From the Trenches of Sendbird to the Edge of Autonomous Commerce
Chase Kim arrived at Sendbird before the company's most consequential bet. By 2024, Sendbird - the messaging SDK company running the chat infrastructure behind DoorDash's delivery updates, Reddit's community threads, and Hinge's match conversations - was making a pivot. Not a soft pivot. A complete identity shift: from communications infrastructure to AI agent platform. Kim was in the room where it happened, and then some. He led it.
His role as Head of Forward Deployment put him directly at the intersection of customer problems and product ambition. Forward deployment is the unglamorous, indispensable work of taking a product to a customer, watching it fail in real conditions, and figuring out what it actually needs to be. At a company mid-pivot into AI agents, that job is everything. Kim wasn't writing product specs from a distance. He was the one watching enterprise teams try to build AI customer service on top of Sendbird's platform, seeing where it broke, and feeding that intelligence back into the product.
"A new kind of consumer business is possible - one capped by compute, not headcount."
Sendbird at the time of Kim's tenure was a unicorn with serious infrastructure under the hood. Founded in 2013 as a location-based app for moms, then pivoted to chat SDK after a hackathon experiment, then accepted into Y Combinator's Winter 2016 batch - the company had already survived multiple near-deaths and emerged as the largest private in-app conversations platform in the world. $270 million raised. SoftBank, Tiger Global, ICONIQ, Emergence Capital. $1.1 billion valuation. 290 employees. DoorDash, Match Group, Paytm on the client roster. Kim joined that story late enough to skip the early chaos, but early enough that the real transformation was still ahead.
The Conviction That Became a Company
At Sendbird, Kim worked alongside Sangha Park, who led the forward deployment team at the company. Two operators, deep in the machine, watching AI agents get deployed inside real enterprise workflows. At some point, the obvious question becomes inevitable: if AI can run customer support for a major consumer brand, what else can it run?
They left Sendbird together to find out. In 2026, Chase Kim and Sangha Park co-founded Light Anchor, accepted into Y Combinator's Spring 2026 batch (YC P26). The thesis is clean and audacious: consumer brands have always scaled by adding humans. What happens if you take the humans out of the operational loop entirely?
Light Anchor - YC P26
Light Anchor builds fully autonomous consumer brands run entirely by specialized AI agents filling the CEO, GM, Marketing, and Engineering roles. The flagship brand, Seoul Dispatch, operates as a real e-commerce store - no human employees. The company is headquartered in San Francisco's Bay Area and backed by Y Combinator, with Harj Taggar as primary YC partner. lightanchor.ai
Kim's background from Penn State's Information Systems Technology program gave him the systems-thinking orientation that forward deployment demands - you're not just solving a tech problem, you're solving a human-plus-tech integration problem. IST is a discipline that sits at the intersection of business process and technology implementation, which is exactly what AI agent deployment is at the enterprise level. It turned out to be good preparation for running a company where the agents handle the integration themselves.
The Sendbird Context: Why It Matters
Understanding what Chase Kim was part of at Sendbird matters for understanding what he's building now. Sendbird didn't just make a messaging SDK. Under CEO John S. Kim (a former professional gamer who built and sold a social gaming company to GREE before co-founding Sendbird), the company built the chat rails for entire industries. When you ordered food and got a real-time update, when you matched with someone on a dating app and the message appeared instantly, when a customer support agent's message landed in a mobile app - Sendbird was often the plumbing. 300 million monthly users across the platform. Invisible infrastructure at massive scale.
Sendbird's 2024 AI pivot wasn't a pivot away from that scale - it was an attempt to move up the stack. From "we handle the messages" to "we handle the agent that decides what messages to send." That's a very different business. And Chase Kim was the person responsible for making that new business work with actual customers, in actual deployments, at actual scale. That experience - watching AI agents succeed and fail in enterprise environments - is the intellectual capital he brought to Light Anchor.
Building Businesses That Run Themselves
Light Anchor's model is worth understanding precisely. The company doesn't sell software to help other businesses use AI. It builds and operates its own consumer brands using AI agents. Seoul Dispatch is a real store, selling real products, with AI handling everything from operations to marketing to decisions about what to stock. The company's two human founders set the direction; the agents execute.
The YC P26 batch is Y Combinator's first "Pioneers" cohort - a name that carries expectation. Kim and Park are betting that the constraints that have always limited consumer brands - how fast you can hire, how many people you can coordinate, whether humans need to sleep - dissolve when the operators are software. A store that can run a campaign at 3am, analyze results by 3:01am, and adjust by 3:02am has a different kind of operational ceiling.
Whether Light Anchor's model scales into a genuine category or remains an interesting experiment will take time to tell. But the conviction behind it is earned. Kim didn't read about AI agents and decide to build with them. He built the deployments that showed enterprise companies how AI agents could transform customer experience. That's a different kind of founder credibility.