$36M ARR in 45 Days—Building a Trillion Token Company | Wen Sang, Co-Founder of Genspark
Upcoming events:
From GTM to Hypergrowth: How Alignment Fuels Scale | Time: Monday, Oct 27 6:00 PM - 9:00 PM PDT | Location: Stanford University | Register: https://luma.com/no1atpxz
Chen Institute Symposium for AI Accelerated Science | Time: October 27–28, 2025 | Location: The Conservatory at One Sansome, San Francisco, CA | Register: https://aias2025.org
Wen Sang is the Co-Founder and COO of Genspark, an AI agentic engine company pioneering the future of work for knowledge professionals. A PhD graduate of MIT’s Mechanical Engineering, Wen went from combustion research in MIT’s automotive lab to founding Smarking, a Y Combinator-backed startup that digitized America’s parking industry before its acquisition by ParkHub in 2022. Now, Wen channels those hard-won insights into building AI agents at Genspark that collaborate, learn, and think like humans—applying the same systems thinking that once revolutionized parking to the frontier of artificial intelligence. Genspark has reached a major milestone as a trillion token company, recently highlighted by OpenAI at OpenAI Dev Day.
Listen to the full interview:
Part 1:
Part 2:
Key Takeaways
The Nine-Year Education
Before Genspark, Wen built Smarking (parking management SaaS) for nearly a decade, digitizing 2,500+ facilities, joining Y Combinator and eventually selling to ParkHub in 2022. It was successful but exhausting—”We delayed every life decision,” Wen reflected. “My wife didn’t say much, but it was hard on her.” The lesson: “Solving the right problem gives you more than solving every problem correctly.” Parking needed digitization, but it was heavy lifting with 500+ ETL jobs and educating an industry where vendors asked “What is an API?” When Wen met Eric Wang (ex-Microsoft, ex-Baidu, $5B company builder) in 2023, they asked: What if we bet on the AI wave instead of fighting it?
The Insight That Changed Everything
Vinod Khosla at YC said: “People who take advantage of what AI could do would do really well—not people who build models, people who use them.” Instead of competing with OpenAI/Anthropic/Google on capital-intensive models, Wen built ON TOP with a 3-layer architecture: model orchestration (routing tasks to GPT, Claude, Gemini), tool layer (building “arms and legs” for AI execution), and work products (delivering finished slides, sheets, documents, code). “ChatGPT chats,” Wen explained. “It hasn’t tapped 1% of AI potential. We don’t talk much—we think, we do, we deliver.”
Why Speed Is the Only Moat
Genspark operates with under 30 people—everyone’s a full-stack developer, 1-2 people own entire projects, and they ship weekly or daily with a completely flat structure. “Large companies can’t organize this way,” Wen said, because AI landscapes shift faster than traditional planning cycles—by the time you specify GPT-4 features, GPT-5 changed what’s possible. Look at the cycle: 2022 was GitHub Copilot, 2024 was Cursor, 2025 is Claude Code—each wave erases the previous moat. “This is the worst version you’ll ever use. We’ll make it better day after day.”
The 45-Day Playbook
Here’s how to go from $0 to $36M ARR: pick painful problems (30M knowledge workers), make it 10x better (”work taking a week done in 2 hours”), start self-serve at $20/month, let users pull you upmarket with enterprise deals, and deliver monetizable value where users make money, not just save time. Users don’t just consume—they generate revenue, which is why word-of-mouth exploded.
The Founder Behind the Growth
Wen’s first company nearly failed in 2017 from “people problems”—”I understood technology better than humans,” he admitted—until an executive coach prescribed meditation (Headspace: “A month later, whoa, I can see things”) and creating space before responding (”Don’t rush to respond—give yourself time to prepare”). His advice to his younger self: “Solving the right problem gives you more than solving every problem correctly” and “Take care of yourself first—physically, mentally—then family, then everything else follows.”
The Future of Work
Will AI replace knowledge workers? Wen disagrees: “When AI does more, humans get elevated—everybody becomes a boss with a fleet of Goldman Sachs associates in your pocket who are competent, smart, fast, never tired.” It’s not AI taking jobs—it’s people using AI taking jobs from people who don’t. Jensen Huang says “I never work, I just dream,” and Wen wants that for everyone: “I want everybody to feel like, damn, I’m dreaming every day.”
Reference:
Wen Sang: https://www.linkedin.com/in/wen-sang/
Jing Conan Wang (Host): https://www.linkedin.com/in/jingconan/
FounderCoHo: https://www.linkedin.com/company/foundercoho/
Timestamps:
Part 1
00:00 - Highlights
01:25 - Introduction and Interview Focus
02:48 - Introduction - Growing up in China, coming to MIT
05:04 - Early MIT experience and academic path
07:15 - The parking problem that sparked the startup idea
09:23 - Turning the idea into a real business
10:17 - Building the initial team with college friends
11:10 - Funding challenges - Boston vs Silicon Valley
15:13 - Y Combinator experience and Sam Altman
18:39 - Developing communication and leadership skills
21:01 - Executive coaching and meditation practice
25:02 - The Smarking business model explained
28:45 - Building the parking data network
30:38 - Scaling challenges and team growth
34:20 - Nine years of dedication to the mission
37:30 - The exit decision and sale to ParkHub
38:50 - Post-exit tinkering year and vehicle project
42:06 - Lessons from hardware startup attempts
43:13 - Outro (Preview to Part 2)
Part 2
00:00 - Highlights
01:29 - Introduction and Interview Focus
02:50 - The Genspark Journey
07:49 - AI Technology Strategy and Architecture
10:14 - Eric’s Background and Impact
14:09 - The 2025 Agentic AI Pivot
16:54 - Explosive Growth: $36M ARR in 45 Days
18:07 - Why PMF is a “Last Generation Term”
20:02 - Business Philosophy: No Moats in AI
23:22 - Customer Success Stories and Use Cases
29:13 - Challenges of building the Best Product
33:51 - Building with AI: Team and Operations
37:57 - Future of Work Vision and Competition
40:10 - Wen’s Vision for Genspark
42:06 - Genspark’s Mission
44:17 - Thoughts on AGI
48:59 - Founder Advice and Closing Thoughts
50:51 - Outro