AI chipmaker Iluvatar CoreX says new GPU matches Nvidia’s H200 as home-grown push accelerates

Chinese artificial intelligence chipmaker Iluvatar CoreX claims that its new Tiangai 300 graphics processing unit (GPU) can match and even surpass Nvidia’s H200 in real-world AI workloads, highlighting efforts by domestic players to close the technology gap with the US chip giant.

In a live demonstration for mainland Chinese media outlets on Wednesday, the Shanghai-based firm showed the Tiangai 300 outperforming the H200 in complex multimodal tasks, including text-to-image and text-to-video generation using frontier AI models such as Z.ai’s GLM-5.3, according to a report by Chinese tech media outlet Xinzhixun.

In one text-to-image test, the Tiangai 300 achieved a 100 per cent completion rate with a throughput of 0.101 images per second, compared with the H200’s 87.5 per cent success rate and 0.086 images per second, although the Chinese chip took 0.64 seconds longer to complete the task, Xinzhixun reported.

The demonstration corroborated findings from earlier third-party benchmark tests conducted by state-backed Shanghai INESA. Those tests showed the Tiangai 300, unveiled in July, reaching 1.05 times the throughput of the H200 “under typical deployment conditions”.

That made it the first Chinese general-purpose GPU to match the US processor in empirical testing, the report said.

A separate report by Chinese financial outlet Wallstreetcn on Thursday noted that the Tiangai 300 beat the H200 in mixture-of-experts (MoE) throughput by 15 per cent, while slashing first-token latency by up to 30 per cent in long-text processing. MoE is an architecture that routes tasks to specialised sub-networks to save computing power.

In AI, training is the resource-intensive process of teaching a model to learn from vast data sets, while inference is the subsequent stage of putting that trained model to work answering live user prompts.

On the AI training front, the Tiangai 300 delivered 1.33 times the effective throughput of Nvidia’s prior-generation H100 when fine-tuning Alibaba Group Holding’s open-source Qwen3-32B model, the report added. Alibaba owns the South China Morning Post.

Iluvatar did not immediately respond to a request for comment on Thursday.

The performance claims come as China’s top GPU contenders – including Iluvatar, Biren Technology, MetaX Integrated Circuits, Enflame Technology and Moore Threads – are racing to deliver viable home-grown substitutes under Beijing’s self-sufficiency drive and tightening US trade curbs.

The H200, which Nvidia launched in late 2023, has become a key focal point in Sino-US tech rivalry. While Washington eased conditions earlier this year under a revised licensing framework to permit H200 shipments to China, Beijing has discouraged domestic tech champions from buying foreign chips to bolster the local ecosystem.

Nvidia said in August that it had commenced its first H200 data-centre processor shipments to China under the new US rules, ending an extended supply freeze.

Iluvatar, which went public in Hong Kong earlier this year, has laid out ambitious targets to counter US dominance. In January, the firm unveiled a multi-year architecture road map aiming to leapfrog Nvidia’s advanced Rubin platform within two years.

To prove commercial viability beyond lab tests, Iluvatar said earlier this month that its Tiangai series had reached 10,000-card-scale deployments, giving it crucial experience in system-level delivery and operations maintenance.

The chipmaker is also expanding beyond cloud-connected data centres into edge computing – running AI directly on local machines, vehicles and connected devices.

Iluvatar’s Tongyang TY1200 processor outperformed Nvidia’s Thor T5000 by 20 per cent in specific large language model tests, according to Wallstreetcn.