{"id":597749,"date":"2026-05-22T09:35:09","date_gmt":"2026-05-22T09:35:09","guid":{"rendered":"https:\/\/www.newsbeep.com\/uk\/597749\/"},"modified":"2026-05-22T09:35:09","modified_gmt":"2026-05-22T09:35:09","slug":"alibaba-just-admitted-its-struggling-to-keep-up-with-rival-chipmakers-and-ai-shops","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/uk\/597749\/","title":{"rendered":"Alibaba just admitted it\u2019s struggling to keep up with rival chipmakers and AI shops"},"content":{"rendered":"

Systems<\/p>\n

Reveals decent new homegrown accelerator and tiny production volumes<\/p>\n

Chinese tech giant Alibaba has revealed a new accelerator and accompanying rack-scale server rig without offering much detail about their performance \u2013 and also admitted it\u2019s only been able to make chips in trivial quantities.<\/p>\n

The new chip is called the Zhenwu M890, and comes from Alibaba\u2019s semiconductor design business T-Head. Neither company has said much about it other than stating it includes 144GB of on-chip memory, possesses \u201c800 GB per second of inter-chip bandwidth\u201d and natively supports precision formats from FP32 down to FP4.<\/p>\n

The Chinese giant didn\u2019t offer any info about performance other than to say it delivers \u201cthree times the performance of its predecessor, Zhenwu 810E.\u201d Based on the specs of the old and new devices, we think the M890 might give Nvidia\u2019s 2024-vintage H200 a run for its money.<\/p>\n

That means the most interesting figure in Alibaba\u2019s announcement<\/a> is 560,000 \u2013 the number of Zhenwu chips Alibaba says T-Head has made to date.<\/p>\n

By way of contrast, Nvidia says AWS alone will rack and stack one million of its GPUs this year. AWS\u2019s spending on AI infrastructure is at similar levels to Microsoft, Meta, and Google, so it\u2019s conceivable that Nvidia will make and sell three or four million GPUs to satisfy those four customers alone.<\/p>\n

Alibaba\u2019s announcement doesn\u2019t offer any information about production volumes for the M890.<\/p>\n

The company did talk up the machines the M890 will run inside \u2013 a new beast called the Panjiu AL128 Supernode Server Alibaba described as \u201ca rack-scale system that packs 128 AI accelerators into a single unit and delivers petabyte-per second internal bandwidth \u2026 designed specifically for the concurrency patterns that agents generate: unpredictable, high-frequency bursts of inference requests that overwhelm conventional compute clusters.\u201d<\/p>\n

It seems Alibaba intends racks packed with M890s plus Panjiu AL128s to handle agentic workloads.<\/p>\n

T-Head has also created a new networking chip called the \u201cICN Switch 1.0,\u201d which we\u2019re told \u201cdelivers up to 25.6 Tbps of aggregate bandwidth and enables congestion-free communication across clusters of 64 accelerators.\u201d Those are specs that Broadcom and Nvidia reached years ago.<\/p>\n

\"Alibaba<\/p>\n

\n Alibaba’s T-Head M890 chip and Panjiu AL128 Supernode Servers<\/p>\n

Alibaba\u2019s chips therefore deliver performance that leaves its cloud well behind its western competitors, which would be a problem if its Chinese peers were buying kit from Nvidia. But despite the US lifting export restrictions on some advanced AI hardware, Beijing has not let local buyers acquire any: Nvidia recently told investors it doesn\u2019t expect to win any revenue from China for the foreseeable future.<\/p>\n

Chinese hyperscalers haven\u2019t announced capex spending at anywhere near the levels of their American rivals. Perhaps we now know why: T-Head can\u2019t get a lot of gear made, probably because accelerators like the M890 require an advanced semiconductor manufacturing process that Chinese companies can\u2019t access from offshore fabs due to US sanctions. Chinese fabs can\u2019t yet match the prowess of TSMC, the source of most high-end GPUs.<\/p>\n

So how are China\u2019s tech giants meeting demand? We\u2019ve previously covered attempts to smuggle Nvidia parts into China<\/a> and Bloomberg yesterday reported<\/a> Taiwanese authorities have cracked down on GPU smugglers. We\u2019ve also seen suggestions that Chinese companies send storage devices across borders to move data into facilities that have Nvidia kit waiting to run training workloads. \u00ae<\/p>\n","protected":false},"excerpt":{"rendered":"Systems Reveals decent new homegrown accelerator and tiny production volumes Chinese tech giant Alibaba has revealed a new…\n","protected":false},"author":2,"featured_media":597750,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[554,733,4308,86,56,54,55],"class_list":["post-597749","post","type-post","status-publish","format-standard","has-post-thumbnail","category-artificial-intelligence","tag-ai","tag-artificial-intelligence","tag-artificialintelligence","tag-technology","tag-uk","tag-united-kingdom","tag-unitedkingdom"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts\/597749","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/comments?post=597749"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts\/597749\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/media\/597750"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/media?parent=597749"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/categories?post=597749"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/tags?post=597749"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}