Susana SchwartzTable of Contents
Taming power spikes in transient AI loads
Greenflash recently closed hybrid tax capital and debt financing for Project Soho – a 400MW / 800MWh standalone battery storage project in Brazoria County, TX – the largest standalone BESS currently under construction in the state. In addition to the innovative financing structure of the project, the scale and capacity of “Project Soho” is unique in not only its massive size, but the innovation of the “energy-ready” infrastructure designed to handle GPU-driven power spikes.
Susana Schwartz
Technology Editor
RCRTech
AI Infrastructure Top Stories
Taming AI workload spikes: Greenflash’s Project Soho requires direct power dispatch into ERCOT. See how FlexGen helps provide the digital backbone for that 5 GW pipeline is FlexGen through modern EMS and “energy-ready” infrastructure.
Viavi on Open RAN: Open RAN provides the cloud-native architecture required for AI workloads, and at RCR Wireless News‘ Intelligent RAN Forum 2026, Viavi examined how openess and network intelligence will bring autonomous networks.
AI Today: What You Need to Know
Anthropic – Nscale $45B deal: Anthropic struck a$45 billion cloud deal with Nscale to rent approximately 460 MWs of compute capacity at an Nscale data center development in West Virgina. Anthropic will leverage Vera Rubin chips at the facility.
Mobileeye raises $220M: Mobileye CEO Amnon Shashua launched his new AI startup, doubleAI, which raised $220 million in total funding. Its product, Maxwell, will act as an autonomous agent to maximize software performance.
Keysight-Broadcom-Alibaba alliance: At OCP Global Summit, the 3 companies are showcasing LLR operation across 200G and 100G SerDes implementations, along with Credit-Based Flow Control at 100G SerDes for Ethernet-based networks.
AI investment MD hired: Goldman Sachs hired former SMBC ABS head Joseph Servi as a managing director for AI infrastructure investment banking with a focus on data centers, GPUs and fiber optics.
Together GPU-hour price: Together AI introduced Together Post-Training, a service that provides GPU allocations (called Model Resources) with a disclosed rental rate starting at $9 per GPU-hour for NVIDIA B200 over first 8 weeks.
Latest RCRtv Interviews
RCR Events
RCR Roundtables AI Infrastructure, October 21st, Dallas, Texas
Join 50 senior data center, energy and AI leaders at the Ritz-Carlton Dallas on October 21 for invitation-only roundtables on powering and scaling AI. Request your invitation
Industrial Wireless Forum, November 10th, Virtual
From private 5G, Wi-Fi and public cellular to edge AI, spectrum, next-generation IoT and sensing, the forum examines the infrastructure choices and practical challenges shaping the future of industrial operations. Register now
AI Infrastructure Forum, December 8th, Virtual
Join RCRTech’s AI Infrastructure Forum to explore how telcos, hyperscalers and data centres can compete or partner as AI shifts to inference, and who captures the value when every token counts. Register now
Industry Resources
RCRTech Trend
Rise of Non-Terrestrial Networks (NTN)
