Cloud Computing Watch: AWS's Concentration Problem Just Hit a New Record, AI Integration Race & The Global Hyperscaler Arms Race

The CODEW Cloud Computing Watch | August 24, 2026

AWS's Concentration Problem Just Hit a New Record — Even as Google Cloud Posts the Quarter's Biggest Growth Surprise

Executive Summary

Two Numbers From the Same Week Tell Opposite Stories About Cloud Risk and Cloud Growth

Cloudflare Radar's latest data, pulled on August 1, shows AWS's us-east-1 region took 44.33% of all AWS requests in July — the highest monthly reading in the tracker's history, with the weekly series ending the month at 46.4%. Northern Virginia's traffic gravity well is still deepening eighteen months after analysts first started flagging it as a concentration risk. In the same window, Google Cloud posted the quarter's biggest growth surprise: $24.8 billion in Q2 2026 revenue, up 82% year-over-year, overtaking both Azure's roughly 40% pace and AWS's 28% — and pushing Google's overall market share to 15%, up from 14% in Q1.

Those two numbers, read together, are this week's real story: the provider carrying the most concentrated single-region risk (AWS) isn't the one seeing the fastest re-acceleration in growth (Google Cloud), and neither of those trends is obviously connected to reliability, which this series has tracked all month separately. This edition updates the concentration data, the growth-rate leaderboard, and a broader shift Constellation Research calls the "AI vertical integration race" — where cloud lock-in is migrating from infrastructure up into models, agents, and data catalogs.

01 — The Cloud Lead

AWS's us-east-1 Concentration Hits a New Record, While Azure's De-Concentration Trend Stalls

AWS's us-east-1 region carried 44.33% of all AWS requests in July, per Cloudflare Radar's Cloud Observatory — its highest monthly reading since the firm began tracking regional traffic share, continuing a trend this series flagged three weeks ago when the figure sat in the low-40s. The weekly series ended July even higher, at 46.4%, meaning nearly half of all AWS traffic globally now routes through a single region.

Azure's picture is genuinely more nuanced than a simple "de-concentrating" story suggests. Year over year, Azure's global routing share is down 8.48 points — a real structural shift as newer regions like Brazil South, Central India, and Poland Central absorb growing real traffic. But month-to-month, that trend has stalled: July's reading of 49.18% is up 1.14 points from June, and the weekly series has crossed back and forth over the 50% line all summer. Azure is still meaningfully less concentrated than AWS on a like-for-like basis, but its own de-concentration curve isn't moving in a straight line.

02 — Hyperscaler Cloud Watch

Updated Q2 2026 figures reshuffle the growth-rate leaderboard meaningfully. AWS held steady at 28% global market share and $37.6 billion in Q1 2026 revenue (+28% YoY, per Synergy Research and Tech Insider), a position CRN's Q2 roundup confirmed held through the following quarter. Azure sits at a steady 21% share. Google Cloud is the real mover: revenue surged to $24.8 billion in Q2, up 82% year-over-year — eclipsing the roughly 40% pace Azure posted earlier in the year and far outstripping AWS's growth rate, while its share climbed from 14% to 15% quarter-over-quarter, the most significant share gain among the traditional Big Three this cycle.

Google's acceleration lines up with aggressive positioning moves: an 8% compute price cut across all regions in Q1, deep NVIDIA GPU partnerships showcased at Google Cloud Next 2026, and continued Gemini integration across its data and database tooling. Oracle remains the outlier on pure percentage growth — infrastructure revenue up 84% year-over-year in fiscal Q3 — but from a far smaller base than any of the traditional three.

03 — The AI Vertical Integration Race

Constellation Research's framing of the current cloud cycle as an "AI vertical integration race" captures a shift this series has been tracking piecemeal all month: AWS's launch of Connect business applications and Amazon Quick, Google Cloud Next's raft of agentic and lakehouse announcements, and continued hyperscaler results all point toward cloud providers racing to own a full AI-optimized stack rather than just compute and storage. Constellation's own risk framing is the more important part: as providers build that full stack, "lock-in shifts from compute, storage and databases to models, agents, vector stores, data catalogs" — a dynamic that gives fresh context to both Alation's confirmed cyberattack and Oakley Capital's acquisition of knowledge-graph platform Graphwise, covered in recent CODEW briefings. The layer where enterprise dependency concentrates is visibly migrating up the stack, from infrastructure toward the data and knowledge layers that make AI agents actually useful.

04 — Enterprise Cloud Adoption

Multi-cloud adoption has reached 89% of enterprises, up from 76% in 2024 — direct, corroborating evidence for the interconnect buildout (AWS-Google, Oracle-AWS) this series covered earlier this month. Separately, an Information Services Group report finds growing AI compute demands, cost, and sovereignty concerns are pushing more enterprises toward hybrid cloud specifically, while Unisys research finds business leaders expect agentic AI to play a key role in cloud management going forward, even though few organizations have moved past pilot-stage deployments. Enterprises are simultaneously spreading workloads across more providers and expecting AI agents to manage that growing complexity — two trends that will only become harder to reconcile as the vertical-integration dynamic in Section 03 deepens.

Three Cloud Signals

1. AWS's regional concentration risk is still deepening, not plateauing. A new all-time-high reading for us-east-1 confirms the trend flagged three weeks ago is accelerating rather than stabilizing, even as AWS's headline market share holds steady.
2. Growth-rate leadership keeps rotating. Google Cloud's 82% year-over-year surge is now the standout number among the traditional Big Three, a genuine reordering from earlier in the year when Azure held the percentage-growth lead.
3. Vertical lock-in is migrating up the stack. As hyperscalers race to own the full AI stack — models, agents, vector stores, data catalogs — the concentration and dependency risk this series has tracked at the infrastructure layer is reappearing one layer up, visible directly in incidents like the Alation breach.

The CODEW Take

This week's data confirms that cloud risk and cloud growth are running on genuinely independent tracks right now. AWS's record us-east-1 concentration hasn't dented its market share or its ability to hold steady growth; Google Cloud's blowout quarter isn't obviously tied to superior reliability or a more distributed infrastructure footprint — it's tied to pricing, GPU partnerships, and enterprise AI positioning. Enterprises appear to be making cloud decisions on cost, capability, and vendor relationship far more than on the resilience metrics this series tracks closely, at least until an incident large enough to change that calculus actually occurs.

What to watch next: whether Google Cloud's 82% growth rate proves durable into Q3 or reflects one-time contract timing, and whether the "vertical integration race" Constellation describes produces its own concentrated-dependency incidents at the model and agent layer — the way Alation's breach did at the data-catalog layer this month — as more of the AI stack consolidates inside each hyperscaler's own full-stack offering.

---

THE CODEW · CLOUD COMPUTING WATCH

Editorial Note

The CODEW Cloud Computing Watch examines the strategic moves of hyperscalers, AI cloud infrastructure, data center capacity, cloud economics, and the competitive battle for the next generation of cloud infrastructure. It focuses on who controls the infrastructure layer on which the AI economy runs.

Coverage is based on company announcements, public disclosures, industry reporting, and other publicly available information. Analysis reflects the reporting period and should be considered in the context of the sources and developments cited.

```