Goldman Sachs Bullish Outlook on Microsoft Stock and AI Growth

The financial world is currently grappling with a stark divergence between corporate performance and market sentiment. By any traditional metric, Microsoft is a juggernaut in its prime, recently posting quarterly revenue of $81.3 billion a 17% year-over-year increase. Yet, the equity markets have responded with a cold shoulder. Over the last six months, the stock has shed up to 25% of its value, leaving many short-term traders “sketchy” about the company’s massive overhead in the artificial intelligence arms race.

While the market panics over quarterly capital expenditure (Capex), Goldman Sachs is doubling down. The investment bank recently reiterated its “Buy” rating with a $600 price target implying a massive 55–57% upside. Are investors missing the forest for the trees? To understand why Wall Street’s most rigorous analysts remain bullish, we have to look past the quarterly jitters and into the “secular growth story” of an infrastructure meant to last centuries.

The Capex Paradox: Why Spending is the New Winning

The primary driver of recent market anxiety is Microsoft’s aggressive capital expenditure. In its most recent quarter, the company spent $37.5 billion, putting it on a trajectory to reach a staggering $148 billion for the 2026 fiscal year a fourfold increase in just four years.

While critics see this as a high-risk cost, Goldman Sachs views it through the lens of a “Capex Paradox.” This is not merely a cost of doing business; it is a fundamental shift in internal R&D aimed at capturing structurally better unit economics. Crucially, not all capex translates directly to Azure revenue growth. Instead, Microsoft is prioritizing its first-party (1P) applications like Copilot which offer higher profit potential than simply reselling infrastructure. To support this, Goldman expects the proportion of Microsoft’s computing resources dedicated to internal R&D to rise from 10% to 20%.

“Capital expenditure on internal R&D is a long-term investment, some of which is not directly monetized at present but may be monetized in the future and remains crucial to Microsoft’s overall development strategy.”

The “Maia” Factor: Breaking the Silicon Monopoly

One of Microsoft’s most potent strategic weapons is the transition to custom silicon. The Maia 200 chip an updated custom AI inference accelerator is designed to decouple Microsoft’s margins from the expensive third-party hardware market.

By diversifying its silicon footprint, Microsoft is pursuing several strategic advantages:

• Cost Leadership: Maia is reportedly 30% cheaper than other AI silicon currently on the market.
• Margin Recovery: Custom ASICs allow Microsoft to target AI gross margins that approach the 60% levels seen in legacy CPU-based Azure workloads.
• Operational Leverage: Internal hardware allows for tighter integration with software, driving higher efficiency.

However, a seasoned analyst must note the limitations. While Maia is a breakthrough, Microsoft still faces a steep climb in deepening its software ecosystem to compete with Nvidia’s entrenched “CUDA moat.” For the near term, Nvidia maintains leadership in the accelerator market, but Microsoft is playing the long game by building a footprint that limits its downside to any single vendor.

Project Silica and the 10,000-Year Memory

Beyond immediate AI gains, Microsoft is pioneering a breakthrough in data storage that redefines the concept of “long-term.” Project Silica, recently detailed in a paper published in Nature, involves encoding data in ordinary borosilicate glass the same material used in kitchen cookware and oven doors.

This technology is a quantum leap over traditional media:

• Extreme Durability: Resistant to water, heat, and dust, it can preserve data for 10,000 years without the degradation seen in hard drives or tape.
• Commercial Viability: By moving from expensive fused silica to widely available borosilicate glass, Microsoft has addressed the primary hurdles to the commercialization of permanent storage.
• Density: The new technique stores hundreds of layers of data in glass only 2mm thick.

This isn’t just a science project; it’s an impactful move toward creating a permanent, low-cost archive for the world’s most critical data, turning a “supply-constrained” storage market into a stable, century-scale utility.

The OpenAI Double-Edged Sword

The deepest point of contention for investors is the “concentration risk” revealed in recent disclosures. Approximately 45% of Microsoft’s $625 billion commercial backlog (remaining performance obligations) is now tied directly to OpenAI.

This is a double-edged sword. On one hand, it grants Microsoft a first position in compute AI and a 27% stake in the industry leader. On the other, the high OpenAI reliance creates a dependency that some fear could leave Microsoft vulnerable. However, the strategist’s view is that this relationship allows Microsoft to “pick and mix” models to maximize price-to-performance, ensuring they maintain the premier LLM position without being locked into a single technological path.

The Margin Recovery Narrative

The crux of the $600 price target lies in the rapid recovery of Azure AI margins. In fiscal year 2024, these margins were a dismal -50% as the company built out its infrastructure. By fiscal year 2025, they had leaped to 17% all while operating in a supply-constrained environment.

Goldman Sachs points to a “maniacal focus” on efficiency across hardware, software, and human capital. This operating leverage is expected to drive earnings per share (EPS) to potentially double by the 2028 fiscal year. As Microsoft moves from the infrastructure-build phase to a compounding AI product cycle, the ability to scale profits without a linear increase in costs becomes the dominant narrative.

Conclusion: The Next Fifty Years

Microsoft is currently transitioning from building the foundation to harvesting the crop. While the market remains jittery about the billions of dollars leaving the balance sheet today, the infrastructure being built is designed to serve a generation of compounding AI cycles.

As the company enters its second half-century, it is clear they are not managing for the next quarter, but for the next century. In an era defined by quarterly jitters and 24-hour news cycles, the question for investors remains: In an era of quarterly jitters, are we prepared to value a company that is building infrastructure meant to last 10,000 years?