How Cloud and AI Actually Integrate (in the real world)
1) Data gravity → model gravity
- The public cloud centralizes an organization’s data. That “data gravity” pulls AI workloads (training, fine-tuning, inference) to the same cloud regions to minimize latency and egress costs.
- Result: tighter loops between data pipelines, vector stores, feature stores, and model endpoints—making AI apps cheaper, faster, and easier to operate at scale.
2) Reference architectures that ship
- Ingestion → Lakehouse → Feature Store → Model (LLM/GenAI) → Orchestration → Observability/Security.
- Managed services (serverless compute, GPUs/TPUs, vector DBs, event buses) mean teams spend time on product logic, not plumbing.
3) Multicloud + sovereign AI
- Many enterprises blend providers to chase price/perf, specialized silicon, or data-residency guarantees. Expect “model portability” (fine-tuning in one cloud, deploying in another) to become normal.
4) Economics that compound
- Cloud turns capex into opex; AI turns workflows into software. Together they raise gross margin potential for customers—and for the platforms selling the compute, storage, and AI SKUs.
Where the Growth Is (and why)
Microsoft Azure: Azure growth has re-accelerated, with recent quarters citing ~33% y/y for “Azure and other cloud services,” and management calling out ~12 points from AI services; summer updates suggested Azure growth nearing the high-30s as capacity comes online.
AWS (Amazon): AWS remains the scale leader and is re-accelerating as AI services (Bedrock, Trainium/Inferentia) kick in; Q2 2025 AWS revenue +17.5% y/y to $30.9B, with heavy AI capex near-term pressuring margins.
Google Cloud: Clear AI tailwind; Q2 2025 Google Cloud +32% to $13.6B with strong AI infra and gen-AI solutions; backlog expanding and capex rising to meet AI demand.
NVIDIA: The hardware flywheel; Q2 FY26 revenue $46.7B (+56% y/y) with Blackwell data-center strength, fuel for every cloud’s AI expansion.
Oracle (OCI): A breakout second wave; cloud revenue up ~27–28%, with management guiding ~77% FY26 OCI growth and RPO soaring, benefiting from multicloud database + AI GPU demand.
Risks & Realities (Greenhouse view)
GPUs & power constraints: Compute scarcity and power availability can cap near-term AI revenue recognition.
Unit economics: AI margins improve as utilization rises and inference shifts to cheaper silicon; until then, capex and COGS can compress cloud margins.
Vendor concentration: Multicloud and open tooling reduce lock-in but also spread spend, which can blur who “wins” any single quarter.
Stocks to Watch: Cloud + AI Leaders with Momentum
Not financial advice. We look for durable growth, strong balance sheets/FCF, and visible AI catalysts.
Microsoft (MSFT) — AI demand lifting Azure
- Azure growth mid-30s with ~12 pts from AI; recent reporting pointed to an acceleration toward high-30s. Microsoft Cloud margins diluted near-term by AI buildout, but scale benefits are kicking in.
Amazon (AMZN) — Scale leader, AI flywheel
- AWS +17.5% y/y to $30.9B (Q2 ’25); heavy AI capex weighs on margins short-term but underpins higher-value Bedrock and custom silicon adoption.
Alphabet (GOOGL) — AI-first cloud catching up
- Google Cloud +32% to $13.6B (Q2 ’25); strong AI infra & gen-AI uptake; capex lifted to ~$85B for data centers and GPUs.
NVIDIA (NVDA) — Picks & shovels of AI
- $46.7B in quarterly revenue (+56% y/y) with data-center growth from Blackwell; direct lever on every cloud’s AI capacity.
Oracle (ORCL) — Multicloud database + AI infra surprise
- Cloud revenue up ~27–28%, OCI guided +77% FY26; RPO exploded to $455B; multicloud DB + AI GPU demand driving bookings.
ServiceNow (NOW) — Agentic AI workflows at scale
- Q2 ’25 subscription revenue $3.11B (+~23% y/y); raised full-year guide; customers with $20M+ ARR up ~30%. AI (“Now Assist/agentic”) is a demand catalyst.
Datadog (DDOG) — Observability/SecOps for AI-era stacks
- Q2 ’25 revenue $827M (+28% y/y); strong large-customer growth and lifted FY guidance; expanding AI observability products.
Snowflake (SNOW) — Data cloud feeding AI apps
- FY’25 product revenue +$943M in Q4 (+28% y/y); NRR 126%; AI features (e.g., Cortex) driving consumption and guidance raises into FY26.
Bonus: CrowdStrike (CRWD) sits where cloud and AI meet security: ARR $4.66B (+20% y/y) with robust net-new ARR—useful exposure to securing AI workloads.
What to Watch Next (for readers/investors)
Utilization > headlines: Track GPU capacity, region launches, and AI service utilization (not just capex) to gauge revenue conversion.
AI SKUs becoming “standard”: As AI copilots/agents get bundled into suites (M365, Workspace, NOW), pricing power and stickiness rise.
Data moat > model hype: The defensibility moves to high-quality proprietary data pipelines and governance across clouds.
Greenhouse Takeaway
Cloud + AI isn’t a “feature” it’s the new operating system for business. The leaders above marry scale compute with AI platforms, and the downstream ecosystem (observability, data clouds, security) captures recurring spend as enterprises turn workflows into software.
Leave a Reply