Public-cloud usage accelerated dramatically in the second quarter of 2026, according to a fresh report from alternative-data firm UncoverAlpha. The firm’s channel checks and expert interviews suggest year-over-year growth climbed to 25–35% in Q2, up sharply from 10–20% in the previous quarter. The report, published July 17, also indicates that hyperscaler data centers remain effectively sold out — a signal that demand continues to outstrip supply across the industry.

The analysis, based on aggregations from multiple alternative-data providers and platforms like AlphaSense, captures sentiment from enterprise buyers and industry insiders ahead of official earnings releases. “The sentiment from expert interviews this quarter is very bullish,” the report states. While the free preview does not name which cloud vendor saw the strongest momentum, it describes “strong usage across the board,” suggesting that no single provider is driving the uptick.

That broad-based acceleration has immediate implications for anyone relying on Microsoft Azure, Amazon Web Services, or Google Cloud. If usage is truly rising this fast, capacity that was already tight could become even scarcer — especially for AI-related workloads that have been the primary growth engine for cloud providers over the past 18 months.

A sudden leap, measured in percentages

The numbers at the center of the report leap off the page: growth estimates moving from 10–20% year-over-year in Q1 to 25–35% in Q2. UncoverAlpha compiled these figures by analyzing hundreds of expert interviews drawn from AlphaSense, supplemented by supplier checks and alternative-data feeds that track everything from job postings to credit-card transactions.

The sell side had already been optimistic heading into the quarter. Analyst consensus typically clusters near the high end of the guidance ranges that Microsoft, Amazon, and Google issued after their Q1 calls. But UncoverAlpha’s work stands out because it goes beyond desktop estimates and into the messy, real-world signals that precede official revenue numbers.

Sell-side expectations for cloud revenue are generally clustered near the top ends of guidance ranges already issued by the major providers, according to the report. Its interview data shows strong usage across public cloud, not merely a rebound at a single vendor.

That distinction matters heading into quarterly results. Hyperscaler earnings are increasingly shaped by whether they can turn AI-related demand into recognized revenue quickly enough, rather than by demand generation alone. If capacity remains constrained, strong bookings or usage signals may not immediately translate into reported revenue; power delivery, data-center completion, server availability and customer deployment schedules can all become limiting factors.

The report also flags Microsoft Copilot adoption as part of the paid analysis, but provides no public figures, named customer data or specific adoption rate. Readers should therefore treat the reference as an indication of the report’s focus, not independent evidence that Copilot revenue or seat growth materially changed during the quarter.

What it means for your cloud strategy

For enterprise IT teams running workloads on Azure or any public cloud, the immediate takeaway is that capacity is not getting looser anytime soon. If the 25–35% growth figures hold true, the hyperscalers could be even more constrained heading into the second half of 2026 than they were in the first.

Organizations that need to spin up large GPU clusters for AI training or inference should start conversations with their cloud account teams now. Lead times for specialized instances have already been stretching; a fresh wave of demand could push them out further. While the report doesn’t confirm that any specific Azure region or SKU is unavailable, the pattern of broad, accelerating usage suggests that planning for scarcity is the prudent move.

Developers building on Azure AI services should keep a close eye on service availability and regional capacity. During previous periods of high demand, some AI models and APIs experienced throttling or limited availability in certain datacenters. If Copilot and other AI-driven services are indeed seeing faster adoption — even if the exact numbers remain under wraps — that would only compound the pressure on underlying infrastructure.

For IT professionals who manage Microsoft 365 environments, the mention of Copilot’s momentum is a reminder that AI features are becoming embedded deeper into everyday tools. Even without specific growth figures, an accelerating cloud market signals that the platform underneath these services — Azure — will need to scale to match. That could influence M365 feature rollouts and performance in the months ahead.

How we got here: a multi‑year capacity crunch

The cloud industry has been riding an AI-fueled wave since early 2025, when enterprise adoption of generative AI tools began to outpace data-center construction. Hyperscalers have responded with historic capital spending. Microsoft alone has committed tens of billions of dollars to expand its global Azure footprint, yet demand continues to run ahead.

The Q1 2026 earnings season revealed that even with aggressive investment, capacity was “sold out” at multiple providers. Guidance calls from Microsoft, Amazon, and Google consistently emphasized that revenue growth was being gated by supply, not demand. CFOs described order books that stretched well beyond what they could deliver in a single quarter.

UncoverAlpha’s Q2 data slots into that narrative cleanly. If usage growth has indeed accelerated from a 10–20% pace to 25–35%, it suggests that enterprises are not only sticking with their multi-year cloud commitments but are also accelerating their migrations and new AI projects. The underlying drivers — AI model training, inference at scale, data modernization — show no signs of slowing.

What’s new in the Q2 report is the speed of the change. A jump of that magnitude in a single quarter is unusual and points to a market inflection. Whether it reflects a one-time surge or the start of a new baseline won’t be clear until the hyperscalers report their own numbers and offer guidance for the second half.

What to do now: practical steps before earnings

The next major checkpoint will be the Q2 earnings disclosures from Microsoft, Amazon, and Google, likely starting in late July or early August. Those reports will either validate the alternative data or show that usage signals did not fully convert into recognized revenue. Until then, here’s what you can do:

  • Reassess your own cloud demand. If your organization is planning any significant Azure deployments for Q3 or Q4, reach out to your Microsoft account team to understand regional capacity and potential lead times. Include GPU‑accelerated instances in your assessment, even if you’re not yet training large models.
  • Review reservation and savings plans. Locking in compute capacity for 1‑ or 3‑year terms not only offers cost benefits but can sometimes provide access to capacity that might otherwise be reserved for strategic customers. If you have workloads you know you’ll run for the next 12 months, converting to reserved instances can add a layer of assurance.
  • Watch for Copilot announcements. While the UncoverAlpha report doesn’t provide hard Copilot adoption data, Microsoft may use the earnings call to share seat growth or revenue figures. If your organization is rolling out Copilot for Microsoft 365, a strong quarter for the service could accelerate product updates and feature releases.
  • Consider multi‑cloud flexibility. If capacity in your primary region becomes tight, having a secondary cloud provider can offer a pressure release. Even a light commitment to a secondary provider can give you negotiating leverage and an insurance policy against local supply shocks.
  • Temper expectations for near‑term price drops. With demand outpacing supply, deep discounting on cloud services is unlikely. Expect pricing to remain stable or even firm up in high‑demand areas.

The outlook: earnings season as the reality check

The 25–35% growth estimates from UncoverAlpha represent the bullish edge of what alternative data can capture. As the firm itself notes in its methodology, these signals come from a sample of experts and channel partners—not from exhaustive financial filings. Respondents may be clustered in particular industries or regions, and “usage” can mean different things to different sources.

Still, the signal is strong. If the hyperscalers report Q2 revenue that aligns with this acceleration, it will confirm that cloud adoption has entered a new gear. For Microsoft, that would likely mean Azure growth rates comfortably above 30%, possibly pulling other parts of the business — including Copilot and security services — along with it.

If, instead, the reported revenue lags the usage signal, it would point to deeper supply‑chain bottlenecks or slower customer deployment cycles. Either way, the Q2 earnings calls will offer the first tangible look at whether the cloud market has truly stepped into a higher growth plane — and whether enterprises need to adjust their strategies accordingly.