[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fVerVmyLoe46P6cj-AprnuNu3eN-bmdVc-CVEhaNrUaA":3},{"article":4,"related":19},{"id":5,"slug":6,"title":7,"seo_title":7,"description":8,"keywords":9,"content":10,"category":11,"image_url":12,"source_guid":13,"published_at":14,"created_at":15,"updated_at":16,"source_url":17,"source_name":18},1332,"deepseek-orders-160000-huawei-chips-for-inference-cluster","DeepSeek planned 160,000-chip Huawei inference cluster","DeepSeek reportedly planned a Huawei inference cluster in Inner Mongolia. The proposal points to workload-specific sourcing, with delivery still uncertain.","[\"DeepSeek\",\"Huawei\",\"Ascend-950DT\",\"AI inference\",\"Nvidia\",\"chip supply\"]","\u003Cp>DeepSeek planned to deploy at least 160,000 Huawei Ascend-950DT processors at a data center in Inner Mongolia, \u003Ca href=\"https:\u002F\u002Fthe-decoder.com\u002Fdeepseek-plans-the-largest-known-huawei-chip-cluster-with-160000-processors-in-inner-mongolia\u002F\" rel=\"noopener noreferrer\">the-decoder.com reported\u003C\u002Fa> on September 4, 2026, citing Bloomberg. The proposed cluster would serve inference workloads. DeepSeek continued to use Nvidia hardware for training, according to the report.\u003C\u002Fp>\n\u003Cp>The account described a prospective deployment, not an operating cluster. It said production constraints and memory shortages would probably prevent Huawei from supplying the full order for more than a year. Those specifications and delivery expectations were reported claims, not independently verified findings.\u003C\u002Fp>\n\u003Cp>The useful distinction for AI infrastructure buyers is between replacing a supplier for one workload and replacing it throughout a development process. This proposal suggests DeepSeek was pursuing the former: Huawei for the planned inference capacity, with Nvidia still used for training. That split does not establish that Huawei cannot support training, nor does it demonstrate that the proposed inference system would meet DeepSeek’s needs.\u003C\u002Fp>\n\u003Cp>For a team considering a similar sourcing decision, the missing evidence concerns usable capacity. A processor count alone does not establish throughput, operating cost or the work required to run a particular model. The supplied report provides none of those comparisons. The proposal therefore supports a narrower conclusion about DeepSeek’s intended procurement direction, while leaving the practical case for other buyers unresolved.\u003C\u002Fp>\n\u003Cp>\u003Cem>Correction, September 19, 2026: This article was revised to attribute reported claims, remove unsupported statements and clarify the limits of the available evidence.\u003C\u002Fem>\u003C\u002Fp>","Gadgets & Hardware","https:\u002F\u002Fseedwire.co\u002Fapi\u002Fimages\u002Farticles\u002F1788567170424-286i84driov.png","26afe5085e8df54b12698e58a238ffdc4445535a3203d480a07b85f072ea33dc","2026-09-04T14:19:14.000Z","2026-09-05T00:12:51.196Z","2026-09-19T19:28:34.469Z","https:\u002F\u002Fthe-decoder.com\u002Fdeepseek-plans-the-largest-known-huawei-chip-cluster-with-160000-processors-in-inner-mongolia\u002F","the-decoder.com",[20,27],{"id":21,"slug":22,"title":23,"description":24,"category":11,"image_url":25,"published_at":26},1334,"agility-unveils-digit-5-with-a-claim-of-fence-free-work","Agility unveils Digit 5 with a claim of fence-free work","Agility says Digit 5 can work beside people without safety fences. Its warehouse record offers context, but the new model's safety claims need closer scrutiny.",null,"2026-09-15T12:57:10.000Z",{"id":28,"slug":29,"title":30,"description":31,"category":11,"image_url":32,"published_at":33},1322,"openais-jalapeo-chip-shows-fast-inference-at-scale","OpenAI reports Jalapeño inference gains, with rollout to follow","OpenAI reported faster, more power-efficient inference with Jalapeño. Its deployment timeline and missing benchmark detail limit what AI buyers can conclude.","https:\u002F\u002Fseedwire.co\u002Fapi\u002Fimages\u002Farticles\u002F1787702487045-retumd9ylw.png","2026-08-25T14:22:04.000Z"]