Particle.news
Download on the App Store

AI Push Reorders Finance and Big Tech as Banks Tout Gains, Meta Plans Closed Model

Companies are investing heavily to build capability before expecting efficiency-led payoffs.

Overview

  • U.S. bank leaders reported measurable productivity gains from AI, with JPMorgan citing a rise to 6% from 3% and executives flagging potential role reductions in some functions.
  • Goldman Sachs’ internal OneGS 3.0 memo, reported by Reuters, signaled potential year-end layoffs and slower hiring tied to AI-driven efficiency plans.
  • A new Bloomberg Intelligence survey found over 70% of financial firms expect higher operating costs over the next three years and about two-thirds foresee initial headcount increases before efficiencies materialize around 2027–2028.
  • Bloomberg reported that Meta is pivoting to a closed-source commercial model codenamed Avocado, said to distill capabilities from Google’s Gemma, OpenAI’s gpt-oss and Alibaba’s Qwen, with leadership changes and a target launch in early 2026.
  • SpaceX and startup Starcloud placed Nvidia H100 GPUs in orbit and demonstrated in-space AI by running Google’s Gemma and training Andrej Karpathy’s NanoGPT, with plans to add more GPUs, including Blackwell, in a 2026 mission.