Microsoft has scaled its capital expenditure (CAPEX) with a multi-billion dollar bet on its Azure AI infrastructure that’s setting up the tone for Wall Street’s valuation of the Big Tech giant.
Microsoft is no longer selling an AI story to the financial bull, but is building its Azure AI infrastructure through cloud deals, such as the $750 million partnership with Perplexity AI to lock in compute, scale, and control over how modern AI products actually run.
Microsoft Corporation (NASDAQ:MSFT) may be facing market skepticism after its latest earnings, but beyond the stock volatility, Microsoft is betting quietly on its AI backbone that powers others’ ambitions while monetizing scale, long-term cloud contracts, and data center dominance.
The Azure-parent is experiencing a surge in spending, highlighted in recent financial disclosures, coinciding with the Perplexity AI deal. Microsoft wants to dominate the next generation of search and discovery, and this pivot in aggressive fiscal posture has caught the attention of influential market analysts.
“We are only at the beginning phases of AI defusion, and already Microsoft has built an AI business that is larger than some of uor biggest franchises,” said CEO, Satya Nadella, in a statement about the Microsoft Azure AI vision.
Earnings Show AI Costs Cloud Growth
On January 29, Microsoft announced to the world its unprecedented spending on AI, but on the expense of slower cloud growth.
In its fourth quarter (Q4) earnings, Microsoft shares slid by 6.5% in late trading as investors looked past a revenue beat, focusing on the soaring cost of maintaining a place in the AI era.
Despite a 39% climb in Azure’s cloud sales – topping analysts’ estimates – Microsoft’s total revenue of 17% to $81.3 billion was eclipsed by a 19% increase in expenses. It’s all for Azure AI infrastructure.
Driving this expense spike was a colossal $37.5 billion in capital investment, a 66% Year-over-Year (YoY) jump, dominantly driven by the giant’s aggressive purchase of advanced computing chips. Microsoft spent $37.5 billion on data centers and equipment in the quarter, adding nearly one gigawatt of capacity. Demand, however, still outpaces supply.
Adjusted earnings per share reached $4.14, Azure growth of 39% slightly missed forecasts.
Microsoft disclosed it now has 15 million paid Microsoft 365 Copilot seats. While impressive, UBS analysts needed to evaluate the cloud computing company Microsoft azure on AI, highlighting that that “M365 revenue growth is not accelerating due to Copilot,” raising questions about return on massive compute investments, compared to Azure.
Microsoft Copilot vs Perplexity
Perplexity, by contrast, represents a different bet. Rather than pushing a single branded assistant, Microsoft spending on AI gains visibility into how modern AI search operates across models, queries, and user behavior.
Is the Microsoft AI infrastructure also collecting large telemetry on how modern LLM search operates? Is the Perplexity deal a move to own the plumbing of every AI search player? The Azure AI infrastructure suggests the answer may be yes.
As for the Microsoft Azure Open AI feud, for now, it remains fragile.
Nearly 45% of Microsoft’s $625 billion cloud backlog comes from OpenAI’s $250 billion commitment, which is concerning when assessing concentration risk. Jefferies analyst, Brent Thill, questioned whether OpenAI can meet those financial obligations.
OpenAI moved aggressively into politics and public influence and ran into problems and that doesn’t really fit the path Microsoft’s choosing to walk on, as it’s pivoting away from that exposure, anchoring neutrality in the Azure AI infrastructure.
Microsoft understands what the obvious truth should be, that this is not sector versus sector, but company versus company. Whoever wins the AI race will be the one with the strongest data centers and compute capacity.
Designing and Implementing a Microsoft Azure AI solution
According to Bloomberg, the Perplexity deal allows the AI search company to deploy models from OpenAI, Anthropic, and xAI through Microsoft’s Foundry service.
Strategically, this diversifies Perplexity away from its primary partner, Amazon Web Services, while reinforcing Azure AI platform as a neutral one that can host competing AI models, side by side.
These matters because Microsoft is not trying to win AI by building every application itself. Instead of focusing on stock prices, this shows Microsoft as a quiet AI backbone, making money by hosting others’ ambitions through cloud deals, heavy spending, and long-term scale.
The company must decide whether to allocate compute to first-party products like Microsoft 365 Copilot or to external customers such as Perplexity, who bring predictable, high-margin volume.
This is where the Perplexity deal becomes strategic insurance with the deal it made to enhance Microsoft Azure AI services. Through the Perplexity deal Microsoft is relying on it as the quiet backbone to succeed in AI. It ensures the Azure AI infrastructure first, not politically entangled, and commercially neutral unlike OpenAI, which has increasingly faced scrutiny and friction with US regulators. Microsoft understands what the obvious truth should be, that this is not sector versus sector, but company versus company. Whoever wins the AI race will be the one with the strongest data centers and compute capacity.
Inside Telecom provides you with an extensive list of content covering all aspects of the tech industry. Keep an eye on our Intelligent Tech sections to stay informed and up-to-date with our daily articles.