Who Will Win the AI Browser Race?

Microsoft released a new batch of features for its AI assistant Thursday, including an ambitious project that builds artificial intelligence directly into one of its most central products. More than a simple extension, the enhanced Copilot Mode of Microsoft’s Edge browser represents the company’s evolution into what CEO of Microsoft AI Mustafa Suleyman calls a “dynamic, intelligent companion” that can see tabs, summarize information, and take actions like booking hotels or filling out forms.

The announcement comes just two days after OpenAI showed off its new Atlas browser. While Copilot Mode in Edge first launched in July 2025 with features like a streamlined new tab page and multi-tab reasoning, Thursday’s expansion adds capabilities that bear striking resemblance to Atlas’s offerings. Neither company invented AI-assisted browsing, but the visual and functional similarity between the two products is hard to ignore.

The Competitive Landscape

The timing seems significant. OpenAI announced Atlas on Tuesday, October 21, 2025, launching it worldwide for Mac users with agent mode limited to paid Plus and Pro subscribers. Microsoft’s enhanced feature rollout came just 48 hours later, positioning the two tech giants in direct competition for control over how people navigate the internet.

The AI browser market has quickly become a battleground, with startups like Perplexity launching Comet and The Browser Company introducing Dia. Both Google and Microsoft have scrambled to retrofit AI capabilities into their legacy products, though Microsoft’s partnership with OpenAI adds complexity to the competitive dynamics.

What’s New in Microsoft’s Update

Starting October 23, Microsoft brought Copilot Actions and Copilot Journeys to Edge, both designed to make the browser more intelligent. Copilot Actions allows the AI to take control and interact with the web autonomously, while Copilot Journeys remembers browsing history and creates topic-based “journeys” that users can resume from anywhere.

Microsoft’s examples include having Copilot unsubscribe from email newsletters or make restaurant reservations. The system can analyze browsing habits to create personalized journeys that appear as boxes on the new tab page, providing one-click resume capabilities for related browsing sessions.

The enhanced capabilities also include natural voice navigation, allowing users to speak directly about tasks without extensive typing. Currently, Copilot Actions in Edge is available for free in a US-only limited preview.

Moreover, the company aims to evolve its AI browser to make it “more humanist”. Mustafa Suleyman in this post on X

“All of today’s @Copilot announcements boil down to one core idea: we’re betting on humanist AI. An AI that always puts humans first”.

Two Peas in a Pod

The functional similarities between Atlas and Edge’s enhanced Copilot Mode are substantial. Both feature AI assistants that understand multi-tab context. Both offer agent modes that can autonomously complete web-based tasks. Both track browsing history to deliver personalized responses. Both present simplified search interfaces with AI-generated responses by default.

The visual resemblance is equally notable, with the main differences being Edge’s darker background, text instead of a logo, and Windows-style window controls versus MacOS conventions. Part of this similarity stems from functional constraints as there are only so many ways to integrate a chatbot into a browser’s new tab screen. For users, the primary difference lies in the underlying AI models powering each experience.

The Microsoft Advantage

Despite the similarities, Microsoft brings significant advantages to the competition. Mustafa Suleyman serves as CEO of Microsoft AI after co-founding both DeepMind (acquired by Google) and Inflection AI. Under his leadership, Microsoft has positioned Copilot as the most personable AI companion, emphasizing personality and tone as key differentiators.

Edge also benefits from Microsoft’s existing infrastructure. The browser already reaches millions of Windows users by default, and Microsoft emphasizes that all features operate under the company’s standards of security, privacy, and performance.

Privacy Concerns and User Control

Both browsers raise similar privacy questions. Copilot Mode requires explicit permission to access browsing history and provides richer insights using past browsing data, if users choose to enable it. Microsoft’s blog post stresses that browser data is protected under the company’s privacy statement and that Copilot “only collects what’s needed to improve your experience,” with clear visual cues when Copilot is active.

OpenAI similarly states it won’t use Atlas browsing data to train models unless users opt in, and won’t use business users’ data at all. Both companies recognize that to function effectively, their AI browsers require access to sensitive information including login credentials, payment details, and personal data.

A Larger Strategy at Play

The rapid-fire browser launches reflect broader competitive tensions in the AI industry. Microsoft’s announcements also included introducing Mico, an expressive avatar that serves as Copilot’s visual presence, offering a warm and customizable face that “listens, reacts, and even changes colors to reflect your interactions”.

The browser push represents more than just feature parity. For Microsoft, integrating AI deeply into Edge strengthens its ecosystem and reduces dependence on distribution partnerships. For OpenAI, Atlas offers direct access to users rather than relying on platform owners who could restrict access at any time.

The Road Ahead

The high stakes of the AI race and the tense state of play between the two companies make it significant that both browsers launched in the same week. While browsers have largely looked the same for years, the underlying models and execution quality will ultimately determine which AI browser wins user adoption.

Whether these AI browsers can realistically challenge Chrome’s 3.45 billion users remains uncertain. The technology faces a “better must be significantly better” problem, as switching browsers requires overcoming powerful inertia from established habits. Early versions of web-browsing AI agents have shown promise on simple tasks but struggle with the complex, multi-step processes users most want to automate.

For now, the message seems quite clear. In the battle to control how humans interface with digital information, Microsoft and OpenAI are willing to launch remarkably similar products within days of each other. The competition has only just begun.