Who's Who of AI

SemiAnalysis

675 trust @semianalysis_ · 149,292 followers
Why they matter

Tracked through public AI activity and peer connections inside the directory.

AI signals
11
past 30d
Sources
5
distinct domains
Discussions
0
past 30d
Latest signal
4d ago
View every signal from SemiAnalysis →
What they're sharing

Articles & links

Or OpenAI's blog https://t.co/rnSrr68wy0 And attend today's Jalapeño Hot Chips session with Richard, Ravi, and Chris from OpenAI too. (7/7)

openai.com
View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 7 from the directory shared this · 26d ago

Resources mentioned: 🟠 Using group theory to explore the space of positional encodings for attention (blog): https://t.co/ciWX1bLAWu 🟠 Positional Encodings and Group Theory (video): https://t.co/16BvKWPBU2 🟠 RoFormer: https://t.co/ZXMDVyvCf4 (7/7)

arxiv.org
View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 8d ago

Everyone Says Datacenter Moratoriums Are Killing the US Buildout. We disagree. 300+ moratoriums mapped, 20GW sits inside a restricted local boundary, 1,525MW actually slips, 2.3GW nationwide including New York https://t.co/lYwyxj3XE7

Everyone Says Datacenter Moratoriums Are Killing the US Buildout. We Mapped All 300 of Them newsletter.semianalysis.com
AI Weekly's analysis →
  • SemiAnalysis estimates only 2.3 GW of US datacenter capacity is genuinely delayed by moratoriums, against a projected 38 GW of new IT capacity in 2027.
  • Of 20 GW planned inside restricted boundaries, just 1,525 MW is directly blocked at the local level, or 7.6%; New York's permit halt accounts for most of the rest.
  • Michigan leads with 45 enacted moratoriums and Ohio has 40 (83% adopted in 2026), even as voters view datacenters unfavorably by 46% to 29%.
Read full analysis →
View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 2 from the directory shared this · 5d ago

AgentX - InferenceXv3: Does CUDA Moat Hold up in Agentic Inferencing? $3 Million USD dataset open sourced, 1 Mil+ Context Length, Multiturn, Sub Agents 95%+ KVCache HitRate, GB300 NVL72, MI355, B200 https://t.co/PgtKNYXdFr

AgentX - InferenceXv3: Does CUDA Moat Hold up in Agentic Inferencing? newsletter.semianalysis.com
AI Weekly's analysis →
  • SemiAnalysis's AgentX 1.0, built on 393 anonymized Claude Code traces at 1M+ context, cost more than $3M and used ~2MW across 1000+ chips.
  • On Qwen3.5 SGLang the report puts Nvidia at 'over 20x better performance' at 90 tok/s/user; B300 FP4 shows '12x better performance per dollar' vs H100.
  • AMD's ATOM stack shows strong single-GPU kernels but almost no production adoption — only one Alibaba ad unit runs it live, the authors say.
Read full analysis →
View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 2 from the directory shared this · 27d ago

TokenBudgeting: Our Conversations with Enterprises on Token Spend Was Widespread TokenMaxxing Ever Really Here? https://t.co/tgNssBBGVk

TokenBudgeting: Our Conversations with Enterprises on Token Spend newsletter.semianalysis.com
AI Weekly's analysis →
  • Meta employees consumed over 60 trillion tokens in a 30-day window in early 2026, with one individual alone accounting for about 280 billion.
  • Monthly per-employee caps now range from $250 at an aerospace and defense manufacturer to $2,000 at Workday and Stripe, with no cross-industry consensus.
  • Ramp data cited by SemiAnalysis shows 99th percentile customers spend about $90,000 per employee per year while the median customer spends $136.
Read full analysis →
View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 2 from the directory shared this · 82d ago

A Brain Too Big to Carry — On-Device vs Datacenter Inference Robot Models, Silicon & DRAM Efficiency, Jetson Thor vs. B300 TCO, Deployments, The Network Wall https://t.co/FzaOQl22RP

Where Does a Robot Think — On-Device vs Datacenter Inference newsletter.semianalysis.com
AI Weekly's analysis →
  • Boston Dynamics offloads its System 2 planner to Google TPUs; the model runs at hundreds of billions to a trillion parameters, too big for a robot.
  • For 96 robots, aggregate TCO is $14.97/hr on-device with Jetson Thor, $15.61 on RTX 6000 Pro offload, and $18.63 on B300.
  • Sunday Robotics pivoted to on-device inference after home WiFi jitter proved unreliable; its ACT-2 model now reports 99.1% laundry-folding success.
Read full analysis →
View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 6d ago

To learn more, please check our networking model (7/7) https://t.co/kqt0IiSiFi

AI Networking Model semianalysis.com
AI Weekly's analysis →
  • SemiAnalysis is offering device-level tracking of AI cluster networking across five fabric layers, with data running 2023 to 2026.
  • Coverage spans 80+ hyperscaler configuration panels for Microsoft, Google, Meta, Amazon, Oracle, X.AI and neoclouds, tied to specific accelerator SKUs.
  • 25+ suppliers are tracked, including Nvidia, Arista, Broadcom, Cisco, Coherent and Lumentum, across 200G to 1.6T transceiver speeds.
Read full analysis →
View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 7d ago

Nvidia’s Backstop Universe Heads I Win, Tails Who Loses? The $11T AI Buildout, Nvidia’s Backstop Economics, and the Limits of Nvidia’s Balance Sheet https://t.co/cpGvPR0sUf

Nvidia’s Backstop Universe – Heads I Win, Tails Who Loses? newsletter.semianalysis.com
AI Weekly's analysis →
  • Nvidia's 2Q F1/27 10-Q discloses $530B in gross off-balance-sheet guarantees, up from $184B the prior quarter.
  • A single $108.5B line covers SB Energy's PORTS-Pike campus in Ohio: 4.25 GW leased to OpenAI for twenty years.
  • AICP take-or-pay floors are disclosed for Firmus ($21.1B), SharonAI ($4.2B) and an estimated $2.2B for GMI.
Read full analysis →
View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 9d ago

Are you SemiAnalysis? Show it.

Add the Who’s Who of AI badge to your site or bio. It links back to this profile.

Listed in AI Weekly's Who's Who of AI

Markdown: [![Listed in AI Weekly's Who's Who of AI](https://aiweekly.co/modules/custom/aiweekly_whoswho/images/whoswho-badge.svg)](https://aiweekly.co/whos-who/person/semianalysis)