Morgan Stanley: Cloud growth accelerates in the third quarter as AI drives demand for core services and software

Why did growth at the three largest cloud providers accelerate again?

In the third quarter of 2026, combined cloud revenue at AWS, Azure and Google Cloud grew approximately 56% year over year, up from 48% in the second quarter. The Azure figure uses Microsoft's reporting basis after its business resegmentation.
AWS grew 39% year over year at constant currency, compared with approximately 37% in the second quarter, marking its sixth consecutive quarter of accelerating growth. Azure grew 48% at constant currency, up from approximately 42% in the second quarter. Microsoft attributed the growth to improved monetization of Foundry and CPUs, while also mentioning delays in third-party capacity.
Google Cloud grew 103% year over year, up from approximately 82% in the second quarter, with TPU system sales making a noticeable contribution. Morgan Stanley estimates that core GCP growth excluding TPUs was approximately 71%, broadly unchanged from the second quarter.
AI demand is also driving traditional cloud services: Amazon said revenue from core compute and storage grew at its fastest rate in years, while Microsoft and Alphabet also reported strength in both AI and core cloud services.

Amazon said inference now accounts for more than half of AWS's AI revenue; the number of agents deployed by AgentCore customers in the third quarter was twice the combined total for the preceding three quarters. Morgan Stanley believes that more agent activity will increase the calls, traces and token costs that need to be monitored.
Alphabet said building agents has driven database usage because agents need data as context for inference. On that basis, Morgan Stanley is positive about the overall demand for operational databases as agents enter production.

Most of Snowflake's customer consumption takes place on AWS, so Morgan Stanley sees AWS's acceleration as a favorable backdrop for its product revenue. However, Amazon's acquisition of DuckLabs and Google's reduction of BigQuery query-processing costs show that cloud providers are competing for analytics business through openness and pricing.
In knowledge work, Alphabet and Microsoft are also targeting the same enterprise users as Snowflake CoWork. Morgan Stanley sees this as medium-term competition at the user entry point and context layer, while also believing that agents querying existing data could increase Snowflake consumption.
Datadog's core observability business tracks overall cloud consumption. The bank's channel checks indicate that partners performed above plan in the third quarter and customer consumption exceeded initial commitments; however, its largest AI-native customers remain a separate variable for results.
In its earnings preview, Morgan Stanley identified approximately 33%–34% year-over-year revenue growth for Datadog in the third quarter and a fourth-quarter growth guide in the upper half of the 20% range as the threshold. Microsoft's positioning of Agent 365 as an AI observability platform also creates competition.

MongoDB previously identified using enterprises' own data to provide real-time context for AI systems as a key use case. Increased database usage driven by agents supports demand for this business.
Competition is also intensifying: Microsoft said its AI-optimized databases are being rapidly adopted for agent data and context, with PostgreSQL growth accelerating for four consecutive quarters.