Anthropic confirmed its first profit, a $67B energy mega-merger was driven entirely by AI power demand, and India unveiled AI in every school by 2027.
Anthropic's first ever profit: $559M on $10.9B in Q2 revenue, and a $30B raise closing
Anthropic told investors it expects revenue for the April–June quarter to hit at least $10.9 billion: more than double the $4.8 billion it made in Q1, and generate a $559 million operating profit. That would be the first time the company has ever turned a profit. At the same time, it is closing a $30 billion funding round at a valuation above $900 billion, backed by enterprise customers including Goldman Sachs, PwC, and Blackstone.
Why it matters: This is the moment AI companies move from burning cash to generating it. When the fastest-growing software business in history turns its first profit, it changes the conversation about whether AI is a bubble or a business.
NextEra buys Dominion in a $67B deal. The stated reason: AI is eating the power grid.
NextEra Energy announced it is acquiring Dominion Energy in an all-stock deal valued at $67 billion, creating the world's largest regulated utility. The merger documents state explicitly that the driver is surging AI data centre electricity demand. US electricity consumption is rising faster than at any point since World War II, and AI data centres are projected to consume 15–25% of US electricity by 2030. Dominion alone has already connected more than 450 data centres.
Why it matters: When the biggest utility merger in US history happens because of AI's appetite for power, energy and AI policy are no longer separate conversations. Electricity bills, climate targets, and infrastructure spending are all downstream of this.
OpenAI's Codex now works on your Mac while the screen is locked, you monitor it from your phone
OpenAI's Codex coding agent added a feature called Locked Use: it can operate Mac applications after your screen locks, and you can trigger and monitor those long-running tasks from your phone while you're away from your desk. It uses an Apple authorisation plug-in with tight safeguards, the machine relocks the instant keyboard or mouse input is detected. The feature is unavailable in the EU, UK, and Switzerland at launch.
Why it matters: This is one of the clearest signs yet of AI moving from "tool you use" to "agent that works while you're not watching." The line between assistant and autonomous worker just shifted.
India makes AI mandatory in all schools and colleges, starting the 2026-27 academic year
The Government of India confirmed that AI will be integrated into teaching and learning at all education levels from the 2026-27 academic year under the updated National Education Policy. The plan includes AI-aligned Content Creator Labs in 15,000 schools and 500 colleges. A Centre of Excellence in AI for Education is being established at IIT Madras with a ₹500 crore budget. GPU access for students and researchers expands to 58,000 units at ₹65 per hour, roughly one-third the global average cost.
Why it matters: This is systemic. If you are in school in India, or have children in school, AI is about to become a core subject, not a coding club. The gap between students who engage early and those who wait is about to widen.
OpenAI launches Rosalind Biodefense: AI built specifically for pandemic preparedness
OpenAI launched Rosalind Biodefense, a new product giving trusted developers access to AI capabilities designed for biosurveillance, outbreak detection, and early-warning systems. Named after DNA pioneer Rosalind Franklin, it targets pandemic preparedness applications that government labs have historically handled.
Why it matters: AI companies are moving into high-stakes scientific domains that were previously the exclusive territory of government research. Whether this is a public health breakthrough or a governance risk depends entirely on who gets access and how it is overseen.
Rhoda AI raises $450M and launches from stealth, a robot that learns by watching video
Rhoda AI publicly launched after 18 months in stealth with a $450 million Series A, unveiling its FutureVision platform, a robotic intelligence system that uses video-predictive control to teach machines how the physical world works. It targets industrial and logistics applications where text-based AI has never been enough.
Why it matters: Most AI progress has happened in the digital world. Rhoda is betting the next frontier is physical intelligence, machines that understand cause and effect in the real world. With $450M in backing, that bet just became very serious.
Who pays for AI's electricity bill?
The NextEra-Dominion merger is the starkest signal yet that AI is not just reshaping software, it is reshaping the physical infrastructure of entire countries. AI data centres are projected to consume 15–25% of US electricity by 2030, and that infrastructure requires billions in upgrades. The question now sitting on every policymaker's desk: does that cost land on electricity consumers, on the cloud companies building the data centres, or on governments? Nobody has answered it clearly. Watch for the first serious policy proposal to name a number.
Martha Stewart's AI home startup + OpenAI named Gartner Leader in coding agents + India's GPU milestone
Martha Stewart co-founded Hint, an AI home management startup that predicts and prevents household problems before they break, it raised $10 million from Slow Ventures. Gartner named OpenAI a Leader in its first-ever enterprise coding agents ranking, the closest thing the industry now has to an official performance standard. And India's subsidised GPU capacity officially crossed 58,000 units, one of the largest public AI compute pools in the world outside of the US and China.