Microsoft Perplexity deal puts Azure multi-model AI strategy under scrutiny

• Microsoft (NasdaqGS:MSFT) has entered into a $750 million, multi-year partnership with AI startup Perplexity.
• Perplexity uses Microsoft Azure’s Foundry service to run OpenAI, Anthropic, and xAI models on Azure infrastructure.
• This agreement emphasizes Azure’s positioning as a multi-model AI hosting platform, rather than a single model provider channel.

If you’re an investor focused on Microsoft (NasdaqGS:MSFT), this deal sits at the intersection of cloud and generative AI, two areas central to the company’s long-term strategy. Azure provides computing power and positions itself as a neutral location with one-stop access to various model providers. This could be important for customers who need flexibility across OpenAI, Anthropic, xAI, and potentially others without committing to a single stack.

The Perplexity deal also highlights key questions investors are asking about Microsoft’s AI spending, including capital intensity, customer concentration, and timing of potential benefits. As Azure signs on to more multi-model workloads, the mix of customers and use cases using AI infrastructure is likely to become a key factor to watch alongside the key numbers for overall cloud demand.

Stay up to date with the most important news stories about Microsoft by adding Microsoft to your Watchlist or Portfolio. Or explore our community and discover new perspectives on Microsoft.

How Microsoft stacks up against its biggest competitors

Perplexity’s three-year, US$750 million commitment signals Microsoft’s intention to make Azure a neutral, multi-model hub where AI customers can mix and match models such as OpenAI, Anthropic, and xAI on one platform. The takeaway for you is that this will add new large-scale AI workloads to Azure, especially at a time when investors are watching to see how Microsoft translates its significant AI infrastructure spending into contract usage, especially as rivals like Amazon Web Services and Google Cloud are also courting AI startups and enterprises.

How does this fit into the Microsoft AI story?

The deal with Perplexity is consistent with Microsoft’s existing claims that it wants Azure to be the default place for enterprises to build AI-powered applications, in addition to Copilot, GitHub, and sector-specific services such as healthcare and finance. It also expands the picture slightly beyond OpenAI, which is important given recent concerns that a large portion of Microsoft’s commercial backlog is tied to a single AI partner.

Risks and benefits investors should consider

• Another avid AI customer supports the use of Azure at a time when their remaining commercial performance obligations are already significant.
• Multi-model access can make Azure more attractive compared to Amazon or Google for companies that don’t want to be locked into one AI stack.
• The economic value of US$750 million will depend on contract-specific terms, including when capacity is available, which models are ready for production, and which regions can charge per token.
• In a multi-cloud setup, such as Perplexity using both AWS and Azure, margins are not guaranteed for this type of workload as network and data transmission costs can increase.

What to watch next

From here, it will be worth tracking whether Microsoft discloses any larger multi-model Foundry deals, how quickly such contractual commitments turn into reported Azure revenue, and whether customer concentration on OpenAI eases over time. If you’d like to see what other investors think about these trade-offs, you can check out the community narrative around Microsoft through this dedicated page and compare different long-term views on this new partnership.

This article by Simply Wall St is general in nature. We provide commentary using only unbiased methodologies, based on historical data and analyst forecasts, and articles are not intended to be financial advice. This is not a recommendation to buy or sell any stock, and does not take into account your objectives or financial situation. We aim to provide long-term, focused analysis based on fundamental data. Note that our analysis may not factor in the latest announcements or qualitative material from price-sensitive companies. Simply Wall St has no position in any stocks mentioned.

Do you have feedback on this article? Interested in its content? Please contact us directly. Alternatively, email editorial-team@simplywallst.com.