Where Deep Computer Science Meets Mainstream Culture: Kaon AI Closes Series B to Define Content Engine of the Generative AI Era

Backed by Leading Investors Including B Capital, Redpoint Ace, Goodwater Capital, and DCM, Kaon's Berkeley-Educated Founders are Outpacing Big Tech in Building a Vertically Integrated AI Entertainment Empire.

SAN FRANCISCO, July 8, 2026 /PRNewswire/ -- Kaon AI, a generative AI platform and research lab pioneering the next era of interactive entertainment, today announced it has closed a Series B round to advance what is quickly becoming one of the most engaging generative AI content experiences. Backed by marquee tech and consumer investors, including B Capital, Redpoint Ace, Goodwater Capital, and DCM, the capital milestone validates an up-and-coming, Berkeley-educated founding team quietly re-architecting how millions of everyday users consume digital media.

While Big Tech races to automate the workplace, Kaon AI is capturing the most valuable commodity of the generative era: attention. The future of entertainment lies in long-horizon, hyper-personalized consumer content: a shift Kaon has capitalized on to build an AI-native ecosystem that is currently generating an eight-figure ARR. Led by CEO Jay Dang, CTO Alex Xi, and COO Lifan Wang, the company's flagship consumer product, Emochi, already commands an average daily session time of 150 minutes among more than two million daily active users. By treating deep computer science as a vehicle for mainstream culture, Kaon is achieving sustained user engagement metrics that are simply out of reach for most general-purpose AI startups today.

The Founders Rewriting the Economics of Trillion-Token Media
Traditional media is losing its audience. As AI drives the cost of content production toward zero, Kaon's Berkeley-bred founding team saw a different problem worth solving: not how to make more content, but how to make content that actually responds to the person consuming it, turning passive viewers into co-creators powered by AI

By focusing on content generation rather than distribution, Kaon inverts the legacy media model. Instead of matching a user to an existing video or story, Kaon's proprietary architecture serves as a "customization engine" that writes, illustrates, and evolves immersive multimedia narratives frame-by-frame based on real-time human behavior.

To make this vision economically viable, the technical team is building a custom full-stack inference infrastructure that manages more than a thousand GPUs across dedicated partners like Nebius and DigitalOcean. The result is an operation that processes trillion-level daily token volumes at roughly 10x lower cost than a legacy cloud structure.

"We believe interactive AI entertainment is emerging as the next distinct consumer category," said Daisy Cai, General Partner, B Capital. "Kaon stands out because Jay, Alex, and Lifan have built a full-stack powerhouse that owns its data pipeline from the ground up. That level of integration creates a compounding advantage in performance, data, and product iteration that is difficult to replicate, and we are excited to support the Kaon team as they help define this new category."

The Strategy Bridging Deep Computer Science and Mainstream Digital Culture
As model access commoditizes, Kaon's founders' insistence on a token-to-application structure has become a compounding competitive advantage. While most competing platforms rely on a partner-oriented open structure, the founding trio controls the entire ecosystem end-to-end, ranging from front-end consumer interactions and token economies to core inference infrastructure. This autonomy turns what is typically a massive operational bottleneck for startups into a highly defensible, low-cost moat.

This token-to-application ownership also shapes the team's approach to AI research. Their new research arm, Kaon Labs, will leverage Kaon's hyper-engaged, millions-strong consumer base as its live scientific sandbox, enabling rapidly iterative models to be brought to market in days, rather than months.

"Every streaming platform today recommends what could be your next favorite show from a long list of preset options. In the generative AI era, though, this finite catalog is poised to become the 'Blockbuster' of the 2020s," added Jay Dang. "At Kaon, we're building a new type of content engine to create individualized narratives that truly resonate, moving the industry past static text prompts and into real-time multimedia."

Scaling the San Mateo R&D Footprint for Personalized Content Generation
With deep roots in the Bay Area research community and at its founders' alma mater, UC Berkeley, Kaon Labs is actively in discussions with academic advisors, collaborators, and other top-tier research talent as they look to expand to a new R&D hub in San Mateo. For more information on Kaon Labs, including initial research documentation, technical pillars, and the core mission statement, visit kaonlabs.com.

About Kaon AI
Kaon AI is building the customization engine for AI-native entertainment. Moving beyond the era of static recommendations, Kaon AI uses real-world user behavior to generate personalized stories, characters, and video experiences in real-time. Headquartered in San Mateo, Kaon AI powers a global ecosystem of creators and consumers through its flagship products FlowGPT, Emochi, Tickle, and Branches. For more information, visit www.kaon.io.

MEDIA CONTACT:
Michael McKay
(406) 369-2093
[email protected]

SOURCE Kaon AI