Compute & Chips

Anthropic has committed to spend $11.6 billion on Akamai cloud infrastructure over seven years, with the arrangement potentially growing to about $20 billion. Akamai is also offering Anthropic a potential stake of up to roughly 5%, tied to spending milestones.

Google will send four TPUs into orbit aboard a refrigerator-sized experimental satellite to test its Suncatcher concept. Limited thermal capacity means the chips can run for only about 15 minutes at a time.

Qualcomm has introduced the Snapdragon 8 Elite Gen 6 and Elite Extreme Gen 6, with a focus on local AI agents, voice interaction, personalization, and camera processing. The Extreme model is said to run a 30-billion-parameter mixture-of-experts model on-device.

Huawei says its next-generation Ascend 960DT AI chip is expected to be ready in the first quarter of 2027, earlier than its previous third-quarter target. The company is also building large AI systems that link accelerator cards through its own interconnect architecture.

Apple is reportedly developing an enterprise AI server using two or four future M8 Ultra chips. If it ships around 2029, the system would mark the company’s return to the server market after the Xserve era.

PyTorch researchers extended FlashAttention-4 with MXFP8 forward and backward passes on Blackwell, combining block-scaled MMA with fused quantization and jagged-input processing.

Arm’s Neoverse CSS N4 supports up to 128 cores per die and adds LPDDR6 and PCIe 7.0. The launch reflects a broader shift in AI infrastructure, where CPUs, memory and I/O must keep pace with accelerators.

Nvidia is investing $3.5 billion in MediaTek and bringing the Taiwanese chipmaker into its NVLink Fusion ecosystem. The move suggests Nvidia is prepared to accommodate custom silicon as long as it remains central to the surrounding AI infrastructure.

As AI clusters scale toward gigawatt-level deployments, performance depends on more than adding processors. Nvidia is extending its advantage into the orchestration of data, memory, storage and networking around the GPU.

Apple has refreshed the Mac mini and Mac Studio with new M6 and M5 Ultra chips. Their memory capacity and high-bandwidth interconnects position them as increasingly capable platforms for local model development and inference.

As GPUs become rentable production assets rather than scarce hardware, compute price volatility is creating demand for financial hedging. Moves by CME, Silicon Data and the CFTC suggest that GPU futures are moving toward a market infrastructure role.

As spending on GPUs and data centers continues to rise, compute is becoming a major cost that businesses may need to hedge. Silicon Data wants to create a reference price for GPU rentals and bring compute futures to Wall Street.

AMD plans to acquire Toronto-based AI chip startup Taalas, whose unusual approach is to place model weights directly in silicon rather than in HBM or other memory. The deal points to a broader race to improve inference efficiency beyond conventional GPU scaling.

Anthropic has confirmed that it is hiring a custom silicon team to explore chips for running its AI models. The company says it will still use a multi-chip approach rather than relying solely on its own hardware.

AMD’s new MI455X accelerator and Helios rack system show that the company is no longer competing only at the GPU-card level. It is trying to sell a full AI infrastructure story built around advanced chiplets, HBM4 memory and open interconnects.

A routine grid fault near Washington, DC became a warning sign when gigawatts of data center load disappeared almost at once. The incident shows how AI infrastructure can amplify, not merely consume, power system stress.

AMD is positioning Helios as a rack-scale AI system for the world’s largest model builders and cloud data centers. With deployment plans from Microsoft, Anthropic, OpenAI and others, the company is trying to compete with Nvidia at the system level, not just the chip level.

AI chip startup Etched has raised a $300 million Series C at a $10.3 billion valuation. The company is betting that purpose-built inference systems can reduce dependence on GPUs while running a wide range of model architectures.

Qiymor used WAIC 2026 to showcase a full-stack interconnect plan for domestic supernodes, an IBGDA demo linking a domestic GPU to a domestic RDMA NIC, and its roadmap toward optical interconnects. The message is clear: China’s AI stack is moving from chip-level progress to system-level integration.

At WAIC 2026, Arm China presented edge AI as a systems challenge rather than a raw-compute race. Its roadmap spans embedded CPUs, configurable NPUs, machine-oriented video processing and an open AIOS initiative.

Intellifusion outlined a new inference-chip roadmap built around three DeepVerse processors optimized for Prefill, Decode and Decode FFN workloads. The strategy points to a broader industry shift toward heterogeneous clusters and lower token-generation costs.