The Idea: Eric Jing spent nearly a decade at Microsoft building Bing Search for Asia and creating Xiaoice - Microsoft's conversational AI that earned him the nickname "Father of Xiaoice." He then joined Baidu, where he eventually co-founded Xiaodu Technology (think: China's Amazon Echo) and scaled it to a $5.5 billion valuation. But in 2023, he saw something bigger. Traditional search was broken. Users didn't want links - they wanted answers. And the new generation of AI models finally made it possible to deliver them. So Eric left Xiaodu, partnered with his former CTO Kay Zhu, and moved to Silicon Valley to start fresh. This was the beginning of Genspark.
The Execution:
- Late 2023: Eric and Kay incorporate MainFunc (named after the core function in programming) and begin building in Palo Alto. Team of just 4 people.
- June 2024: Launch Genspark's beta and immediately close a $60M seed round led by Lanchi Ventures at a $260M valuation. The platform goes beyond traditional search, creating dynamic, synthesized "Sparkpages" from web sources instead of just links.
- Early 2025: Hit 5 million monthly active users. But Eric notices something in the data - users aren't just asking questions, they're trying to command the platform to do tasks. They want the business plan written, not instructions on how to write one.
- March 2025: Quietly close a $100M Series A, valuing the company at ~$550M. Begin pivoting from search to "Super Agent" - an autonomous AI platform that delivers finished work.
- September 2025: Launch Genspark AI Workspace publicly. Hit $36M ARR within 45 days of launch - with a team of just 20 people. One of the fastest revenue ramps in AI history.
- November 2025: Close $275M Series B led by Emergence Capital at a $1.25 billion valuation. Existing investors double down. Total funding reaches $435M.
- January 2026: Now targeting the billion+ knowledge workers globally with the pitch: "Where your work gets finished autonomously." Team is around 50 people and it took them just 9 months to hit $100M in ARR.
The lesson? Sometimes the best pivot is killing what's working. Eric had 5 million users on a successful AI search engine - and he walked away because the data told him users wanted more. Most founders would've optimized what they had. Eric rebuilt from scratch to give users what they actually needed.