Polar, a new browser startup built around AI and aimed at knowledge workers, launched on July 29, 2026 with a $5.7 million seed round led by Madrona, according to TechCrunch.

The company was founded by CEO Kevin Jiang, who previously worked on Perplexity's Comet browser, an AI-focused browser that treated web pages not just as things to display but as material to search and summarize. Polar is available now at polarbrowser.com. It is free to use with a daily allowance of AI task credits, and paid subscription plans starting at $20 per month unlock higher usage limits.

Madrona partner Sabrina Albert, who is backing the company, framed the opportunity in straightforward terms: automating tasks for knowledge workers is a large market. The seed round drew a roster of angel participants that spans enterprise software, semiconductors, and AI research: Thomas Dohmke (ex-GitHub CEO), Phoebe Gates (Phia founder), Erik Bernhardsson (Modal CEO), Benjamin Spector (Flapping Airplanes co-founder), Rob Wachen (Etched co-founder), and Robert Yang (Fundamental Research Labs co-founder and CEO).

Polar enters a browser market that is becoming one of the more contested spaces in consumer and enterprise software. Arc, from The Browser Company, showed that a venture-funded browser can attract a devoted user base even without building its own underlying rendering engine. Perplexity's Comet, where Jiang worked on this category, pushed the idea of a browser as a tool that retrieves and synthesizes information rather than passively displaying web pages. Polar's focus on knowledge workers specifically, rather than general consumers, narrows the target audience but increases the potential value each user gets from the product.

The investor list is worth noting for what it signals about conviction rather than capital. Bernhardsson's participation connects Polar to the AI infrastructure layer through Modal, a platform for running compute-heavy workloads. Wachen's ties to Etched, which is building custom chips designed specifically for AI inference, place a bet that AI-native browsers will be compute-intensive enough to benefit from hardware-level optimization. Dohmke's involvement brings credibility on the developer-tools side, where knowledge work and code workflows overlap. These are not passive angel checks; they are operators with direct stakes in adjacent layers of the AI stack.

The freemium model, with daily AI task credits and a $20/month paid tier, mirrors pricing patterns that have become standard across AI-native consumer products. The assumption underneath is that AI task execution, not rendering or tab management, is the core unit of value, and that heavy users will hit credit limits fast enough to upgrade to a paid plan. Whether that assumption holds for knowledge workers, whose workflows are often varied and unpredictable, remains an open question that only usage data will resolve.

What is not yet clear from the available information is Polar's underlying engine, its approach to managing context across tabs and sessions, or how it handles the privacy and data-retention questions that any AI browser must address when processing work-related content. These are the details that will determine whether Polar differentiates on substance or on interface polish. The browser category has a long history of compelling demos that failed to survive sustained daily use.

In my view, the narrower focus on knowledge workers is a sounder bet than chasing the general browser market, where Chrome and other incumbents with default-installation advantages are nearly impossible to displace. Knowledge workers already pay for tools that reduce friction in their workflows. If Polar can make browser-based research, summarization, and task automation meaningfully faster, the $20/month price point is well within the range that professionals expense without hesitation. The risk is not the price; it is whether the AI execution layer is reliable enough that users trust it with real work rather than novelty queries.

The broader context here is whether the browser is the right foundation for AI-mediated knowledge work at all. Competing approaches range from AI agents that operate across the entire desktop to copilots embedded inside code editors and other applications where work actually happens. Polar's bet is that the browser, as the universal interface to web-based information, remains the natural place to coordinate AI-assisted tasks. That bet has historical precedent on its side, but the competitive landscape is denser than it was even a year ago.

For now, Polar has capital, a credible founding team with direct category experience, and a focused go-to-market. The seed round buys time to prove that daily AI task credits translate into retained, paying knowledge workers.