OpenAI is building a super app, Google's Gemini is becoming your phone's OS, and the US government just demanded early access to AI models before public release.
OpenAI is building a "super app": ChatGPT, Codex, and a browser all in one
OpenAI is merging ChatGPT, its coding agent Codex, and its developer API into a single product team, with the goal of shipping one unified app that also integrates the Atlas browser. The idea: one place where you chat, code, search, and automate, without jumping between tools.
Why it matters: OpenAI is no longer just a chatbot company. This is a bet that AI becomes the interface itself (your browser, your assistant, your developer tool) all in one. If it lands, it changes how most of us use the internet daily.
Google Gemini is becoming your phone's operating layer, not just an app
Google is pushing Gemini deeper into Android so it can move across your apps, read your screen, pull from Gmail, build shopping carts, book reservations, and handle tasks that used to require jumping between five different apps. A smarter Chrome on Android and a redesigned Android Auto are also part of this rollout.
Why it matters: This is the clearest sign yet that Google sees Gemini as an OS-level layer, not a side feature. Your phone will start anticipating tasks, not just responding to them. If you use Android, this affects you whether or not you actively use AI.
Anthropic's Orbit turns your scattered inbox and calendar into a daily brief
Anthropic released Orbit, a feature that automatically generates a daily briefing by pulling from your Gmail, Slack, GitHub, Calendar, Google Drive, and Figma. It flags which emails matter, which pull requests need your attention, which meetings to prepare for, and which messages are piling up. No setup required beyond connecting accounts.
Why it matters: This is the move from "AI that answers questions" to "AI that manages your attention." For anyone drowning in notifications across work tools, this is a practical answer, and a preview of what Claude is becoming.
US regulators now get early access to AI models before the public does
The US government has reached agreements with major AI companies (including Microsoft and xAI) to provide regulators with early access to new AI models before public release. This is a significant shift: governments are no longer reacting after the fact but now have a formal preview window to assess risks.
Why it matters: For the first time, regulators sit at the table before a model ships. This could slow some releases, but it also builds the kind of trust that makes broader AI adoption possible. Other countries are watching closely.
Deepfake laws are moving from "being discussed" to "being enforced"
The US Senate unanimously passed a law giving victims of non-consensual, sexually explicit deepfakes the right to sue creators, distributors, and platforms, with a 48-hour removal requirement. Meanwhile, the EU's AI Act, which requires all AI-generated content to be clearly labelled, reaches full applicability in August 2026. A new threat has also emerged: AI-generated "synthetic job candidates" passing video interviews in real time.
Why it matters: Deepfake harm is no longer just a policy conversation, it now has legal teeth in both the US and EU. If you work in hiring, media, or content, these new obligations land on your desk, not just on tech companies.
Krutrim posts its first-ever profit: India's AI moment is arriving
Krutrim, India's first GenAI unicorn, reported its first annual net profit, with revenue tripling to roughly ₹315 crore ($31M) in FY2026, driven by enterprise clients adopting its Hindi-first AI models. This comes as the IndiaAI Mission continues expanding GPU access and the government's ₹10,000 crore AI fund is under discussion for a potential doubling.
Why it matters: A homegrown Indian AI company turning profitable (not just raising money) is a real signal. It shows there is commercial demand for AI built for Indian languages and contexts. The infrastructure and the market are aligning.
The race to be your AI operating system, and who wins it
OpenAI, Google, and Anthropic all made moves this week to become less of a tool you open and more of a layer that runs your digital life. OpenAI is building a super app with a browser. Google is embedding Gemini across Android at the OS level. Anthropic's Orbit wants to be your morning brief. Only one of these will feel indispensable to most people. Watch which one earns that position, because whoever does, everyone else becomes optional.
Oracle's $50B AI data centre push + Mistral's 128B flagship model
Oracle announced plans to raise up to $50 billion to expand its global AI data centre network, sending its shares lower as investors weighed the scale of the bet. Separately, Mistral launched its new 128B flagship model with async cloud coding sessions and a dedicated "Work" agentic mode in Le Chat, the most capable open-weight model Mistral has shipped to date.