A combined memory and silicon shortage is fueling hot demand for Apple's new Macs (up to 0.5TB memory) for local AI processing, while investors are rotating toward memory stocks as the more…

Nvidia is seeding an AI ecosystem through investments and consortia akin to John Malone's TCI strategy, while frontier labs like OpenAI build custom inference chips (Jalapeno) with Broadcom…

The IPO market is gearing up for a truncated post-Labor Day window (6-7 weeks before midterms) with multiple mid-cap ($10-20B) filings expected, which would broaden market participation bey…

Lambda, a neocloud provider with privileged access to GPUs, is raising up to $3B at a $12B+ valuation ahead of an expected IPO next year, highlighting the growing capital intensity and inve…

Gatik focuses on middle-mile autonomous trucking (not robotaxis) with an asset-light model where vehicles stay off its balance sheet; multi-year contracts with PepsiCo and other retailers p…

OpenAI's Jalapeno chip aims to cut inference cost per token by ~1.8x via a novel HBM4+SRAM architecture co-designed with Broadcom and TSMC; the multi-generational roadmap reflects a strateg…

OpenAI's Jalapeno achieves industry-first simultaneous high throughput and ultra-low latency by integrating HBM4 and SRAM in a blank-slate architecture that minimizes data movement; the des…

Richard Ho argues that OpenAI's development of multi-generational custom silicon represents a long-term bet to lower overall AI infrastructure costs, suggesting that in-house chip design co…

Noam Schwartz argues that AI safety must stay ahead of bad actors by ensuring models do only what they're intended to do, highlighting a growing need for safeguards as AI systems become mor…