Taking Japan-Born AI Products to the World: An Interview with Sakana AI’s VP of GTM

Sakana AI spent years as a research shop. In 2026, it started acting like a product company, rolling out Sakana Chat, Sakana Marlin, Sakana Fugu, and Sakana Translate, then formally launching its go-to-market team in September to sell them and push them into wider use.

The person leading that push is Hidekazu Kakishima, who has a habit of arriving when a technology is still finding its feet. He has done that with cloud, IoT, drones, and now AI. His pitch for Sakana AI is simple: the company has strong research, it is growing fast, and it wants to take Japan-born technology to the global market without waiting around.

That sounds neat until you get to deployment. Sakana AI is already working with industries like finance and defense, but Kakishima says a product launch does not mean a customer can use it cleanly on day one. That is why the company is hiring Forward Deployed Engineers, who will work at customer sites, feed what they learn back into the product, and slowly reduce the number of cases that need that extra hand-holding.

The company’s GTM approach is also more hands-on than a simple product sale. Sakana AI wants to start with the product, see how customers react, and then shape the deployment around the actual problem. Sometimes that means direct API use. Sometimes it means extra security work. Sometimes it means a partner, especially system integrators or technology companies with their own ecosystems, will carry the product out into the market.

One product drawing attention is Sakana Fugu, which uses multiple frontier models together on a single task instead of just letting users switch between them. Kakishima says customers are increasingly wary of depending on one model alone, especially after political or contractual disruptions can make a model unavailable. That concern, along with growing interest in AI sovereignty and data sovereignty, is helping Sakana AI get inquiries from a broad mix of customers, including smaller businesses.

The harder part is that AI sales are not cloud sales and not IoT sales. The market is bigger, customers are less sure what they need, and the models keep changing fast enough that nobody wants to lock in too early. Sakana AI’s answer is to build in Japan, keep product and development close together in Tokyo, and expand in parallel at home and overseas instead of waiting for one market to finish before starting the other.