Microsoft Build 2026: The 7 Biggest Announcements
Satya Nadella took the stage at Build 2026 and, true to form, the announcements came fast and came heavy — new Surface hardware, in-house AI models, an always-on assistant, and more. Microsoft is clearly done being OpenAI’s well-dressed distribution channel and wants to be taken seriously as an AI company in its own right. Whether that ambition matches the execution is the question the next 12 months will answer.
Microsoft’s First Advanced Reasoning AI Is Here
MAI-Thinking-1 is Microsoft’s first flagship reasoning model, and it marks a genuine inflection point: the company that spent years as OpenAI’s biggest customer is now competing with its landlord. The timing isn’t subtle — this dropped right after Microsoft and OpenAI renegotiated their partnership to loosen the ties. Nothing says “we value this relationship” like building your own replacement.
Microsoft Wants to ‘Make People Addicted’ to Its New AI Assistant, Internal Documents Reveal
Planning documents for Microsoft Scout explicitly use the word “addicted” — as in, the goal is to hook users before layering on new features. To be fair to Microsoft, every tech company engineers for engagement; it’s just rare that someone writes it down in those exact words. Points for honesty, I guess, minus several points for strategy.
Trump Signs Executive Order to Review AI Models Before They’re Released
The White House signed an executive order creating a “voluntary framework” for AI companies to share frontier models with the federal government before public release. “Voluntary” is doing a lot of work in that sentence — it’s hard to imagine the administration being relaxed if a major lab opts out. This lands as the administration is, per Wired’s reporting, actively at war with itself over how much to regulate AI at all, which gives the whole thing a certain we’ll figure it out later energy.
Google’s Phone App Will Tell You If a Scammer Is Impersonating One of Your Contacts
Google is rolling out AI-powered detection that flags calls from scammers pretending to use a number already in your contacts. This is a genuinely useful, unglamorous application of AI — no flashy demo, just a feature that might stop your grandmother from wiring money to someone pretending to be you. More of this, please.
Nvidia and Microsoft Researchers Say AI Agents Don’t Care About Safety or Reliability
Researchers from Nvidia and Microsoft published findings that AI agents are largely indifferent to safety and reliability constraints — they compared the behavior to Mr. Magoo stumbling through hazardous situations without realizing it. The irony of this warning dropping on the same day Microsoft announced a fleet of new AI agents at Build is either perfect timing or a cry for help.
Google Is Quietly Buying Code From Play Store Developers to Train AI
Google is running a confidential program to purchase code directly from Android developers — presumably to feed into AI training. The word “quietly” in the headline is carrying some weight here; if you have to keep a data acquisition program confidential, it probably says something about how you expect it to be received.
Uber Caps Employee AI Spending After Blowing Through Budget in 4 Months
Uber encouraged its employees to use AI tools liberally, and they did — so liberally that the company burned through its annual AI budget by April. Now the same company that told staff to go wild is rationing access. The lesson here isn’t that AI is too expensive; it’s that “use it as much as you want” is not a budget strategy.
Bottom Line
The AI industry is simultaneously promising to make you more productive, trying to get you addicted, and quietly admitting the agents doing all this work don’t particularly care about safety — and we’re all just nodding along.