Microsoft’s aggressive, multi-billion-dollar push into synthetic intelligence was speculated to be a flawless victory. The mixing of Copilot into Home windows 11, Microsoft 365, and GitHub was designed to usher in a new period of agentic computing. But, beneath the polished keynote displays and large infrastructure investments, a dramatically completely different actuality is what we noticed.
As first reported by Home windows Newest, in keeping with a extremely revered former Microsoft government, the corporate’s AI technique is basically failing to attach with actual customers, spurring requires a large inner “manufacturing facility reset.”
The manager in query is Mat Velloso, who was most lately the Vice President of Product for the Developer Platform at Meta’s Superintelligence Labs. He additionally led AI developer merchandise at Google DeepMind (together with the Gemini API and Google AI Studio). However earlier than his stints at Google and Meta, Velloso spent over 12 years at Microsoft, the place he served as a Companion Director managing AI innovation in Home windows and, apparently, spent 4 years because the Technical Advisor to Microsoft CEO Satya Nadella.
When somebody with Velloso’s resume, having noticed the AI arms race from the very best ranges of Microsoft, Google, and Meta, says Microsoft has “missed the AI wave,” it may well rip the lid off the deep tensions inside Redmond.
Ex-Microsoft exec says Microsoft missed the AI wave
Microsoft’s conduct over the previous couple of months is nothing wanting surprising. Each the Home windows and Xbox divisions out of the blue began prioritizing person suggestions and implementing requested options after years of ignoring them.
It’s additionally not a small process to assemble and set up OEMs, ODMs, and chipset distributors in an occasion like WinHEC that had its final prevalence nearly a decade in the past (2018).
Explaining this sudden pivot to listening to prospects, Velloso remarked that regardless of making Bing the corporate’s greatest AI guess, it did not seize a single share level of search market share from Google. Extra damning is the state of Copilot.
In keeping with Velloso, lower than 3% of paying customers actively use Copilot, though Microsoft has pre-deployed it instantly into the Home windows 11 taskbar and throughout the Workplace suite.
Out of Microsoft’s 450 million Microsoft 365 person base, the corporate has solely managed to transform roughly 15 million paid Copilot seats. This implies a staggering 96.7% of customers are rejecting the premium AI options, yielding only a 3.3% paid adoption fee. When seen towards Microsoft’s estimated $37.5 billion quarterly AI spending, that is an alarmingly low adoption fee.
But it surely’s not simply software program; Velloso additionally referred to as out the present state of AI {hardware}. Over the previous yr, Microsoft has closely pushed OEMs to incorporate Neural Processing Models (NPUs) of their newest laptops to energy superior Home windows 11 capabilities. We now have tracked Microsoft’s push for NPU-powered AI options in Home windows 11, however as Velloso famous, OEMs invested closely in NPUs solely to search out out that “no one cares as a result of not a single invaluable usecase was constructed for these in Home windows/Workplace.”
Moreover, he highlighted that GitHub, a platform that must be thriving because the centerpiece of the AI coding revolution, has seen its Service Degree Settlement (SLA) reliability drop beneath 90%. Mix this with rising Price of Items Bought (COGS) and shareholders starting to ask troublesome questions, Microsoft was pressured to start out listening to prospects as a result of the AI guess isn’t paying off as easily as anticipated.
Microsoft’s government exodus
All this friction seems to be taking a toll on Microsoft’s management. Lately, information broke that Julia Liuson, the extremely revered head of Microsoft’s Developer Division (DevDiv), was retiring after 34 years with the software program large.
Whereas official channels framed this as a regular retirement, Mat Velloso critiqued the information, saying, “Appears like Microsoft simply went from hit refresh to hit manufacturing facility reset.” He additionally listed a large string of high-profile departures and reassignments throughout the corporate, together with leaders from Xbox, GitHub, AI Infrastructure, Groups, and OneNote.
This public commentary drew the ire of Frank X. Shaw, Microsoft’s Lead Communications government. Shaw replied to Velloso, defending the departing executives and accusing Velloso of jamming a “destructive body” onto regular company retirements.
All Velloso needed to do was level out the tough monetary realities that the market is at the moment digesting. He moved from Microsoft to Google in early 2024, and whereas Google’s shares surged by roughly 230%, Microsoft’s inventory development remained primarily flat at 0%.
“I suppose the entire market can also be improper about this,” Velloso fired again. “However what do I do know? I solely noticed each corporations, how they function, their strengths and their flaws. Microsoft missed the web wave, the cellular wave and now it missed the AI wave. It’s what occurs if you hold doing the identical issues anticipating completely different outcomes.”
He urged Microsoft’s management to cease “gaslighting,” deflecting blame, and burying skeletons, arguing that hitting a whole inner “manufacturing facility reset” is the one pragmatic method to repair the arduous issues plaguing the working system and enterprise software program stack.
OpenAI is slicing out the intermediary (Microsoft)
Aside from the interior challenges, Microsoft is more and more being affected by its closest allies. The corporate has staked its complete generative AI future on its multi-billion-dollar partnership with OpenAI. Nonetheless, OpenAI is quickly constructing out its personal enterprise infrastructure, threatening Microsoft’s historic dominance within the company sector.
Simply days in the past, OpenAI formally launched the “OpenAI Deployment Firm” (DeployCo), a brand new enterprise unit backed by over $4 billion in preliminary funding from international corporations. This new enterprise options 150 “Ahead Deployed Engineers” (FDEs) tasked with embedding instantly into Fortune 500 corporations to assist them construct and deploy customized AI options.
Traditionally, this hands-on, enterprise-level consulting was Microsoft’s bread and butter. As Velloso reminisced about his early profession as a advisor, he famous that Microsoft’s unbelievable penetration into giant enterprises was constructed on “armies of individuals spending time, listening, understanding enterprise targets and fixing them with expertise in each trade vertical.”
Now, high AI labs like OpenAI are replicating this similar playbook. They now not need to delegate enterprise deployment to hyperscalers like Microsoft Azure. By slicing out the intermediary and proudly owning the direct relationship with companies, OpenAI is stepping instantly into the extremely profitable companies layer the place the actual enterprise AI {dollars} are headed. This presents an enormous structural menace to Microsoft’s long-term enterprise income, particularly as customers proceed to push again towards pressured Copilot internet app integrations.
Why Microsoft is way from useless: “The moat is unbreakable”
Regardless of these extreme criticisms, Velloso defended his former employer towards apocalyptic tech media narratives. When a distinguished tech publication lately claimed that “AI is killing Microsoft” and in contrast their present trajectory to the disastrous 2008 period, Velloso stepped in to close the narrative down.
“Nope, they aren’t dying,” Velloso acknowledged. “I do know I criticize them loads and that’s as a result of I care, however boy when you suppose Microsoft is dying you haven’t watched what number of occasions they recovered from issues.”
He identified that whereas AI startups and labs is likely to be constructing flashy deployment corporations, fully changing legacy enterprise software program is extremely troublesome. When requested about corporations claiming they will absolutely automate software program companies, Velloso really helpful speaking to Fortune 500 CIOs to see how reasonable that actually is.
“There’s a cause why all the highest AI labs are hiring giant consulting groups,” he defined. “The final mile is the toughest and Microsoft has the perfect distribution for that. Their moat is unbreakable.”
That is the final paradox of Microsoft in 2026. They’re concurrently struggling to persuade customers to undertake their new AI options, but their foundational grip on the enterprise ecosystem remains to be strong. Hundreds of thousands of companies rely on Energetic Listing, Azure, Workplace, and Home windows. Whereas it’s simple to spin up an AI agent in Python, integrating it securely right into a legacy company atmosphere requires the type of large distribution community that Microsoft has spent 4 a long time constructing.
A mandatory wake-up name
All of the “new” options in Home windows 11, together with the much-awaited movable taskbar and resizable Begin menu, are a results of all the things that went improper within the firm and management lastly doing one thing about it.
When an organization realizes that pushing a 3.3% adoption-rate AI chatbot down customers’ throats isn’t translating to inventory development or person satisfaction, they’re compelled to rethink their technique.
We’re lastly seeing the outcomes of this “manufacturing facility reset” in actual time. Microsoft is dismantling sluggish internet wrappers, committing to native UI efficiency for Home windows 11 apps, overhauling driver high quality requirements, and explicitly prioritizing long-ignored person suggestions.
Microsoft’s pivot again to platform fundamentals proves that whereas AI is likely to be the long run, you can’t construct that future on a crumbling basis. The market pressured Microsoft to pay attention, and for annoyed Home windows customers, that’s arguably the perfect factor that would have probably occurred.
