Introducing Amazon SageMaker HyperPod Inference Gateway17m ago The 'Godfather of AI' backs a new watchdog plan to track OpenAI and Anthropic's AI risks from the inside26m ago Huawei says AI will burn 100,000 times more tokens by 203532m ago AI stock swings could affect your 401(k). Here’s what retirement savers need to know34m ago OpenAI is offering engineers up to $500,000 in salary as it pushes into robotics37m ago Huawei’s chairman says Chinese AI can’t yet see the risks it should fear38m ago How Can Robotaxis Handle Demand Surges?41m ago Open-weight models now handle a majority of tokens on Vercel’s AI Gateway. But Anthropic still takes 64% of the spend.43m ago Meta's personal AI agent Muse soars to the top of the App Store47m ago Oracle exec tells workers that its own AI rollout didn't go so smoothly51m ago Starting a Career in Data Science in the Age of AI56m ago AI Governance Will Separate Tomorrow's Market Leaders From Everyone Else56m ago Brookfield’s Flatt Says AI ‘Slowing Down Anyway’ as Builders Can’t Keep Up57m ago Context and Coordination: The Missing Layer in Enterprise Automation57m ago

Introducing Amazon SageMaker HyperPod Inference Gateway17m ago The 'Godfather of AI' backs a new watchdog plan to track OpenAI and Anthropic's AI risks from the inside26m ago Huawei says AI will burn 100,000 times more tokens by 203532m ago AI stock swings could affect your 401(k). Here’s what retirement savers need to know34m ago OpenAI is offering engineers up to $500,000 in salary as it pushes into robotics37m ago Huawei’s chairman says Chinese AI can’t yet see the risks it should fear38m ago How Can Robotaxis Handle Demand Surges?41m ago Open-weight models now handle a majority of tokens on Vercel’s AI Gateway. But Anthropic still takes 64% of the spend.43m ago Meta's personal AI agent Muse soars to the top of the App Store47m ago Oracle exec tells workers that its own AI rollout didn't go so smoothly51m ago Starting a Career in Data Science in the Age of AI56m ago AI Governance Will Separate Tomorrow's Market Leaders From Everyone Else56m ago Brookfield’s Flatt Says AI ‘Slowing Down Anyway’ as Builders Can’t Keep Up57m ago Context and Coordination: The Missing Layer in Enterprise Automation57m ago

Chips & Compute Machine Learning5 min reading time

Huawei’s answer to Nvidia isn’t a faster chip, it’s a much bigger machine

The Next WebPublished 2 hours ago

Huawei announced its Ascend 960DT AI processor arriving in early 2027, alongside a new Atlas 960E SuperPoD system using near-packaged optics for high efficiency and reliability. The system integrates 4,096 NPUs delivering 8 EFLOPS at FP8 precision, emphasizing optical interconnects over raw chip speed.

#ai-hardware #ascend-960dt #atlas-960e-superpod #high-performance-computing #huawei #optical-internetworking-forum

More in Chips & Compute

Chips & Compute

The Startup That Built OpenAI’s Biggest Data Center Is Now Making Tiny Ones

Covered by 4 sources1d ago

Chips & Compute

Huawei Accelerates Launch of New AI Chip to Take On Nvidia

Covered by 2 sources2d ago

Chips & Compute

SoftBank Adds $21 Billion to AI Firepower in New Debt Deals

Bloomberg8h ago