Microsoft's Most Bizarre AI Pivot Yet

Microsoft's Copilot movement shows signs of not keeping up. 2025 are sad times for the Behemoth.

While Microsoft announced a $80 billion investment in artificial intelligence (AI) for fiscal year 2025 of capex, some of its business decisions are highly questionable now and it has lost the race for ML Researcher and product talent. The investment will be used to build AI-enabled data centers, train AI models, and deploy AI and cloud-based applications.

But wait for it, that’s not all.

In classic Satya Nadella fashion, the failed Copilot mission keeps getting a new spin. Microsoft’s developer division is now all about AI, as it reorgs teams to be ready for a platform shift.

It’s beyond cringe how “AI” things are getting and the how high the stakes are for CEO bonuses like Satya’s pay day. Dubbed "CoreAI - Platform and Tools," the new division rolls the existing AI platform team and the previous developer division (responsible for everything from .NET to Visual Studio) along with some other teams into one big group.

After acquiring Inflection (fairly shady means) and its team it would appear Microsoft’s consumer AI approach failed significantly in 2024 to move the needle. Not a big surprise considering how they operate.

Microsoft is no longer competitive in the high stakes talent wars between companies like OpenAI, Anthropic and Google. Jay Parikh, previously VP and global head of engineering at Meta, will lead the new division. He’ll report to Microsoft CEO Satya Nadella and oversee groups, including the company’s AI platform and developer teams.

Parikh worked on technical infrastructure and data center projects at Meta. Before joining Microsoft in October, he was appointed the CEO of cloud security startup Lacework. This is all very confusing to me.

Microsoft is pretending it will be an agentic AI leader in another wishful thinking move. Microsoft is building “the end-to-end AI stack – tools, platforms, and infrastructure to enable innovation and deliver real-world impact across industries.” Together, the company claims, to be helping customers and partners move beyond first-generation AI applications into a world of dynamic, agentic applications – applications designed to act on their behalf with memory, adaptability, and intelligence at their core. CoreAI “must lead” in defining and building what software application development for humanity will look like, according to a recent LinkedIn post.

With Meta’s CEO saying even some mid level SWEs are replaceable the engineering community has reason to be worried about o3 level models. Microsoft who have insider data to OpenAI’s top models must be thinking along similar lines. Just like OpenAI is pushing journalism to more AI via a shady Axios deal, Microsoft has always been pushing its devs in the Github Copilot direction. But is it even good business strategy?

Microsoft is neither here nor there in a lot of its moves. Trying to convince the world AutoGen is the future AI agentic framework is not helping in the credibility department.

The new org, called CoreAI — Platform and Tools, is actually a combination of Microsoft’s existing Dev Div and AI platform teams

Microsoft has the capability to invest $80 Billion into AI datacenters and connecting these sites yet how many layoffs do you think it will undertake in 2025? It could be a considerable amount.

Monopoly Capitalism is full-on shareholder value and stock market bubble moves, and less on actually innovating as of the mid 2020s. Microsoft’s equity in OpenAI is nice to have but reliance on their models has made them weak.

Microsoft has been undergoing a significant transformation as it pivots towards artificial intelligence (AI) in software development. But whatever happens now to the company is purely self-inflicted. Betting on OpenAI has been hugely expensive and has given rise to a formidable competitor where AGI is just a number of revenue generation ($100 Billion apparently!).

It’s just all so mind blowing and dumb.

A world where Sam Altman has grifted Satya Nadella and his greed. This is all backfiring in stupendous ways for Microsoft. Where OpenAI somehow has the money to build humanoid robots internally even as it will record by far the biggest cash deficit in 2025 of its short history.

Was Jay Parikh the best Microsoft could find for this idea? A real head scratcher. So many executives at Microsoft in AI, the former Google one keeps promoting his book on LinkedIn, sounds about right.

Microsoft has been overhauling its entire product lineup around artificial intelligence and needs to manage complex infrastructure projects that can help it wring greater power and efficiency from its networks. But why the constant pivots? Because what it is doing isn’t working. The Datacenter expansions is good for the future of Azure, but the in-house AI business looks very distressed and customers are not happy - a bit like consumers and Apple’s lame Apple Intelligence efforts.

CEO Satya Nadella emphasized that this shift is crucial for Microsoft to remain competitive in the rapidly evolving tech landscape, particularly in the realm of generative AI.

But in truth, Microsoft is not competitive, even with access to OpenAI’s best models. This means it’s failed in Product, execution, growth and sales and on so many different levels, it’s too late to fix. Somehow the DOJ let Microsoft acquire Github, LinkedIn and so many businesses it should never have been allowed to acquire - like in gaming too. But it’s gotten too big to even execute.

Microsoft's pivot to generative AI is fueled by its rich history in IT and software development. But their own customers hate them with a passion for a reason. Customers feel locked in to their poor Copilots and astoundingly poorly implemented AI solutions.

Satya Nadella keeps talking about the same things somehow immune to the reality: with various mantras of an “AI-first app stack” inside Microsoft. In the meantime Anthropic, Google and others are building legit Enterprise AI products.

Jay is just another Facebook leader. Parikh joined Facebook in 2009 and spent more than a decade there, working on technical infrastructure and data center projects that helped the company grow into the world’s largest social network. He left the company now known as Meta Platforms Inc. in 2021 and became CEO of cloud security startup Lacework Inc.

This new division will bring together Dev Div, AI Platform, and some key teams from the Office of the CTO (AI Supercomputer, AI Agentic Runtimes, and Engineering Thrive), with the mission to build the end-to-end Copilot & AI stack for both our first-party and third-party customers to build and run AI apps and agents. This group will also build out GitHub Copilot, thus having a tight feedback loop between the leading AI-first product and the AI platform to motivate the stack and its roadmap.

Github Copilot is a good tool, but others in the lucrative space are already catching up and are more useful to most SWEs using AI like Cursor with Claude and other competitors now.

Parikh will lead this new group as the executive vice president of CoreAI - Platform and Tools, after previously being instrumental to Meta’s engineering efforts for more than a decade. But where does Microsoft expect all these new re-orgs to lead honestly?

The fact that the blog post doesn't say anything about .NET or Visual Studio, instead emphasizing GitHub Copilot and anything and everything related to agentic AI shows Nadella is chasing hype, a strategy which was a loser in 2023 and 2024 for Bing AI and Copilots. Satya made a lot of good decisions around Azure and the Cloud but is clearly out of his depth when it comes to AI, as Microsoft Research has also lost some momentum, credibility and talent since 2022.

In the Generative AI era, Microsoft has lost its identity.

What is Microsoft CoreAI?

Microsoft CoreAI is a newly established division within Microsoft, officially launched in January 2025. This initiative is part of a broader strategy to enhance the company's capabilities in artificial intelligence (AI) and streamline its development processes. The CoreAI division is specifically focused on integrating AI technologies into Microsoft's existing software and tools, thereby creating a more cohesive and efficient development environment for AI applications.

This sounds a lot like automating the work of employees with Generative AI and new o3-like models. Pretty much what you would expect from a monopoly like Microsoft.

So-called AI agents are applications that are given specified boundaries (action spaces) and a large memory capacity to independently do subsets of the kinds of work that human office workers do today. But AutoGen isn’t even that good and Microsoft has little realistic chance to make a dent with AI Agents, either. We know already how all of this turns out by history. History being the biggest predictor of the future behavior of corporations too!

I rest my case.

Or maybe not, a number of poor decisions have led Microsoft here. Supporting OpenAI into its own monopoly status of Generative AI has been a big mistake it’s increasingly looking like.

CoreAI is designed to improve the synergy between Microsoft's software engineering and AI divisions, allowing for a more integrated approach to AI application development.

But are more VPs of AI the answer?

Microsoft is no Nvidia and agile group-think, Microsoft is among the worst bureaucracies in BigTech.

“In this world, Azure must become the infrastructure for AI, while we build our AI platform and developer tools — spanning Azure AI Foundry, GitHub, and VS Code — on top of it,” says Nadella.

I don’t know what Microsoft must become, but I know what’s likely to become as it relates to AI. It’s already history.