EN
Back to the archive

The archive · AI & Models · Product decision · 2023–2026

Genspark's search-to-super-agent bet: ex-Baidu execs to a $1.25B unicorn

Ex-Baidu's Xiaodu CEO built an AI search engine, pivoted to autonomous 'Super Agents' that finish work, and reached a $1.25B valuation in 18 months.

Genspark

The betThat users want finished work, not links: a multi-agent platform that autonomously delivers reports, apps and decks would outgrow AI search engines and chatbots.Scaling

What the business is

Palo Alto-based AI company that runs Genspark, an AI search engine with AI-generated Sparkpages, and Genspark AI Workspace, a Mixture-of-Agents platform that orchestrates 30+ models and 150+ in-house tools to autonomously complete business tasks.

Starting capital$60M seed at $260M post (June 2024); $100M Series A (March 2025); $275M Series B at $1.25B post-money (Nov 2025) — roughly $435M raised in 18 months

How it started

Eric Jing, ex-CEO of Baidu's Xiaodu and a former Microsoft Bing development manager, and Kay Zhu, ex-Google and ex-Baidu search architect, left Baidu in 2023 and founded Genspark (Mainfunc Inc). They launched the beta in June 2024 with a $60M seed from Lanchi Ventures, betting search would move from keyword links to AI-generated answers.

What happened

In March 2025 Genspark raised a $100M Series A at roughly a $540M valuation. In April it launched Super Agent, a multi-agent system that beat Manus and OpenAI on the GAIA benchmark; annualized revenue hit $10M in nine days, $36M in 45 days and over $50M within five months, with 2M+ monthly active users. On November 20, 2025, Emergence Capital led a $275M Series B at a $1.25B post-money valuation, alongside the launch of Genspark AI Workspace and entry into Microsoft's Agent 365 ecosystem.

How it ended up

Still scaling: as of November 2025 Genspark is a $1.25B unicorn with more than $50M annualized revenue, an enterprise-focused AI Workspace, and integration into Microsoft 365 via Agent 365.

Background

Genspark was founded in 2023 by Eric Jing, the former CEO of Baidu's Xiaodu and a Microsoft Bing veteran, and Kay Zhu, an ex-Google search architect who had worked with Jing for 11 years. Their bet was that search was moving from lists of links to AI-generated answers, and they launched the product in June 2024 with a $60M seed from Lanchi Ventures at a $260M valuation.

The product grew through Sparkpages — AI-generated single-page answers to queries — but Jing and Zhu saw a bigger opportunity than search: users didn't want information, they wanted finished work. In April 2025 Genspark launched Super Agent, a Mixture-of-Agents system combining eight LLMs and 80+ tools that could research, build and deliver complete outputs. It beat Manus and OpenAI on the GAIA benchmark, and annualized revenue hit $10M in nine days and $50M+ within five months.

On November 20, 2025, Emergence Capital led a $275M Series B at a $1.25B post-money valuation, with SBI Investment, LG Technology Ventures, Pavilion Capital and Uphonest Capital joining and all existing investors doubling down. The round coincided with the launch of Genspark AI Workspace, which orchestrates 30+ models and 150+ tools to autonomously produce board decks, financial models, apps and reports.

The company frames itself as the 'autopilot' for the world's billion knowledge workers, and announced integration into Microsoft's Agent 365 ecosystem at Ignite the day before the Series B — a signal that enterprise distribution, not consumer search, is now the battleground.

What has to be true

  • AI search engines face monetization and publisher-cannibalization pressure, so Genspark moved up the value chain to agents that complete work enterprises will pay for.
  • The founders' search pedigree — Microsoft Bing, Google ranking, Baidu — gave the team credibility and a data-quality edge from day one.
  • Launching Super Agent as a benchmark leader created a viral proof point: $10M annualized revenue in nine days showed demand, not just hype.
  • The AI Workspace and Microsoft Agent 365 integration target the 1B+ knowledge-worker market, converting a consumer search app into enterprise infrastructure.

What can be applied

A wedge that builds users and data before the real business exists: free AI search drew an audience and sharpened evaluation, making the paid Super Agent credible at launch.

Aftermath

As of November 20, 2025, Genspark is a $1.25B unicorn with over $50M in annualized revenue five months after Super Agent's launch, 2M+ monthly active users, and a new enterprise product, Genspark AI Workspace, positioned as end-to-end autonomous work execution. It announced participation in Microsoft's Agent 365 ecosystem at Ignite 2025.

Sources

spotted an error? The archive wants to know.

Your turn

You just read one. Describe what you are building, and see who is betting on the same thing.

Free account · 3 free questions · no card

Related cases