Microsoft Corp. (MSFT) turned 50 in 2025. It's been public since 1986. It's been scrutinized, debated, and reinvented more times than most companies get the chance to try. And yet, according to Goldman Sachs, the next decade could look a lot like its best one.

In a note released Monday, Goldman analyst Gabriela Borges, CFA, resumed coverage of Microsoft with a Buy rating and bumped the firm's price target from $630 to $655. That implies about 37% upside from where the stock closed on January 9. The thesis? Microsoft isn't just riding the AI wave—it's positioned across the entire stack, from silicon to software, in a way that lets it compound growth across multiple product cycles simultaneously.

The Market Is Still Skeptical, Goldman Says That's the Opportunity

Borges argues that investors are still viewing Microsoft through an outdated lens. The prevailing worry is that AI spending is a near-term drag with uncertain payoffs. But Goldman sees something different: a company that's systematically building leverage across every layer of the AI stack, from infrastructure to applications.

Even after five decades, Goldman describes Microsoft as "still one of the best secular growth stories in Technology." The next phase of value creation, Borges suggests, isn't about a single blockbuster product. It's about compounding adoption across the entire ecosystem.

There are still "areas of discovery value" in the stock, she notes—places where investor sentiment is more pessimistic than the competitive reality warrants. Goldman also points out that Microsoft has structured its AI strategy to "win in multiple ways while limiting its downside to any one particular vendor or approach," essentially optimizing what the firm calls its "sharpe ratio of earnings power."

Azure AI: Growth, Margins, and the Long Game

Goldman projects Azure AI revenue will grow at a compound annual rate of 66% through fiscal 2030. That's not a typo. The firm believes the growth will be supported by ongoing investments and a strategic tilt toward inference—real-time model execution—rather than training, which is capital-intensive and delivers thinner margins.

Microsoft estimates that roughly 30% of its fiscal 2026 compute capacity will monetize directly as Azure AI revenue, with another 45% tied to traditional Azure workloads.

Margins are improving too. Borges notes that Azure AI gross margins have climbed from negative 50% in fiscal 2024 to about 17% in fiscal 2025. The long-term target? A return to pre-AI Azure gross margins near 60%. That would be a significant inflection point.

The Capex Question Everyone Keeps Getting Wrong

One of the biggest reasons Goldman thinks the market is mispricing Microsoft is the way investors interpret capital spending. The common assumption is that capex is a direct leading indicator for Azure revenue growth over the next couple of quarters. Borges pushes back hard on that.

She highlights Microsoft's own guidance that "not all capex translates directly to Azure revenue growth." In a supply-constrained environment, the company is also allocating capital to internal AI projects given their strategic importance, and to first-party applications, which Borges estimates have "structurally better unit economics than Azure today."

Goldman forecasts Microsoft will spend $148 billion on capital expenditures in fiscal 2026—a fourfold increase from 2022. But the point is that this spending isn't just about Azure. It's about building optionality across the entire AI stack.

Can Margins Really Recover in an AI-Heavy World?

Goldman believes the answer is yes, and that Microsoft's focus on efficiency is intensifying. Borges says supply constraints and the company's leading position in compute AI are driving a "maniacal focus on hardware, software, and human capital efficiency."

Microsoft believes Azure AI gross margins can eventually return to the levels seen in the traditional cloud business before the AI build-out. Goldman estimates margins could climb from the low 30% range today to the high 60% range over the next five to seven years.

Borges also notes that Microsoft bears "minimal LLM token costs beyond the compute needed to process them," thanks to its relationship with OpenAI and its ability to "pick and mix models based on price/performance." That flexibility matters.

Goldman views the AI strategy as a blend of vertical integration and strategic diversification. Microsoft holds a 27% stake in OpenAI and has also invested in Anthropic, giving it multiple paths to value creation.

What's the Upside Case?

In its bull scenario, Goldman estimates Microsoft could grow earnings per share to over $35 by fiscal 2030. That would represent more than 20% annual growth—well above the mid-teens EPS growth typical of peers with trillion-dollar-plus market caps.

Borges points to the breadth of investments, including AI and quantum computing, as factors that "set Microsoft apart from the broader software ecosystem." While many of these initiatives are still early, the company's "vision and capital to make strategic investments today" support confidence that it can continue capturing "the next waves of secular growth in technology."

After 50 years of navigating major technology shifts—from personal computing to the internet to the cloud—Goldman's message is straightforward. The next decade might not be a slowdown. It could be another chapter of compounding growth, powered by a company that's learned how to reinvent itself without losing momentum.