AWS is putting route diversity at the center of its subsea strategy. Its new Sta’O’Nuk cable will connect the United States and Japan. The system aims to add 420 Tbps of capacity across 20 fiber pairs.
The cable should enter service in 2029. It will use Hogan’s Corner, Washington, and a separate landing point in Japan. AWS wants to avoid traditional landing hubs in California, Oregon, Chiba, and Shima.
That decision matters for network resilience. When many cables land in the same area, one incident can cause broad disruption. Earthquakes, landslides, ship anchors, or political events can affect several routes together.
David Selby, director of global network planning and acquisition at AWS, explained the strategy clearly.
“Sta’O’Nuk addresses a concentration risk in today’s transpacific cable landscape. Existing systems between the U.S. and Japan land at traditional hubs in California, Oregon, and Japan’s Chiba Peninsula and Shima. By choosing new landing locations, Hogan’s Corner, Washington and a diverse point in Japan, Sta’O’Nuk creates geographically separated network paths that reduce the risk of a single event disrupting multiple systems,” he told RCR Wireless News.
Recent outages underline this point. In 2024, cable damage near Ivory Coast disrupted several systems. Separate issues also affected Red Sea routes. AWS says its network diversity helped Cape Town customers keep cloud access during that period.
Meanwhile, AI is raising expectations for global connectivity. Distributed model training moves data between many computing sites. Real-time inference needs fast responses and steady performance. These workloads need more than just bigger pipes.
“AI workloads like distributed model training and real-time inference need more than raw bandwidth. They need low latency, predictable performance, and resilient paths across continents. But AI is just one driver. Cloud computing, enterprise workloads, streaming, and edge computing are all fueling demand for more capacity and more resilient international connectivity,” Selby said.
For telecom engineers, the message is familiar. Capacity matters, but path design matters just as much. A high-capacity cable can still create risk if geography remains concentrated.
However, subsea builds now face longer delivery cycles. Permitting has become more complex across coastal and national jurisdictions. AWS says average timelines have moved from about three years to five to seven years.
“One of the most significant challenges facing the subsea industry is the growing complexity of the permitting process. Development timelines have extended from an average of 3 years to 5-7 years, largely driven by permitting requirements. We believe thoughtful regulation plays an important role in protecting security and resiliency, and we’re committed to working collaboratively with regulators to ensure the process supports, rather than slows, the buildout of critical infrastructure,” the executive said.
The project also reflects a wider industry shift. Hyperscalers now shape many major subsea routes. Telecom operators still remain vital partners for reach, operations, and regional knowledge.
For enterprises, Sta’O’Nuk promises stronger cloud access between North America and Japan. For the wider market, it shows where infrastructure planning is heading. Future digital economies will depend on diverse, secure, and scalable international networks.
