CloudSky, a leading GPU-native Edge Cloud provider, participated in Web Summit Rio, one of the world’s largest technology conferences, where Kevin Xu, Co-founder of CloudSky, delivered a keynote speech titled “The Last Mile of Intelligence.”
During the session, Kevin shared his perspective on the evolution of computing infrastructure and discussed how edge cloud is becoming increasingly important as intelligent applications move into real-world environments.
According to Kevin, the rapid growth of intelligent applications is driving a fundamental shift in computing demand.
“Over the past two years, daily token consumption in China has grown from approximately 100 billion to 140 trillion, representing a 1,400-fold increase,” Kevin said during the keynote.
Industry forecasts suggest that inference workloads will account for more than half of total computing demand by 2025, grow to four to five times training demand by 2026, and drive a global inference market exceeding US$250 billion by 2028.
“As intelligent applications expand beyond data centers into real-world scenarios—including digital humans, robotics, autonomous driving, and real-time content generation—the challenge is no longer simply building intelligence,” Kevin said. “The challenge is delivering it where and when it is needed.”
Kevin noted that future computing infrastructure will increasingly rely on distributed architectures capable of delivering low-latency, real-time computing services closer to users, devices, and machines.
“Centralized clouds create intelligence. Edge clouds deliver intelligence,” Kevin said.
As computing demand becomes increasingly real-time and distributed, edge infrastructure is expected to play a growing role in supporting next-generation digital experiences.
Founded in 2016, CloudSky specializes in distributed GPU infrastructure and edge computing services. Powered by its proprietary StackGPU software-defined computing technology, the company has built a global edge computing network spanning more than 300 cities and serving over 600 million users.
Today, CloudSky’s infrastructure supports a wide range of latency-sensitive and compute-intensive applications, including cloud gaming, interactive video, digital humans, video generation, and real-time rendering.
As demand for real-time computing continues to grow, CloudSky remains committed to expanding its global edge computing network and working with ecosystem partners to support the next generation of intelligent applications and digital services.
Explore AITechPark for the latest advancements in AI, IOT, Cybersecurity, AITech News, and insightful updates from industry experts!