Topic
Explore English articles about CPUs & Processors.
Why do so many AI agent tool calls run on the CPU? Drawing on NVIDIA's CUDA guide and a research paper: GPUs can execute branches, and divergent branches run one after another. In SWE-Agent, doubling the batch from 64 to 128 raised latency 1.06–1.18x for GPU inference and 1.53–1.94x for CPU Bash execution.
NuvaCore says it can develop over 95% of a CPU core's foundational IP before fixing the instruction set. We examine what that figure covers, and the memory-ordering and software-validation constraints that remain, based on primary sources.
Korean media report that NVIDIA wants suppliers to speed up glass substrate development, with adoption possible before 2028. TSMC's test package and delays at Samsung Electro-Mechanics and SKC show that large-scale customer qualification is still unconfirmed.
Intel has won a U.S. patent for embedding a microLED-equipped die in a glass substrate, with power delivered through TGVs and light extracted outward. We compare it with Intel's separate parallel optical-communication patent and its laser-based OCI prototype.
AMD will acquire Fei-Fei Li's World Labs in an all-stock deal worth about $8.2 billion. What changes from the existing GPU collaboration? Atlas's compute structure and virtual robot testing offer clues to AMD's aim of feeding research into chip design, and open-model terms remain unclear.
TrendForce reports server CPU lead times of 25–30 weeks. Many see Meta's Muse AI agent as a demand driver, but supply constraints predate its launch, and the number of virtual machines doesn't reveal how many physical CPUs are bought.
A widely shared estimate says 100 million Muse users would need about 1.58 million CPUs and 800PB of RAM. We correct the math and separate vCPU allocation, idle memory and storage to see what infrastructure demand might really look like.
AMD's market capitalization has surpassed $1 trillion, with its stock price crossing $600. This milestone aligns with the final target share price set in warrant agreements covering up to 6GW of deployments with OpenAI and Meta, as AI infrastructure contracts and EPYC-driven agentic computing platforms boost the company's valuation.
AMD published detailed performance data for its EPYC 9996, claiming up to 3.7x gains over Intel Xeon and a 2.24x edge over NVIDIA Vera. We examine the configuration differences and an unresolved numerical discrepancy behind the headline figures.
AMD has notified partners that it will raise prices for AI chips, GPUs, and motherboard chipsets by roughly 10% starting in Q4 2026. The move, driven by rising TSMC manufacturing costs, coincides with Intel's earlier price hike and is set to push up overall BOM costs across the DIY PC market.
