Catherine Sbeglia Nin
Chips, cooling, and capital
AI’s infrastructure era has arrived, demanding chips, cooling, capital — and unprecedented levels of coordination. It’s shocking just how quickly the AI infrastructure ecosystem is scaling, and how much is at stake. First, the rise of “neoclouds:” According to JLL, these upstart providers are growing at an eye-catching 82% CAGR by offering GPU access and cost advantages over hyperscalers. Yet their rapid expansion comes with risks: shorter contracts, heavy capital requirements, and questions about long-term resilience.
On the chip side, Huawei has set out a three-year roadmap for its Ascend processors, with each new series — 950, 960, 970 — doubling compute capacity to sustain AI demand. In parallel, LG and SK are tackling the power and cooling burden, co-developing AI-driven systems to cut emissions and costs while packaging energy efficiency as a global service.
Scale is the common denominator. OpenAI and NVIDIA plan 10 GW of AI data centers backed by up to $100B in investment, while governments worldwide pour billions into chips and compute capacity. Analysts warn the energy required is “biblical.” In a video interview with CNBC, CoreWeave’s CEO put it bluntly: Trillions in public-private funding will be needed. Meanwhile, Microsoft and Amazon are positioned to capture the lion’s share of the U.K.’s $42B AI initiative.
Catherine Sbeglia Nin
Managing Editor
RCRTech
AI Infrastructure Top 3
Neoclouds surge on AI: Neoclouds are expanding at an 82% CAGR, fueled by AI demand and GPU access, with JLL highlighting rapid growth, cost advantages over hyperscalers, and investment risks tied to short contracts and capital intensity.
Huawei AI chip roadmap: Huawei has unveiled a three-year roadmap for its Ascend AI chips at Connect Shanghai, with new 950, 960, and 970 series designed to double compute capacity each cycle and sustain global AI demand.
LG, SK target AI cooling: LG Electronics and SK Innovation will co-develop AI-driven cooling and energy systems for data centers, aiming to cut emissions and costs while expanding globally with integrated Energy-as-a-Service offerings.
AI-Powered Telecom Infrastructure
Supermicro, in collaboration with NVIDIA, delivers AI-powered infrastructure tailored for telcos, enhancing operational efficiency, network management, and customer experiences. Explore now
AI Today: What You Need to Know
Future-proofing AI-ready DCs: AI workloads are reshaping the economics and engineering of data centers, from power and cooling demands to network design and financial planning. Here’s what operators need to evaluate before upgrading.
AI needs trillions: In a video interview with CNBC, CoreWeave CEO Michael Intrator says AI infrastructure will need trillions in public-private investment, comparing the surge to the dot-com era.
AI adoption uneven: A PwC survey of 500-plus manufacturing executives has found most see AI as a competitive edge, though adoption is uneven, amid wider shifts in workforce, reshoring, realignment, and product strategies.
MediaTek’s Dimensity 9500: MediaTek’s Dimensity 9500 flagship chip boosts on-device AI, gaming, and power efficiency. It delivers 32% higher single-core, 17% higher multi-core performance, and up to 30% lower power consumption under load.
Mega-trends in AI testing: Spirent flags two AI testing mega-trends: embedding AI inside network hardware and building networks to support AI workloads. Enabled by digital twins, synthetic data, and continuous live network testing.
NVIDIA backs OpenAI: OpenAI and NVIDIA agree to deploy 10 GW of NVIDIA systems for next-gen AI infrastructure, with NVIDIA investing up to $100 billion. First gigawatt launches in late 2026 on Vera Rubin platform.
AI’s “biblical” energy demands: Governments are pouring billions into chips and AI infrastructure, fueling an arms race that demands “biblical amounts of energy” as compute power becomes a measure of national strength.
Millions for enterprise AI infra: Invisible Technologies raised $100M led by new firm Vanara Capital — total funding now $144M. The enterprise AI infrastructure platform is scaling fast across data, workflow, and automation solution layers.
Microsoft, Amazon poised for windfall: The U.K.’s $42 billion AI infrastructure initiative is expected to benefit Microsoft and Amazon most, as both firms possess cloud, AI, and platform strengths aligned with national strategy.
AI in Marines Corps: The Marine Corps is launching training to upskill all Marines on AI, with digital transformation teams driving service-wide adoption, AI lead Capt. Christopher Clark announced at FedTalks 2025.
