The 2026 World Artificial Intelligence Conference (WAIC) was held from July 17 to 20, where more than a dozen domestic GPU vendors, including Huawei, Biren, and Moore Threads, collectively showcased their full-stack capabilities from inference to training.
Huawei publicly unveiled the Atlas 950 supernode for the first time at WAIC, consisting of 1,024 Ascend cards—the largest‑scale supernode in the industry—with cluster scalability up to 500,000 cards.
Oriental Computing globally launched its first high‑performance AI chip DF1000 on the eve of WAIC, built on a domestic 14nm process and claiming performance comparable to 4nm chips.
On July 24, the Lijian‑1 Yao‑15 rocket launched the Jitian Star A‑04 satellite, carrying a GPU computing payload from Moore Threads, marking the first in‑orbit deployment of domestic GPU computing power.
Industry data show that China’s AI computing demand grew 417% year‑on‑year in Q1 2026, while supply grew only 128%, widening the supply‑demand gap.
NVIDIA reported its FY2026 results, with quarterly revenue of $81.6 billion (up 85% YoY) and full‑year revenue of $215.9 billion, maintaining an 80–90% share of the AI data center GPU market.
AMD launched its rack‑scale AI system “Helios” on July 23, featuring the new Instinct MI455X GPU (2nm process, 432GB HBM4) with 40.26 PFLOPS peak compute at MXFP4 precision.
OpenAI announced large‑scale deployment of AMD Helios racks starting in the second half of 2026, and is co‑developing the next‑generation MI500 series AI accelerators with AMD.
Intel posted Q2 2026 revenue of $16.1 billion (up 25% YoY), with its Data Center and AI segment generating $6.3 billion (up 59% YoY), marking its fastest quarterly growth in nearly 15 years.
AMD CEO Lisa Su forecast that the AI accelerator market will reach $1.4 trillion by 2030, with GPUs capturing the vast majority of that share, driven by step‑function growth in compute demand from agentic AI.