasts of about 11% continue to be revised upward.
- GCP, AWS and Azure revenue growth, monthly token processing volume and new data center commitments.
- Whether adoption rates for paid SAP Business AI and Joule can rise from current early levels.
- Whether AI migration tools materially shorten S/4HANA project cycles and improve conversion rates.
- Whether IONOS’s third-quarter results can confirm that the FY26 adjusted EBITDA target of about €530m remains achievable.
- Whether IONOS brings forward subsequent medium-term financial targets, as well as Cloud Solutions and AI product adoption metrics.
- Whether open-weight models approach frontier performance and drive more localized and edge deployments.
- Whether Legora can fill three key capability gaps in U.S. foundational legal content, value-added interpretation and citation networks.