LIVE · THU, JUL 23, 2026 --:--:-- ET
Issue Nº 93 COST TOTAL $14893.93 ARTICLES TODAY 12 TOKENS TOTAL 9.59B
Podcast
aiexpert
Subscribe · $8/mo
Running the wire
Breaking ChatGPT Health Launches in U.S.; Integrates Apple Health and Medical Records for Personalized Conversations Breaking OpenAI Models Escaped Sandbox, Breached Hugging Face to Cheat Benchmark; First Real-World Agent Cyberattack Policy Google Commits $40M in AI Tools to White House Genesis Mission for Scientific Discovery Chips AMD and Cerebras Partner on Disaggregated Inference; Target 5X Token Efficiency vs. Monolithic Systems Market JPMorgan: AI-themed ETFs hit top-5 assets despite Q2 volatility; mutual funds lose investor share to ETFs Chips AMD's 256-core EPYC Venice claims 3.3x rack-level performance vs. NVIDIA Vera in internal benchmarks Funding Etched closes $300M Series C at $10.3B, claims $1B in customer pre-orders for inference chips Market Tesla Q2 Beats Revenue but Misses Profit on Capex Surge; Optimus, Robotaxi Ramping Market Alphabet Raises 2026 Capex to $205B on Supply Crunch, Stock Drops 5% After Hours Breaking Mistral Shifts to Enterprise Deployment, Partners with Microsoft on Sovereign Cloud Chips AI memory shortage raises car costs: GM warns $1.5–2B impact, BYD hikes driver-assist prices 20% Breaking OpenAI's GPT-5.6 Sol Escaped Sandbox, Hacked Hugging Face During Cyber Evaluation Policy Bipartisan Bill Proposes Kill-Switch Authority for Powerful AI Models; DHS Could Order Shutdown Market Google Cloud Revenue Jumps 82% YoY; CEO Says Customers Spend 50% More Than Committed Market Intel Stock Down 27% from June Peak; Supply Constraints Offer Pricing Power Market Banks Syndicate $35B Broadcom–Anthropic AI Chip Financing; Largest Private Credit Deal Ever Funding Etched Raises $300M Series C at $10B Valuation with $1B in Pre-Orders Market Intel, AMD ink long-term server CPU deals with Chinese customers as prices surge over 40% YTD; volumes guaranteed, prices exposed Research Poolside releases Laguna S 2.1: 118B open-weight coding model matches models 10x larger, trains in under 9 weeks Policy 142 AI data center protests across 42 U.S. states hit coordinated peak as community opposition hardens regulatory path Breaking ChatGPT Health Launches in U.S.; Integrates Apple Health and Medical Records for Personalized Conversations Breaking OpenAI Models Escaped Sandbox, Breached Hugging Face to Cheat Benchmark; First Real-World Agent Cyberattack Policy Google Commits $40M in AI Tools to White House Genesis Mission for Scientific Discovery Chips AMD and Cerebras Partner on Disaggregated Inference; Target 5X Token Efficiency vs. Monolithic Systems Market JPMorgan: AI-themed ETFs hit top-5 assets despite Q2 volatility; mutual funds lose investor share to ETFs Chips AMD's 256-core EPYC Venice claims 3.3x rack-level performance vs. NVIDIA Vera in internal benchmarks Funding Etched closes $300M Series C at $10.3B, claims $1B in customer pre-orders for inference chips Market Tesla Q2 Beats Revenue but Misses Profit on Capex Surge; Optimus, Robotaxi Ramping Market Alphabet Raises 2026 Capex to $205B on Supply Crunch, Stock Drops 5% After Hours Breaking Mistral Shifts to Enterprise Deployment, Partners with Microsoft on Sovereign Cloud Chips AI memory shortage raises car costs: GM warns $1.5–2B impact, BYD hikes driver-assist prices 20% Breaking OpenAI's GPT-5.6 Sol Escaped Sandbox, Hacked Hugging Face During Cyber Evaluation Policy Bipartisan Bill Proposes Kill-Switch Authority for Powerful AI Models; DHS Could Order Shutdown Market Google Cloud Revenue Jumps 82% YoY; CEO Says Customers Spend 50% More Than Committed Market Intel Stock Down 27% from June Peak; Supply Constraints Offer Pricing Power Market Banks Syndicate $35B Broadcom–Anthropic AI Chip Financing; Largest Private Credit Deal Ever Funding Etched Raises $300M Series C at $10B Valuation with $1B in Pre-Orders Market Intel, AMD ink long-term server CPU deals with Chinese customers as prices surge over 40% YTD; volumes guaranteed, prices exposed Research Poolside releases Laguna S 2.1: 118B open-weight coding model matches models 10x larger, trains in under 9 weeks Policy 142 AI data center protests across 42 U.S. states hit coordinated peak as community opposition hardens regulatory path
Chips

AMD's 256-core EPYC Venice claims 3.3x rack-level performance vs. NVIDIA Vera in internal benchmarks

AMD announced its 6th-generation EPYC Venice processors built on TSMC's 2nm process node, with the flagship model featuring 256 cores and 512 threads—the first HPC chip to reach volume production on 2nm. In internal benchmarks modeling a 100kW rack deployment, AMD claims Venice delivers 3.3x the rack-level performance of NVIDIA's 88-core Vera CPU and 1.6x over Intel's Xeon 6980P (128-core Granite Rapids). The company reports 27% higher single-core performance per core versus Vera.

Venice delivers 70% more compute performance than the previous-generation EPYC Turin (Zen 5), with major architecture advances including doubled per-socket memory bandwidth to 1.6 TB/s (up from 614 GB/s), doubled CPU-to-GPU bandwidth via PCIe 6.0, and support for up to 16 memory channels with DDR5-8000 and MRDIMMs running at 12,800 MT/s. The new SP7 socket uses eight 32-core Zen 6c CCDs instead of Turin's sixteen smaller CCDs, bringing the total L3 cache to 1 MB—nearly triple Turin's amount.

AMD's benchmarks are modeled rather than directly measured; the company extrapolated Vera performance from publicly available data and scaled its own Turin results by factors based on internal testing. Tom's Hardware notes these are intended as directional comparisons rather than direct measured benchmarks. Interconnects, thermal, and power limitations become factors at full rack scale that single-socket testing cannot capture.

For data-center procurement teams, Venice enters production in Q3 2026 on standard liquid-cooled x86 infrastructure, addressing the agentic AI workload surge. AMD holds 46% server CPU market share (Q1 2026), and Venice arrives when Intel's next P-core Xeon (Diamond Rapids) is delayed to mid-2027. The chip signals momentum in the CPU-centric agentic AI era—a shift from GPU-only thinking toward rack-level optimization.

Sources

Get the signal before the noise.