AnswersPublished17 sources
What Peng Cheng Laboratory Announced About Cloud Brain III
Peng Cheng Cloud Brain III has put 10,000P of domestic AI computing into trial operation, but the facility is not yet fully built; its planned capacity is 16,000P. The October 15–16, 2026 conference in Shenzhen also released white papers on global computing networks and computing–power coordination.
Published byEdited with GPT-6 LunaImages generated with GPT Image 2
Research answer
Key takeaways
- Peng Cheng Cloud Brain III has put 10,000P of domestic AI computing into trial operation, but the facility is not yet fully built; its planned capacity is 16,000P.
- The October 15–16, 2026 conference in Shenzhen also released white papers on global computing networks and computing–power coordination.
AI Prompt
openai.comCreate a landscape editorial hero image for this Studio Global article: What did Peng Cheng Laboratory announce about its Shenzhen-based Peng Cheng Cloud Brain III at the 2026 World Computing Network Conference i. Article summary: Peng Cheng Laboratory announced that Peng Cheng Cloud Brain III’s 10,000P of domestic AI computing capacity had entered trial operation, moving the facility into a “building while operating” stage rather than announcing . Topic tags: general, general web, education. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, watermarks, charts with fake nu
Peng Cheng Laboratory announced that 10,000P of domestic computing capacity at Peng Cheng Cloud Brain III had entered trial operation during the 2026 World Computing Network Conference in Shenzhen. The milestone puts the facility into a “building while operating” phase—not full completion. Once built out, the system is planned to reach 16,000P, or 16 EFLOPS. 4
6
12
What the announcement confirms
The conference took place in Shenzhen on October 15–16, 2026. At its opening, organizers announced the trial operation and released two white papers alongside an update on Cloud Brain III. 2
4
The 10,000P figure describes the capacity put into trial operation, while 16,000P is the project’s planned total. A Shenzhen district government report describes the planned facility as the country’s largest single computing facility of its kind; that is a description of its planned scale, not evidence that the full capacity is already running. 6
12
Media coverage says the trial capacity is already being used for domestic large-model training and scientific-AI research. Reports also attribute a Huawei Ascend 384 supernode architecture to the system. That hardware detail is reported by media coverage; the supplied conference announcement does not independently establish a chip-supplier disclosure or verify the claimed role of the architecture in reducing data-transfer bottlenecks. 1
6
Two white papers focus on networks and energy
The conference released the World Computing Network Development White Paper and the Computing–Power Coordination Technology Development White Paper. 2
4
The global-network paper describes computing moving toward intelligent, clustered and cross-regional coordination, while identifying uneven resource distribution, ecosystem barriers and energy constraints as challenges. The computing–power paper addresses closer connections between computing and electricity systems, including coordination challenges such as mismatches between supply and demand across time and location. 4
19
What remains unclear in the available reports
The supplied coverage does not substantiate several more detailed claims associated with the announcement: a specific AI-for-science research network; a four-year national C2NET pilot with Cloud Brain III as a core node; cross-district links and proposed high-speed connections to Hong Kong and Macao; or Gao Wen’s cited ZPG targets and three-stage roadmap. The reports also do not confirm the white papers’ alleged coauthors or the specific 2026–2030 and 2031–2035 planning phases.
Earlier equipment milestones are not fully reconciled in the available reporting. Peng Cheng Laboratory’s research-progress page describes a 4,500P power-on test on December 29 but does not give a year in the excerpt; other reports say first-phase equipment was powered on in March 2026. The sources do not establish whether these refer to separate steps or explain the date difference. 1
10
The clearest takeaway is the distinction between capacity already in trial operation and the larger planned build-out: 10,000P has entered service on a trial basis, while 16,000P remains the stated full-capacity target. 4
6
12
Studio Global AI
Continue your research
This page includes a source-backed answer you can continue inside Studio Global.
People also ask
What is the short answer to "What Peng Cheng Laboratory Announced About Cloud Brain III"?
Peng Cheng Cloud Brain III has put 10,000P of domestic AI computing into trial operation, but the facility is not yet fully built; its planned capacity is 16,000P.
What are the key points to validate first?
Peng Cheng Cloud Brain III has put 10,000P of domestic AI computing into trial operation, but the facility is not yet fully built; its planned capacity is 16,000P. The October 15–16, 2026 conference in Shenzhen also released white papers on global computing networks and computing–power coordination.