There is no defensible year when cloud computing is expected to stop growing. Published forecasts point to continued expansion through at least 2028, with AI and hybrid-cloud use adding demand. The likelier change is how growth happens: after years of migration, more organizations will focus on choosing the right place for each workload and controlling its cost. Power, budgets, governance and skills may limit where growth occurs, but that is not the same as demand disappearing.
What does “cloud growth” mean?
Cloud growth can refer to several different measures: providers’ revenue, customer spending, computing capacity, or the share of workloads running in cloud environments. They are related, but they do not move in lockstep. The most directly relevant figures for the headline are forecasts of spending on public-cloud services; data-center capacity and electricity use help explain the infrastructure needed to serve that demand.
A forecast is also not a promise. It reflects assumptions available at a particular date and can change as analysts revise their baselines. A forecast of positive spending growth is evidence against a predicted near-term market contraction, but it does not mean every provider, service or workload will grow at the same rate.
What do current forecasts say about when growth might stop?
They do not identify a stop-growth year. Gartner’s forecasts remain positive over the periods they cover, though the estimates changed between releases:
#1 Best Overall
| Forecast and publication date | What it projected |
|---|---|
| Gartner, May 2024 | Worldwide end-user spending on public-cloud services of $675.4 billion in 2024, up 20.4% from $561 billion in 2023. Its table projected $824.763 billion in 2025, a 22.1% total-market growth rate. |
| Gartner, June 2024 | A public-cloud-services market of $1.28 trillion in current U.S. dollars by 2028, with a 20.0% compound annual growth rate in constant dollars from 2023 through 2028. |
| Gartner, November 2024 | A revised 2025 public-cloud-spending forecast of $723.4 billion and 21.5% growth. This differs from the May estimate; it is a later forecast baseline, not evidence that growth had stopped. |
The different 2025 estimates should not be treated as a single settled figure: they came from forecasts published at different times. Taken together, they support continued growth in public-cloud spending through the forecast horizon, not a precise date when cloud computing as a whole will stop expanding.
Why is demand still increasing?
AI adds new workloads
Generative AI requires computing for both model training and the repeated use of trained models, known as inference. Gartner attributed much of the expected increase in public-cloud spending to GenAI-enabled applications being deployed at scale. AI is therefore an important source of demand, although its growth does not guarantee that every AI workload will run in public cloud; cost, hardware access and location can affect placement.
Rank #2
Modernization and distributed systems broaden cloud use
Organizations continue to modernize applications and use distributed, cloud-native and multicloud systems. Those approaches can put computing in more than one environment rather than requiring a complete move to a single public-cloud provider. Gartner’s November 2024 update said hybrid-cloud adoption would remain widespread: it forecast that 90% of organizations would adopt a hybrid-cloud approach through 2027.
What could slow or reshape growth?
Electricity and data-center capacity
Power availability is a potential infrastructure bottleneck, especially for AI-heavy data centers. Gartner’s 2024 analysis forecast that 40% of existing AI data centers could be operationally constrained by power availability by 2027. It also estimated that incremental demand from AI-optimized servers could reach 500 terawatt-hours in 2027, 2.6 times the 2023 level.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →Rank #3
In a June 2026 forecast, Gartner projected data-center electricity consumption of 565 terawatt-hours in 2026, up 26% from 447 terawatt-hours in 2025, and more than 1,200 terawatt-hours by 2030. Gartner said AI capacity was already constrained by power availability. These are forecasts of data-center electricity use, not measures of cloud revenue. They point to a scenario in which grid capacity, permits, interconnection and cooling affect how quickly infrastructure can be added, even while customers want more computing.
Cost control and governance
Cloud use can grow while customers try to reduce waste or move selected workloads elsewhere. In Flexera’s 2025 survey of 759 cloud decision-makers, 84% called managing cloud spend their top challenge; 28% expected cloud spending to increase, 17% said they had exceeded budgets, and respondents estimated 27% of IaaS/PaaS spending was wasted. These survey findings describe reported management concerns and estimated waste, not a measured decline in total cloud demand.
Flexera reported that 21% of workloads had been repatriated, while migration and net-new workloads outpaced exits. That is evidence of selective movement, not an aggregate reversal. Its 2026 report described 73% of organizations as operating hybrid estates, 58% as using public-cloud GenAI services, and estimated wasted IaaS/PaaS spend at 29%. Together, the findings suggest more attention to governance, value measurement and workload placement as cloud estates become more complex.
Skills and workload economics
Cloud is not automatically cheaper or better for every application. Costs depend on utilization, architecture, data movement and the services used; operating multiple environments also takes people with the skills to manage them. As organizations mature, they may optimize, renegotiate, redesign or relocate individual workloads rather than keep moving everything to public cloud. That can change the mix and rate of growth without ending the broader market’s expansion.
Recommended Free Tools
Best Value
Does cloud growth mean everything moves to public cloud?
No. Hybrid estates combine public cloud with private infrastructure or on-premises systems; multicloud use spans more than one provider. These patterns can continue growing even when some workloads stay local or return from a public cloud. The best location depends on the workload rather than a universal rule:
- Total cost and utilization: Compare the full operating cost at realistic usage levels, including management and data-transfer costs.
- Latency and data locality: Applications that must respond quickly or keep data near a particular system may benefit from local or regional infrastructure.
- Regulation and sovereignty: Legal or contractual requirements can determine where data and processing are allowed.
- Resilience and portability: Consider failure recovery and the effort required to move or run an application in another environment.
- AI accelerators, power and cooling: Hardware availability and facility constraints can matter as much as the nominal cloud service offering.
- Operational skills: A technically suitable deployment still needs staff and processes to secure, monitor and manage it.
Synergy Research Group counted 1,189 hyperscale data centers at the end of the first quarter of 2025. It said hyperscalers represented 44% of worldwide data-center capacity then and projected their share to reach 61% by 2030, while on-premises capacity fell to 22%. This is a capacity forecast, not a revenue-share forecast: it suggests that cloud infrastructure growth may increasingly be concentrated among large hyperscalers even as organizations use a mix of deployment locations.
What should businesses watch instead of a stop date?
A single date obscures the decisions that determine whether a cloud strategy is working. For planning, track the measures that matter to your organization and revisit them as workloads and prices change:
- Cost per transaction, user or other meaningful unit of output—not just the monthly bill.
- Utilization and idle capacity, alongside commitments and workload growth.
- Performance, availability and recovery requirements for each application.
- Data location, transfer needs and applicable regulatory obligations.
- Power, cooling and accelerator availability for infrastructure-dependent workloads.
- The staffing and governance needed to operate public, private and hybrid environments securely.
For the market-wide question, watch dated forecasts and distinguish spending from capacity and electricity consumption. A later forecast revision may change the expected scale or pace; it does not by itself establish that cloud growth has ended. Any precise claim that the market will stop in a particular year remains speculation unless a dated forecast actually supports it.
Free tools Windows power users keep installed
One-click scans. No signup required.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
