OpenAI Developing AI Browser to Challenge Chrome

OpenAI is gearing up to do more than power chats in other people’s browsers—it is about to ship its own. According to an exclusive Reuters report, the Sam Altman‑led company will release an AI‑native desktop browser “in the coming weeks,” a move that would plant a flag squarely in Google Chrome’s territory and draw a fresh battle‑line in the search‑and‑ads war.

The still‑unnamed browser is built on Chromium, Google’s open‑source engine, but layers a persistent ChatGPT‑style conversation pane over every page so users can ask follow‑up questions, distil long reads or trigger so‑called “agent” tasks—booking tables, filling forms, even completing a checkout—without leaving the tab.

Inside OpenAI the product is seen as the ideal launchpad for its forthcoming Operator agent, which needs deep hooks into web history to act autonomously. The gambit also hands the startup something Google has long guarded: granular, first‑party data on how billions of people roam the internet. Analysts say that data could sharpen OpenAI’s models and super‑charge its advertising ambitions.

Chrome commands about 68 % of the global browser market, while Apple’s Safari sits at 16 % and Microsoft Edge hovers near 5 %. OpenAI, by contrast, has roughly 500 million weekly ChatGPT users it can try to entice with a one‑click import of chat history and plug‑ins.

Even so, prying users away from default browsers is historically tough. Safari ships with every iPhone and Mac; Edge is welded to Windows; Chrome is pre‑installed on most Android phones and has a decade‑old extension ecosystem.

Rivals aren’t standing still. Perplexity just unveiled Comet, an AI browser that performs agentic tasks from a chat box, albeit behind a US $200/month paywall for power users. Microsoft’s Copilot Vision update in Edge lets the assistant “see” whatever is on screen and walk users through actions—a light version of the agent concept. Brave, Arc and Vivaldi have already woven AI summarisation and ad‑blocking into their Chromium forks.

What OpenAI might do better

• Model freshness: its o‑series models update weekly, so answers and summaries could outpace rivals that fork open‑weight LLMs.
• Plug‑in carry‑over: existing ChatGPT plug‑ins should work on day one, giving the browser a fledgling app store.
• Unified agent framework: early demos seen by partners show Operator chaining multiple web actions—compare this to Edge’s limited “highlight and explain” scope.

Where the road looks rough

• Regulatory glare: U.S. antitrust suits already circle Google’s ad stack; harvesting browser telemetry could invite similar scrutiny of OpenAI.
• Privacy optics: letting a for‑profit AI lab watch every click will test consumer trust, especially in Europe, where GDPR fines can reach 4 % of turnover.
• Monopoly tech reliance: building atop Chromium means any breaking change by Google could hamstring OpenAI features.

Still, history shows browsers can reshuffle market share quickly when they solve a new pain point—Chrome did it with speed, Safari with privacy, Edge with battery life. If OpenAI convinces mainstream users that “agentic browsing” saves real time, Chrome’s decade‑long reign could finally face a challenger built from its own DNA.

OpenAI declined comment, and Google did not immediately respond to requests for reaction. The launch window, leaked by insiders, suggests a debut before the company’s scheduled DevDay conference in August. Until then, Silicon Valley’s newest browser war is on the launch‑pad, engines humming.