For busy readers
- Perplexity signed a $750 million, three-year cloud deal with Microsoft to use Azure and its Foundry AI platform.
- The partnership gives Perplexity access to multiple top AI models (OpenAI, Anthropic, xAI) via Microsoft Foundry.
- Perplexity has not abandoned Amazon Web Services (AWS) as its primary cloud provider — signaling a multi-cloud strategy rather than a full migration.
What the deal actually covers
Originating from a Bloomberg report, the agreement commits Perplexity to spend approximately $750 million with Microsoft’s Azure cloud over three years, primarily through the Microsoft Foundry program — a higher-level AI platform that hosts and sources models from various developers, including OpenAI, Anthropic, and xAI.
In practical terms, this means Perplexity will use Azure not just as “servers,” but as the operational layer where it runs, tests, and scales the AI models that power its products — particularly its advanced generative search services.
Why this matters to Perplexity
Perplexity has gained attention as an AI-driven alternative to traditional search and assistant tools, blending generative models with web retrieval and reasoning. It has continued to build on Amazon Web Services (AWS) as its primary cloud infrastructure, especially for compute and storage, but this Microsoft deal gives it flexibility and access to a broader set of models and tooling.
A spokesperson for Perplexity told Bloomberg that the partnership with Microsoft is intended to expand access to frontier models and does not represent abandoning AWS — reinforcing that many AI companies today deliberately architect across multiple cloud providers.
This kind of diversification helps Perplexity reduce vendor lock-in, balance costs, and pick the best environments for different workloads — from dataset training to inference and customer delivery.
What Microsoft gets out of this
For Microsoft, the deal does more than cash inflow:
- Azure and Foundry become the de facto multi-model AI platform for emerging AI innovators, strengthening their competitive positioning against rivals like Alphabet’s Google Cloud and AWS.
- It brings one of the most closely watched AI startups into its cloud ecosystem — signaling that Azure isn’t just for legacy enterprise workloads, but also for next-generation AI infrastructure.
- Microsoft also continues to demonstrate its commitment to multi-model support — a key CEO Satya Nadella theme — by enabling startups to tap models from various providers under one roof.
Azure’s push for AI customers is strategic: cloud usage is shifting from basic storage and compute to platforms where models are hosted and orchestrated at scale — which brings more value and stickiness than raw server hours.
A broader trend: multi-cloud, not mono-cloud
While the headlines talk about a $750 million “switch,” a more nuanced reading shows a multi-cloud world emerging.
Perplexity’s statement is telling: “AWS remains our preferred cloud provider” — and spending commitments with Amazon Web Services remain intact even as they adopt Azure.
This dual approach reflects a growing pattern among AI startups:
- AWS for core infrastructure and cost-effective compute
- Azure (Foundry) for model access, orchestration, and hybrid workloads
Such strategies help companies manage risk and performance, as well as negotiate better pricing and service levels from providers vying for AI workloads.
Why this is about more than money
The Perplexity-Microsoft deal is part of a larger cloud and AI industry evolution:
? AI workloads drive cloud contracts
Unlike traditional apps, AI systems require not just servers, but:
- specialized GPUs
- model hosting layers
- multi-vendor access
- data pipelines
Companies that win these contracts today could shape how competitors build and deploy AI tomorrow.
? No one cloud rules all
The fact Perplexity still uses AWS while signing a major Azure deal reflects a future where cloud strategies are modular not exclusive — just as enterprises increasingly adopt multi-cloud and hybrid architectures.
? Foundry as a neutral layer
Microsoft’s Foundry platform positions itself not just as a Microsoft service, but as a cross-vendor AI hub, where developers can pick the best model for the job — even from competitors of Microsoft’s own AI partners.
This helps Azure compete on flexibility and choice, not just raw horsepower.
What this means for Perplexity going forward
Perplexity’s ambition is to be a next-generation search and reasoning platform — one that blends generative AI with retrieval and real-world answers. Deals like this provide:
- Infrastructure stability and capacity to scale
- Access to a wider ecosystem of models
- Flexibility in deployment and performance
But it also places Perplexity within a competitive cloud war where every hyperscaler wants to own the stack above and beyond commodity compute.
For Perplexity, the new cloud partnership could fuel growth and innovation — so long as it can juggle multiple providers and keep operational complexity under control.
and Before you go
Perplexity’s $750 million deal with Microsoft isn’t just big — it’s strategic, marking a shift from one-vendor cloud reliance to a more flexible, multi-platform future for AI startups. In an industry increasingly defined by both performance and choice, that kind of arrangement could become not the exception — but the rule.
