I recently watched Biteye’s interview with GOAT Network CEO @kevinliu at KBW, and it really got me thinking.
Whether a project can go the distance often depends on whether its founding team has genuinely built products for users—and whether they’re willing to admit when an approach hasn’t worked and adjust the product accordingly.
@kevinliu didn’t start out in crypto. Before that, he was an entrepreneur in the mobile internet space. The live-streaming app his team built was later acquired by China Radio International and reached around 160 million downloads. 📱
🗓️ After entering Web3 in 2018–2019, the team first built Metis, an Ethereum Layer 2 network, delivering nearly 300x returns to early investors. In 2023, they launched GOAT with a clear thesis: if Bitcoin is merely held and its network lacks real-world use, miner incentives and long-term security will be difficult to sustain, and the foundation of the asset’s value will remain shaky.
The team has experimented with DeFi and BTCFi on Bitcoin Layer 2, but the results were disappointing. Kevin’s assessment is straightforward: the concept of “Bitcoin Layer 2” alone isn’t enough.
Ethereum projects with existing capital, communities, and product traction have little reason to migrate to a new chain. So the team pivoted to Agent payments, integrating Coinbase’s x402 and Stripe’s Machine Payments Protocol, while also supporting fiat.💳
Agents only need to make payments on a chain that supports them; the underlying infrastructure handles cross-chain transactions.
✨ The product behind this is GOAT Flow. When Agents shop today, they’re more likely to find major platforms or existing partners. Products from smaller merchants, or those with unique features or price advantages, are often difficult to discover through search.
GOAT Flow lets merchants enter product information almost as easily as filling in a registration form, then automatically generates payment pages that machines can read. This makes products easier to understand, discover, and purchase. Payments can be made with either cryptocurrency or credit cards.🤖
The product direction has shifted, but the underlying logic hasn’t: final settlement on the crypto side should still return to the more secure Bitcoin network.
The proof at KBW was more concrete than the narrative. The booth featured a process for purchasing GOAT plush toys and Starbucks gift cards through an Agent, and the line at one point stretched all the way to the entrance.🧸
More than 300 items were sold on the first day, including over 100 plush toys and more than 200 gift cards. Over 450 people visited during the event, and Agents made more than 310 purchases.
Kevin @kevinliu believes this is the first time the team has seen a fairly clear signal of approaching product-market fit. Feedback from the Korean market has also exceeded expectations.✨
In the interview, he also mentioned an uncomfortable reality: at some events in Silicon Valley, some Web3 professionals are less willing to openly identify themselves, for fear of being seen as riding the AI hype.
He believes this scrutiny is unsurprising. Many projects in the past, including their own, failed to deliver results that could be used at scale.
But as small, frequent payments between machines become more common, the traditional financial system is ill-equipped to handle this demand. Blockchain is more likely to become part of the infrastructure for AI payments. Whether that proves true will ultimately depend on the products and data.
⭐️ From a consumer internet product with 160 million downloads, to Metis, and now bringing Agent commerce settlement back to Bitcoin, Kevin and his team have consistently tested whether infrastructure is truly needed through real-world use—not just by shifting from one concept to another.
It’s precisely this user-first approach—being willing to change direction while keeping settlement security anchored to Bitcoin—that makes @GOATNetwork worth watching. I’m confident the team will continue to do great things under @kevinliu’s leadership.