Susana Schwartz
Meta layoffs anticipated as it inks deal with Nebius
Today RCRTech takes a closer look at Meta’s increasingly aggressive AI-forward strategy, not only with yesterday’s Nebius deal, but other decisions that show Meta’s willingness to diversify to win the race for AI infrastructure resources. It’s just the latest example of how diversification is going to be critical amid increasingly limited supplies of …well, everything: high-end GPUs; power (equipment, transformers, you name it); specialized memory (HBM, KV Cache, Persistent Agent Memory); labor – high-level tech specialists as well skilled trades like builders, plumbers, mechanics, and electricians. Meta isn’t alone, as Amazon, Google and Microsoft are all starting to up the ante with more open, diversified, multi-layered approaches that involve a blend of their own and external solutions for every layer of the AI tech stack. And, like we wrote not long ago, the greatest achievement in AI could be “frenemy” collaboration among previously unlikely partners, as well as more openness overall and flexibility for the sake of adaptability and innovation. Check out this and other stories, below.
Susana Schwartz
Technology Editor
RCRTech
AI Infrastructure Top Stories
Meta doubles-down on AI: The company continues its high-stakes pivot toward AI infrastructure, doing whatever it takes to build out or acquire data centers, silicon, and software – both reducing its reliance and diversifying its reliance on others.
Iran War halts subsea construction: The 2Africa system will be the largest subsea cable network ever built, spanning 45,000 km, but contractors on the project issued a force majeure notice to stakeholders because of the dangers in the Persian Gulf.
AI-driven autonomy for network operators: AI is moving deeper into the network stack, and operators are experimenting with AI-driven service assurance that automates processes and shifts power from service providers to themselves.
AI Today: What You Need to Know
Nvidia CEO predicts $1T backlog: Nvidia CEO at annual developers conference said he expects to make $1 trillion with AI chips, and heralds ‘inference inflection’ as next phase of AI in which he believes Nvidia will continue to be “indispensable.”
SK Group Memory Warning: SK Group’s Chairman Chey Tae-won said the global memory chip shortage is likely to continue for several more years, potentially lasting through 2030. He spoke on the sidelines of Nvidia GTC 2026 in San Jose, CA.
Nvidia & T-Mobile Partner: Nvidia, T-Mobile, and partners are integrating physical AI applications on AI-RAN-ready infrastructure, utilizing Blackwell server-edition technology to push reasoning and vision AI to the edge.
Potential 8 GW in dedicated AI power: Gov. Morrisey acknowledged Nscale will acquire American Intelligence & Power Corporation, which he attributed in part to his recent Microgrid legislation along with the state’s 50 by 50 energy strategy.
Google seeks cooling solutions: Google is in talks with suppliers like China’s Envicool to acquire data center cooling systems, emphasizing the need for cooling hardware in a tight global supply of components for thermal management solutions.
Alibaba consolidating AI business: Alibaba will bring its AI operations under the “Alibaba Token Hub,“ which will integrate behind the Qwen models with consumer-facing applications, as well as DingTalk and Quark-branded hardware.
