NeuralStackMicrosoft used its Build 2026 developer conference in San Francisco to signal a sharper push into in-house AI, unveiling five major announcements that span prototype devices, desktop hardware, assistants, models, and healthcare. The company’s message was clear: it wants to compete more directly with OpenAI and Anthropic, not just host their systems.
One of the most notable announcements was Project Solara, a new family of AI-focused prototype devices. Microsoft said these devices are designed around AI agents rather than traditional app-centric workflows, with cloud-connected systems handling tasks on the user’s behalf.
The company said the prototypes could take several forms, including smart speaker-sized gadgets and badge-like wearables. That suggests Microsoft is still exploring what an AI-first device should look like, while trying to define a hardware category around agentic computing.
Microsoft also introduced the Surface RTX Spark Dev Box, a new desktop powered by Nvidia’s RTX Spark chip. Satya Nadella described it as a “dream machine,” and the company said it had demonstrated the device running a very large AI model locally, underscoring its interest in bringing advanced AI workloads closer to the PC.
On the software side, Microsoft introduced Scout, a new AI agent that will become part of Copilot. The agent is designed to gather emails, messages, and other items that need attention so users can act faster, placing it in the same broad category of tools that are increasingly trying to do work rather than simply answer questions.
Microsoft also highlighted MAI Thinking-1, its first reading-focused model, which it says can perform at a level comparable to some of the most advanced reasoning systems. Alongside that model, the company said it is also rolling out new transcription and image-generation models as part of its broader AI lineup.
Why this matters
Microsoft is no longer presenting itself only as a platform for other companies’ AI. By building its own models, agents, and specialized hardware, it is trying to control more of the stack and reduce dependence on external labs while offering its own path for customers.
The healthcare partnership with Mayo Clinic shows that this strategy is not limited to productivity software or developer tools. Microsoft wants to position its AI infrastructure and reasoning systems inside high-stakes workflows where trust, speed, and integration matter, which could make it more competitive in enterprise markets.
The broader takeaway is that the AI race is becoming less about isolated model releases and more about ecosystem control. Microsoft’s Build announcements point to a company trying to make its own tools central to how AI is deployed across devices, workplaces, and specialized industries.
For OpenAI and Anthropic, that means facing a rival with deep distribution, cloud reach, and now a growing set of in-house alternatives. Microsoft’s latest moves suggest the competition is moving beyond partnership into direct product confrontation.