-- AI is shifting from model innovation to real-world deployment. How to cross the “Darwinian Valley of Death” between the laboratory and real-world scenarios has become a key topic at WAIC 2026, the World Artificial Intelligence Conference 2026.
On July 19, the WAIC 2026 AWAKEN Forum '26: AI Industry Innovation Network was held in Shanghai. The event brought together academicians and experts including Harry Shum, Guo Yike, Ning Guang and Dong Jiahong, Alvin Wang Graylin, researcher at the Stanford Institute for Human-Centered Artificial Intelligence (HAI), and Prof. Christoph Kletzer, Vice Dean and Professor at The Dickson Poon School of Law, King’s College London, as well as representatives from research, industry and investment, to discuss AI innovation and deployment.
Cloudsky Chief Scientist Dr. Lintao Zhang was invited to join the roundtable discussion titled “From Academic Papers to Commercial Delivery: The Treacherous Leap Across the Valley of Death of Scientist Entrepreneurs.” Together with guests from the investment community, academia and R&D institutions, he discussed how AI achievements can cross last mile of commercialization，and shared Cloudsky’s thinking on intelligence delivery.
Dr. Lintao Zhang is Chief Scientist at Cloudsky and leads Lab 12K, the company’s frontier research initiative focused on the Real-Time Intelligence Fabric. Drawing on Cloudsky’s years of experience in edge computing and distributed GPU infrastructure, his team explores next-generation technologies for Real-Time Intelligence Fabric. An IEEE Fellow and ACM Distinguished Scientist, Dr. Zhang is an internationally recognized researcher in computer systems, artificial intelligence, and cybersecurity.
Models Create Intelligence. Infrastructure Delivers It.
In his remarks, Dr. Zhang noted that many outstanding AI models struggle to reach real-world adoption not because of limitations in model capability, but because intelligence fails at the point of delivery. Academic research tackles difficult problems, but the practical value of those breakthroughs must ultimately be tested by the market and by real-world scenarios.
He framed the leap across the Valley of Death through three dimensions: people, place and timing. People means making technology understandable and explainable to its audience. Place is not only about geography, but also about identifying the right application scenarios and industry entry points. Timing means knowing how to act between technological excitement and market sobriety, which requires patience and judgment.
He emphasized that only continuous technological innovation can build a true moat. The positive cycle of “delivery, innovation and renewed delivery” is the underlying logic for crossing the Valley of Death.
From Edge Nodes to an AGI-Era Digital Nervous System
Guided by this thinking, Cloudsky formally upgraded its positioning this June from an edge intelligent computing service provider and launched its AGI-era strategy centered on building a Real-Time Intelligence Fabric network. In Cloudsky’s view, the past was defined by the idea that “wherever there is a screen, computing power is needed.” In the future, “wherever there is electricity, computing power will be needed.” From phones and glasses to cars and robots, demand for computing power is extending to every physical endpoint.
The next generation of computing infrastructure will operate like the human nervous system. The central cloud will serve as the brain, responsible for deep thinking and complex reasoning. Edge nodes will serve as nerve endings, enabling millisecond-level real-time response. Endpoint devices will act as the hands and feet, executing actions.
Cloudsky has now deployed hundreds of edge computing nodes across more than 300 cities worldwide, keeping end-to-end latency under 10 milliseconds and serving more than 600 million end users to date. These capabilities are already widely used in AI inference, intelligent cockpits, real-time rendering, cloud gaming and other scenarios, with deep collaboration across leading telecom operators, cloud service providers and device manufacturers.
The key to crossing the Valley of Death lies in whether intelligence can truly reach every place where it is needed. Looking ahead, Cloudsky will continue to advance the buildout of its global edge computing network, advance its Real-Time Intelligence Fabric strategy, and work with partners to explore the next generation of real-time intelligent experiences.
About Cloudsky
Cloudsky is a GPU-native edge cloud company providing distributed GPU infrastructure and real-time computing services for latency-sensitive, high-concurrency, GPU-intensive applications.
Powered by its proprietary StackGPU technology, Cloudsky operates a global edge computing network spanning more than 300 cities and serving over 600 million users worldwide.
For more information, visit www.cloudsky.com.
Contact Info:
Name: PR Team
Email: Send Email
Organization: Cloudsky
Website: http://www.cloudsky.com
Release ID: 89198524
In the event of detecting errors, concerns, or irregularities in the content shared in this press release that require attention or if there is a need for a press release takedown, we kindly request that you inform us promptly by contacting error@releasecontact.com (it is important to note that this email is the authorized channel for such matters, sending multiple emails to multiple addresses does not necessarily help expedite your request). Our dedicated team will promptly address your feedback within 8 hours and take necessary actions to resolve any identified issues diligently or guide you through the removal process. Providing accurate and dependable information is our utmost priority.
