cd/sources/semianalysis· home› sources› SemiAnalysis
cat /sources/semianalysis.feed | wc -l → 48

SemiAnalysis

articles 48 domain semianalysis.com → page 1/3 feed RSS
05:55
2026-11-02
semianalysis.com
artificial-intelligence

GDPx: The General Agentic Benchmark

A benchmark called GDPx v1.0 was published to measure frontier model capabilities across 11 domains, built with 14 partners and 206 tasks. The GDPx results are presented on a cost-versus-performance P…

19:06
2026-10-23
semianalysis.com
ai-chips

Amazon’s Trainium Takes a Strike at the Training King

Amazon's AWS Neuron team has rebuilt Trainium's software stack around a native PyTorch experience, moving Trainium3 off the PyTorch/XLA approach and the xm.step() calls that previously limited the chi…

17:34
2026-10-21
semianalysis.com
ai-infrastructure

The Huawei-DeepSeek Symbiosis

DeepSeek released Ascend support for six of its software repositories — TileLang, DeepGEMM, DeepEP, TileKernels, FlashMLA and DeepSelect — on September 30, ahead of China's national day holidays, acco…

11:33
2026-10-20
semianalysis.com
ai-infrastructure

Inference Proft Estimator

Artificial Analysis's Inference tool now includes a beta Revenue & Profit Estimator per GW that models token pricing, cache hit rates, P90 interactivity targets and chip configurations, with Kimi K3 2…

13:36
2026-09-26
semianalysis.com
ai-chips

Intel Panther Lake Teardown

SemiAnalysis's STEEL teardown lab found that Intel's Panther Lake, the first commercial chip to use backside power delivery (PowerVia) and Intel's first gate-all-around RibbonFET transistors on the 18…

18:14
2026-09-21
semianalysis.com
ai-infrastructure

Computation and Data Movement for Inference

Mixture of Experts has changed the structure of AI inference serving by altering which tensors are active per token, what must remain close together, which transfers need strong local bandwidth, and h…

18:19
2026-09-13
semianalysis.com
ai-chips

Long Live the Short King: Why 4-hi HBM Wins

Nvidia's next-generation Rubin Ultra accelerator will ship with 192GB of HBM per GPU, down from 288GB in standard Rubin and B300, according to SemiAnalysis, which first reported the change. SemiAnalys…

20:53
2026-09-09
semianalysis.com
artificial-intelligence

Where Does a Robot Think – On-Device vs Datacenter Inference

Physical AI is moving AI beyond screens into robots, but the field is split on where inference should run: on-device versus in datacenters. Figure runs its Helix model entirely onboard, while Physical…

20:00
2026-09-07
semianalysis.com
ai-infrastructure

TPU Inference Externalization Full Steam Ahead - InferenceX

SemiAnalysis published the first third-party inference results for Google's TPUv7 Ironwood, showing up to 50% better performance per dollar than NVIDIA's B200/B300 in apples-to-apples comparisons. The…

20:14
2026-09-01
semianalysis.com
artificial-intelligence

Korea’s Trillion-Dollar Sovereign AI Investment

South Korea's government launched the 'Independent AI Foundation Model' project in June 2025, a tournament-style initiative to develop a domestic frontier AI model, with five consortiums selected in A…

15:46
2026-08-30
semianalysis.com
ai-safety

Most Neoclouds Suck At Security

SemiAnalysis's ClusterMAX 3.0 testing found that most neoclouds have serious security vulnerabilities, with five frightening patterns identified. The company urges neocloud operators and users to upda…

14:00
2026-08-25
semianalysis.com
ai-chips

OpenAI Jalapeño: Better Than Nvidia Blackwell

OpenAI's new inference chip, Jalapeño, outperforms Nvidia's Blackwell and other competitors in performance per watt across almost all scenarios, according to benchmarks run by SemiAnalysis with OpenAI…

16:40
2026-08-21
semianalysis.com
artificial-intelligence

Are Open Models Catching Up?

Open-source AI models are catching up to closed frontier models twice as fast with each generation, according to a SemiAnalysis analysis that found open models now match closed models on coding and ag…

page 1 / 3 next →