Can Microsoft's AI Bet Pay Off? Investors Are Skeptical.

The $4 Trillion Question: Where's Microsoft's AI Edge?

Satya Nadella is talking a big game on artificial intelligence. But for traders watching MSFT's stock, which has lagged the Nasdaq and its megacap peers this year, the question is sharp: where are the results?

Microsoft’s early, aggressive bet on OpenAI was supposed to cement its AI leadership. Instead, the landscape has shifted. NVDA owns the silicon. Anthropic and OpenAI boast the frontier models. META’s AI agent is in billions of pockets. And Microsoft? It’s still searching for its breakout hit.

Nadella's New Playbook: Tokens, Not Just Subscriptions

Nadella knows the old rulebook is obsolete. The cloud-era mastermind, who transformed Microsoft by pushing subscriptions, is now engineering another fundamental shift—this time toward consumption-based pricing. The new currency? Tokens.

“We went from the code assist to the agentic coding to software factory all within one year,” Nadella told a group of customers recently. His goal: turn clients into “software factories” that consume massive amounts of AI compute, measured in tokens, via Microsoft’s Azure cloud and Copilot tools. Success, he argues, would “further expanding our TAM.” It’ s a bold pivot. But can he pull it off again?

The Copilot Conundrum

The centerpiece of this strategy is Copilot. Yet, after a year of broad availability, it’s not yet pervasive. With 30 million seats against a commercial Office 365 base of over 450 million, penetration is shallow. The fixed $30-per-user price tag is now seen as a limitation, not a virtue.

“I think there needs to be a little bit more maturity of the product before we figure out actually what the best solution or what pricing is,” said Tiffany Wade, co-portfolio manager at Microsoft shareholder Columbia Threadneedle. That sentiment echoes across Wall Street, where “show me” is the mantra.

Microsoft’s response is a new pricing model called Usage-Based Billing (UBB). The first product to use it is Copilot Cowork, designed for larger, complex tasks. The bet is simple: as AI tasks get more ambitious, token consumption—and Microsoft’s revenue—will soar.

“It may have the effect of what we saw when Claude Code, for example, shifted to consumption pricing, and you saw token bills go significantly higher,” said Cowen analyst Derrick Wood. The upside is clear, but so is the risk for budget-conscious IT departments. “There’s a very real concern that you just kind of turn it on and people are going to start draining the bank,” noted one industry advisor.

The Model Problem: Playing Catch-Up

While Nadella downplays the need for a single “frontier model,” the competitive reality is stark. The team under AI CEO Mustafa Suleyman has yet to produce a model that tops independent leaderboards. Microsoft finds itself in the awkward position of relying on and connecting to rival models from Anthropic and OpenAI even as those same companies encroach on its application turf.

The fraying of the OpenAI alliance has left Microsoft with a massive investment but less control. Suleyman’s hire was a hedge, but his responsibilities have already been refined, shifting from broader Copilot experience back to core model development. It’s a sign of the internal recalibration underway.

GitHub's Lesson: Pivot Fast or Lose Share

Microsoft has a recent, painful case study in moving too slowly: GitHub Copilot. As AI-native coding tools like Cursor and Claude Code surged, GitHub lost market share. Microsoft’s fix? A swift shift to consumption pricing. The result? User growth jumped to 50 million from 26 million in under a year, and sales growth accelerated 60% quarter-over-quarter.

“GitHub Copilot went out and learned lessons,” said the unit’s operating chief, Kyle Daigle. Those lessons are now being applied across Microsoft. The message is clear: in the AI race, pricing and packaging are as critical as the tech itself.

What This Means for the Market

For investors, the narrative is in flux. Nadella has a legendary track record of navigating shifts—from cloud to SaaS. But AI is a different beast, moving at a blistering pace and attracting colossal, well-funded competitors.

The stock’s underperformance this year tells the story: the market is waiting for proof that Microsoft can monetize AI beyond Azure compute rentals. The shift to UBB and token consumption is the right strategic answer, but it introduces volatility and uncertainty into future earnings. Will enterprise budgets keep pace with exploding token usage?

Nadella ended a recent talk by emphasizing that “leaders who are hands-on are able to change.” He’s changing the business model—again. The trillion-dollar question for traders is whether that change will be enough to close the gap with the pure-play AI leaders and re-energize a stock that’s been stuck in neutral.