Shanghai, July 18, 2026 — As the bodily boundaries of computing energy are more and more approaching, the place will human computing civilization head to? On the afternoon of July 18, as one of many flagship particular boards of the 2026 World Artificial Intelligence Conference (WAIC), the “Future Computing · Future Computing Power” particular discussion board was held within the Red Hall of the Shanghai World Expo Center. Led by the University of Science and Technology of China and the School of Computing of City University of Hong Kong, the discussion board was co-organized with Fudan University, Peng Cheng Laboratory, Shanghai Computing Power Network Association and different establishments. It gathered prime students, academicians, entrepreneurs and traders from the United Nations, China, the United States, Europe and Japan, to hold out in-depth interdisciplinary, cross-regional and intergenerational dialogues specializing in “Reconstructing the Computing Power Paradigm in the Post-Moore Era”.
The agenda spanning greater than 4 hours was organized as a six-scene journey from “Where We Are” to “Where We Are Going” — protecting coordinate positioning, bottleneck analysis, path outlook, industrial verification, to important considering and ecosystem launch. The discussion board kicked off with the discharge of the primary satellite tv for pc cluster of Shanghai’s “StarPivot Initiative” area computing energy constellation, and concluded with the discharge of three main achievements together with the Quantum Intelligence Industry Alliance, marking that China has entered a important stage of shifting from laboratory analysis to industrialization on the three most important traces of future computing energy: quantum, area, and inexperienced.
I. Starting Point: Computing Power Has Reached the Critical Point of “Paradigm Shift”
The core judgment of the discussion board is a repeatedly reaffirmed consensus — When the expansion of computing energy hits bodily boundaries, the reply lies not in “more”, however in “different”.
Alexander De Croo, Administrator of the United Nations Development Programme and former Prime Minister of Belgium, reminded in his opening speech that synthetic intelligence seems to be digital in nature, however basically depends on bodily infrastructure of vitality, water, supplies and expertise. Without reconstructing the underlying paradigm, the worth of AI can’t be really popularized throughout the globe. Chang Jin, President of the University of Science and Technology of China and Academician of the Chinese Academy of Sciences, took the “Dark Matter Satellite” for example and identified that Moore’s Law is approaching its restrict, and vitality, supplies, warmth dissipation, and water sources have turn into widespread world boundaries. “The physical limit of computing power cannot be addressed by extending the traditional track, but lies in entirely new directions such as quantum computing, space computing, and new physical principles.”
In his opening speech, Wei Mingde, Chairman of the Council of City University of Hong Kong, Member of the National Committee of the Chinese People’s Political Consultative Conference, and Member of the Legislative Council of Hong Kong, supplied a transparent historic coordinate for this judgment — “Advanced computing power and computing are becoming the infrastructure of this era, just like the power grid in the last century and the Internet in the last decade. It truly determines the upper limit of a society’s productivity, and also determines how far a civilization can go.” He additionally reminded all attendees that when the infrastructure undergoes transformation, universities, industries, governments, and capital all have to re-answer the roles they play on this civilization-building undertaking.
In the primary dialogue session, Gao Wen, Director of Peng Cheng Laboratory and Academician of the Chinese Academy of Engineering, and Guo Yike, Vice President of the Hong Kong University of Science and Technology and Academician of the Chinese Academy of Engineering, held a landmark public dialogue on the “China Computing Power Network”. As a first-prize achievement of this yr’s National Science and Technology Progress Award, the China Computing Power Network has for the primary time really engineered and carried out the 20-year-old imaginative and prescient of “grid computing” — “making computing power accessible to thousands of households like electricity”, changing into a national-level computing energy financial base protecting coaching, inference, and trade purposes.
Distinguished Bloomberg Professor at Johns Hopkins University and Academician of the U.S. National Academy of Sciences Professor Daniel Kammen, in his keynote speech titled “The Energy Bottleneck: The First Wall of Computing Power”, pushed the problem to a extra elementary stage — “The next wall no longer lies in the chips, but on the other side of the power grid scheduling table.”
In the master-led session of the third act “Where We Are Going”, Pan Jianwei, Academician of the Chinese Academy of Sciences and Executive Vice President of the University of Science and Technology of China, set the tone for your complete part through on-line video. Starting from the historic coordinate of the one hundredth anniversary of the delivery of Schrödinger’s equation this yr, Academician Pan Jianwei identified that quantum mechanics has spawned many main applied sciences which have modified human life over the previous century, and the vigorous improvement of rising quantum data applied sciences right this moment “may push human civilization to achieve another huge leap”. He quoted Feynman’s most well-known maxim — “If we could pass only one sentence to the next generation, it would be: All things are made of atoms” — as an example the scientific origin of quantum computing: all matter consists of atoms that comply with the legal guidelines of quantum mechanics, and many-body methods can’t be successfully simulated by classical computer systems. Quantum computing originated exactly from Feynman’s pioneering concept of “using controllable quantum bits to simulate complex quantum systems”. Therefore, quantum computing cannot solely crack sure classical cryptographic methods, but in addition assist people deeply perceive the pure world, promote leapfrog improvement in fields comparable to chemistry, supplies science, and life sciences, and convey thrilling breakthroughs in instructions comparable to sustainable progress of computing energy, privateness safety, public security, and life well being.
Academician Pan Jianwei additionally put ahead three calm judgments that set the tone for the quantum narrative of your complete discussion board: First, though a collection of milestone achievements in quantum computing superiority have been made, a general-purpose quantum pc that may shortly clear up sensible issues with particular worth nonetheless requires long-term analysis and improvement efforts. Second, to realize this purpose, steady monetary funding from each the federal government and the non-public sector is important. Meanwhile, “the academic community has the responsibility to guide the public to view the development of quantum computing rationally and guard against industry bubbles”. Third, identical to the event of quantum mechanics within the Heisenberg period a century in the past, “extensive and active international cooperation and exchanges will surely enable quantum information technologies to benefit all mankind faster”.
Shortly afterwards, on the roundtable hosted by Wang Jianping, Dean of the School of Computing of City University of Hong Kong, titled “Exploring the New Realm of Quantum Computing Power, Heading for the Future of Computing”, Lu Chaoyang, Professor of the University of Science and Technology of China and core member of the Jiuzhang Quantum Computing Team, identified with the most recent breakthrough of “Jiuzhang 4” that quantum computing is shifting from “laboratory principle verification” to “surpassing classical performance on dedicated problems”, marking that one other most important line of computing has been opened.
Standing at this coordinate, Lin Jing, tutorial planner of the discussion board and Professor of the School of Computing of City University of Hong Kong, identified in her opening internet hosting remarks: “This is not just six sessions, but also an exploration journey.” Starting from this path, the discussion board gathered clear consensus at 5 ranges for the nation, training, trade, capital, and purposes on how you can launch the “new computing power civilization”.
II. National Level: Plan Computing Power Holistically as a “New Generation Infrastructure”
If the infrastructure of the twentieth century was the facility grid, and the infrastructure of the early twenty first century was the Internet, then the infrastructure of the subsequent decade will inevitably be an clever computing community composed of classical computing energy, quantum computing energy, area computing energy, and inexperienced vitality.
The first core consensus reached by the discussion board is: Computing energy have to be holistically deliberate as a national-level infrastructure, in order that it’s “accessible to all and available on demand” identical to the facility grid.
Top-level planning upfront. The nation ought to, as quickly as potential beneath the fifteenth Five-Year Plan framework, incorporate new types of computing energy comparable to quantum computing energy and area computing energy into the unified format of the nationwide computing energy community, avoiding solely chasing progress on the one monitor of “classical GPU clusters”. Shanghai’s “StarPivot Initiative” with 1000 space-based computing energy satellites, the eight main hubs of the China Computing Power Network, and the State Key Laboratory of Quantum Computing are basically all elements of the brand new technology infrastructure, which must be coordinated and promoted on the identical nationwide blueprint.
Upgrading infrastructure considering. The “cloud-native network” four-element mannequin of the China Computing Power Network proposed by Academician Gao Wen — chips, energy, communication, and storage — gives a brand new financial basis for nationwide computing energy infrastructure. The key to computing energy infrastructure is now not “how many racks to build”, however “whether it can be scheduled like a power grid, interconnected like a highway, and billed like water and electricity”.
Space computing energy initiatives comparable to “StarPivot” and “StarComputing” require national-level coordination on frequency orbits and requirements. Lu Chuan, Chairman of Guoxing Aerospace, reminded on the discussion board that low-orbit frequency orbit sources are “first-come, first-served and non-renewable”. Whoever takes the lead in format will grasp the initiative of future computing energy. This requires strategic ensures from the nationwide stage in spectrum allocation, launch preparations, and worldwide coordination.
Computing-power and vitality coordination and a inexperienced basis. Professor Kammen repeatedly emphasised in his speech that earlier than controllable nuclear fusion achieves breakthroughs, vitality is the ceiling for AI. China has already explored a path of shifting computing energy to areas wealthy in inexperienced energy by way of the “East Data, West Computing” initiative. The subsequent step is to construct “green power + energy storage + computing power” right into a nationwide standardized product, making inexperienced the model of China’s computing energy going world.
III. Education and Academia: Plant the “Next Generation Paradigm” into the Soil 20 Years in Advance
The second core consensus of the discussion board comes from the twin reminders of President Chang Jin and Wei Mingde, Chairman of the Council of City University of Hong Kong — Without the decades-long persistence of universities, there could be no technological frontiers that we are able to focus on right this moment. Professor Franco Nori, Chief Scientist of RIKEN (Japan) and Academician of the European Academy of Sciences, in his keynote speech titled “The Future of Quantum Computing”, used a golden quote to offer a tutorial footnote to this judgment for the worldwide neighborhood — “This is not a 100-meter dash, but a marathon, even a 100-kilometer ultra-marathon.” He systematically reviewed the three “winters” that synthetic intelligence has skilled because the Fifties, reminding all attendees that expertise by no means develops in a straight line, and a technology of applied sciences that actually change civilization requires many years of secure funding and tutorial accumulation.
Moving disciplines ahead. Twenty years in the past, quantum data and inexperienced computing energy have been nonetheless area of interest disciplines; right this moment they’re on the middle of the discussion board, due to “that group of young scholars who were willing to sit on the cold bench”. The course of computing civilization within the subsequent 20 years additionally relies on the younger folks we begin cultivating right this moment. The training system ought to, as quickly as potential, incorporate cutting-edge fields comparable to quantum science, computational physics, vitality system modeling, and AI system structure into undergraduate fundamental programs and interdisciplinary building, as an alternative of treating them as “specialized directions” solely on the postgraduate stage. Professor Nori significantly identified that there are presently 4 mainstream technical routes for quantum computing: superconducting, photonic, ion lure, and Rydberg atoms. “There is no conclusion yet on which one will win, and it is very likely that at least four platforms will need to coexist in the future” — which signifies that when cultivating cutting-edge abilities, universities should preserve a disciplinary format with a number of parallel paths, relatively than betting on a single course.
Interdisciplinary collaboration. Chairman Wei Mingde identified that “every practical problem in computing and computing power cannot be solved independently by any single discipline” — vitality and area require the joint efforts of physics, chemistry, supplies science, and coverage analysis; quantum computing requires coordinated progress throughout principle, engineering, algorithms, and trade. Universities are one of many few locations the place completely different disciplines can have interaction in equal dialogues beneath the identical roof, a worth that no company laboratory or authorities company can change.
International interface. City University of Hong Kong has been rated as probably the most worldwide college on the earth for three consecutive years, with its common quotation fee of educational papers rating second globally. Chairman Wei Mingde emphasised that Hong Kong’s distinctive worth doesn’t lie in how large-scale computing energy it could possibly develop by itself, however in its openness and inclusiveness — it could possibly function a global interface for national-level laborious expertise initiatives, appearing as a connection level for abilities, capital, and tutorial cooperation. This judgment is extremely in step with Academician Pan Jianwei’s emphasis (citing the Heisenberg period) that “extensive and active international cooperation and exchanges will surely enable quantum information technologies to benefit all mankind faster”.
IV. Enterprise Level: Achieve the Utmost of “Pragmatism”
The third core consensus of the discussion board comes from the collective reflection of frontline enterprises in quantum, area, and AI infrastructure sectors — Beware of exaggeration and “digital performance”, and take each step all the way down to earth. This consensus obtained a direct response from the worldwide tutorial neighborhood: Professor Franco Nori sharply requested in his speech — “If you delete all marketing videos, company press releases, and promotional materials, then ask a very simple question: What exactly has been sold today? What practical, useful, and profitable real-world problems have been solved today?” He frankly said that quantum computing presently has thrilling preliminary achievements, however “its scientific foundation is solid, technological progress is real, and its prospects are of great significance — yet valuations often depend on hype overcoming logical gravity”. Such a spot is strictly the fact that every one laborious expertise enterprises have to face.
In the important considering session of the fifth act, a number of tutorial and trade representatives with the identical instructional background (Xiong Hui, Associate Vice President of the Hong Kong University of Science and Technology (Guangzhou); Yao Xin, Vice President of Lingnan University; Wang Bingfeng, Co-Chairman of Digital China; Liang Jia’en, Co-Founder of iFlytek’s CloudWalk) launched a pointy debate on the “token economy”. Vice President Yao Xin’s phrases hit the nail on the pinnacle: “The best intelligence is to use the least resources to accomplish the most things.” Simply speaking in regards to the amount and value of tokens with out addressing the duty worth accomplished per unit token is a type of technological self-indulgence disconnected from actual productiveness.
The motion suggestions for enterprises are clear:
- Do not fabricate numbers for “storytelling”. At the discussion board, many friends implicitly criticized unsubstantiated claims within the trade comparable to “a certain company’s quantum-related business accounts for 40%”. There are not any shortcuts within the laborious expertise sector; each exaggerated share level should be paid again throughout the engineering section. Zhu Xiaobo, Professor of the University of Science and Technology of China and Chief Designer of the “Zhaozhi” quantum computing prototype, said bluntly in his speech : “Do not believe any hype about ‘thousands or tens of thousands of qubits’. Without sufficient precision and reduced error rates, even the largest scale is meaningless. Excessive promotion will only bring devastating blows to the entire field.” His metaphor that “the target to conquer is Mount Everest, and we are still conducting ground experiments” units probably the most easy tone for the engineering posture of home quantum enterprises.
- Follow the trail of “three red lines”. Wang Kangman, Partner at 3C Capital,