SerpApi.Org
Microsoft has had trouble convincing investors that it has a winning artificial-intelligence strategy, but they might be looking at the company’s big AI initiatives all wrong, according to a Morgan Stanley analyst.
In assuming coverage of Microsoft with an overweight rating and a $600 price target on Tuesday, Morgan Stanley’s Adam Wood said that Azure and Copilot are “keys to the stock” and “both are set to inflect.” His price target is about 50% above current levels.
Shares of Microsoft have been down roughly 15% so far this year, though they are up 13% since the stock’s recent low on June 25.
Azure, which is Microsoft’s cloud business, and Copilot, which is its AI assistant, are becoming the “clearest indicators” that Microsoft’s AI investments are translating into “durable value creation,” Wood wrote.
With Azure, Microsoft is showing its ability to monetize infrastructure demand, and Wood believes the market is incorrectly valuing the cloud unit as a “bare-metal” or “commodity” AI infrastructure provider, whereby customers buy compute and inference capacity but not much else.
Wood thinks investors are missing that there is more “pull-through into higher-value Microsoft services” through the company’s business model, and believes Azure can drive adoption across its wider software portfolio.
As for Copilot, he wrote that investors are “increasingly underappreciating” the revenue potential that comes from Microsoft’s pricing model. Copilot-driven average revenue per user is becoming a “three-headed growth driver,” he said — calling out Microsoft’s opportunity to grow its revenue stream through Copilot seat adoption, push user migration towards Microsoft’s M365 E7 subscriptions, and leverage a new consumption-based monetization strategy.
Before, Microsoft’s monetization model was primarily tied to seat expansion and license upgrades. “With Copilot, Microsoft now has the opportunity to monetize both users and usage,” Wood wrote.
“As AI adoption broadens across organizations and customers increasingly leverage agents, reasoning capabilities and workflow automation, Copilot and the broader E7 opportunity have the potential to become one of the most significant ARPU expansion opportunities in Microsoft’s history,” he noted, referring to average revenue per user.
Read more: Meet Kimi K3, the newest Chinese AI model haunting Silicon Valley
D.A. Davidson analyst Gil Luria recently pointed out a related reason why he thinks Microsoft is set to win. He wrote in a note earlier this month that enterprise customers are realizing they may not wish to work directly with frontier AI labs like OpenAI and Anthropic, especially given Anthropic’s tumultuous track record with the U.S. government.
“The solution is becoming obvious — companies will want to build a way to make sure they can switch AI models without disrupting their business,” he said.
The answer to that may be an orchestration layer like Copilot, Luria believes.
“While investors may have thought of Copilot as an AI model in itself, it is already working as a tool that helps funnel queries and tasks to the appropriate model,” he wrote.
Luria explained that users can “choose models, choose whether to look at internal data and choose which agent to use, or allow Copilot to route the prompt in the most efficient and appropriate manner.”
“That is the definition of an orchestration layer,” he noted.
See also: Why cheap Chinese AI models could actually be a boon for Nvidia, Micron and other chip stocks
