Morgan Stanley: Cloud Computing Demand Accelerating, Snowflake, Datadog, and MongoDB Poised to Benefit
nashnova research
Morgan Stanley says the big three cloud providers posted combined Q3 revenue of roughly $109.9 billion, up 55.7% year-on-year, as AI-driven demand spills over into traditional compute and data services — setting a favorable backdrop for upcoming results at Snowflake, Datadog, and MongoDB.
How strong were the big three's Q3 cloud numbers?
Amazon, Microsoft, and Google posted combined cloud revenue of roughly $109.9 billion, up 55.7% year-on-year — accelerating from Q2's 47.9%.
AWS grew 39% in constant currency, its sixth straight quarter of acceleration and the fastest pace in nearly five years. Azure grew 48%, up from 42% in Q2.
Google Cloud surged roughly 103%, but Morgan Stanley flags that the figure includes TPU — Google's custom AI chip system — sales. Strip those out, and core growth was about 71%, roughly flat with Q2. This means → the "triple-digit" headline overstates the underlying business; core growth needs to be read at a discount.
Why does AI spending lift the "old" cloud businesses too?
Morgan Stanley's central thesis: AI infrastructure investment is pulling demand not just for GPUs and specialized compute, but also for traditional compute, storage, databases, and analytics.
In plain terms = companies bought GPUs to run AI models, but those models still need to read data, store results, and be monitored — all of which runs on conventional cloud services.
Amazon disclosed that AI inference now accounts for more than half of AWS's AI revenue. Microsoft's Foundry platform revenue more than doubled year-on-year, and Agent 365 — its AI-agent platform — has registered over 50 million agents, up more than 30% quarter-on-quarter.
Snowflake: what does AWS acceleration mean for it?
Snowflake — a cloud data-warehouse platform — runs the vast majority of its workloads on AWS. Six straight quarters of AWS acceleration directly supports stronger product-revenue demand.
But competition is creeping in: Amazon acquired DuckLabs and plans to make DuckDB a universal SQL query layer for AWS storage analytics. Google's BigQuery improved query speed by over 40% year-on-year while cutting processing costs by more than 50%. Both Google and Microsoft are also building enterprise AI-agent products that overlap with Snowflake's use cases.
Morgan Stanley's read: these competitive moves are more of a medium-term challenge and do not yet indicate that Snowflake's current-quarter demand is being hit. If AI agents can query enterprise data stored on Snowflake directly, platform usage could actually increase.
Datadog: why does the monitoring business expand with cloud spend?
Datadog — a cloud-infrastructure monitoring platform — ties revenue directly to how much cloud customers consume. The big three's combined 56% revenue growth creates a favorable environment for its core monitoring, log-analytics, and APM products.
Morgan Stanley's channel checks show Datadog partners broadly outperformed their Q3 plans: actual customer usage exceeded initial commitments, with strong results in new-customer acquisition, multi-product adoption, and existing-account expansion.
The bank sees Q3 revenue growth of roughly 33–34% year-on-year and a Q4 guide implying close to 30% growth as the key benchmarks. This means → if results fall below that range, the market may question whether cloud-monitoring growth is as elastic as expected.
MongoDB: why do AI agents need an "operational database"?
Unlike conventional chatbots, enterprise AI agents need real-time access to internal company data to understand business context, execute tasks, and deliver accurate results. In plain terms = an agent doesn't just chat — it checks inventory, reads orders, and updates records, all of which require a database capable of real-time reads and writes.
Alphabet confirmed this logic in its earnings: AI-agent development is driving higher database usage as agents pull business-context data for reasoning. Morgan Stanley sees this as consistent with the thesis MongoDB laid out at its investor day.
But Microsoft's competitive pressure is real: its AI-optimized database products — including PostgreSQL, Cosmos DB, and others — are increasingly used by enterprises to store agent-related data, and PostgreSQL has accelerated for four consecutive quarters.
What is the common challenge all three companies face?
Hyperscaler acceleration → improved demand for downstream software companies. That is Morgan Stanley's underlying logic for a constructive view on Snowflake, Datadog, and MongoDB heading into earnings.
Yet each company faces a structural tension: AWS, Azure, and Google Cloud are expanding their own analytics, monitoring, and database product lines — making them simultaneously distribution platforms and direct competitors.
This reflects a trend now materializing across cloud software: AI agents are moving from experiment to production, expanding the total market, but hyperscalers are also pushing their own products into every layer — how much incremental demand reaches downstream companies is the central question for the coming earnings season.
市场有风险，内容仅供研究参考，不构成投资建议。