Radiant

Introducing Radiant. By merging the software and compute capabilities of Ori with the long duration capital and powered-land assets of @Brookfield they created an end-to-end, fully integrated AI infrastructure platform. From shovel to model. Welcome to the age of abundant AI. radiant.co/blog/the-age-o…

We took NVIDIA’s newest open model, Nemotron 3 Super, for a spin on Radiant with vLLM, and what makes it compelling is the problem it is built to solve. As agentic systems scale, teams face a tradeoff. Larger models improve reasoning but are costly to run, while smaller models are efficient yet struggle with complex tasks. At the same time, multi agent workflows sharply increase token volume, amplifying cost pressures. Nemotron 3 Super offers a credible answer. It is a 120B parameter model with only 12B active parameters at a time, enabling long context and deeper reasoning without the typical runtime burden at this scale. In this tutorial, we walk through the steps to run Nemotron Super on a cloud GPU and see how it holds up: hubs.la/Q048012Y0

A year ago, most AI conversations focused on models. Now, more of the discussion is about what sits beneath them. Energy, land, cloud infrastructure, and capital are shaping what can be built, how quickly, and at what cost. NVIDIA’s description of AI as a 5-layer cake captures this transition clearly, framing AI as a full stack that runs from energy and compute all the way to applications. In this article, we unpack how Radiant translates those layers into a deployable system, connecting powered land, compute, capital and software into infrastructure that can scale in the real world. Give it a read: hubs.la/Q047NmvQ0

A first principles view of building sovereign AI 🏗️: Radiant's Chris Galloway debunking Sovereign AI myths and emphasizing the value of a holistic approach towards strategic independence

Starting in a few minutes! Don't miss this insightful session if you're at NVIDIA #GTC. Head over to SJCC 230C (L2) to catch Radiant's CTO Chris Galloway speak about building the right foundation for Sovereign AI.

Truly sovereign AI goes beyond data residency and vendor choice. Today at NVIDIA GTC, our CTO Chris Galloway will speak about Radiant’s multi-dimensional framework for AI sovereignty, spanning jurisdiction, architecture, operations, and control. Drawing on real-world experience designing large-scale AI infrastructure, the session explores how design decisions and control planes shape autonomy, resilience, and adaptability over the full lifecycle of an AI system. 📆 March 19 | 11:00 AM PT 📍 SJCC 230C (L2), San Jose McEnery Convention Center This is a foundational session for anyone building AI systems that need to operate with control and strategic independence. Add it to your GTC schedule hubs.la/Q047zk2Q0 #SovereignAI #AIInfrastructure #NVIDIAGTC

Keynote is rocking and the floor is starting buzz. Can't wait for this thing to pop... #GTC #GTC2026 @nvidia Come see how we build #AIFactories from the capital stack to the software stack. Booth 1901.

The global AI race is hitting a physical wall: the power grid. ⚡️Morgan Stanley is projecting a massive 49 GW shortfall in U.S. data center capacity by 2028, which makes the traditional 3-to-4-year timeline for building AI factories unviable. So, how do we fix the infrastructure bottleneck? In a great new piece, Manesh Patel outlines how Radiant is fundamentally rewriting the deployment playbook, compressing timelines from years to a few quarters. Instead of the flawed, sequential approach of waiting on grid interconnects and piecing together hardware, Radiant is leveraging a vertically integrated utility model to parallelize delivery: 💡 Power First: Pre-secured powered land and behind-the-meter generation to bypass grid queues. 💰 Structural Advantage: Infrastructure-grade capital to absorb upfront costs and deliver predictable pricing. ❄️ High-Density Design: Purpose-built cooling architecture designed for frontier GPUs such as NVIDIA Rubin and Blackwell Ultra. 🧠 Software Intelligence: In-house orchestration achieving maximum GPU utilization. We can't afford to wait for the grid. For nations, enterprises and telcos building AI today, infrastructure deployment velocity is the ultimate strategic advantage. Highly recommend reading the full breakdown here: hubs.la/Q046DNzP0 #AIFactories #DataCenters #AIInfrastructure

It’s easy for AI models to look fast in a vacuum, but low-concurrency benchmarks rarely reflect production reality. To see how they behave under sustained load, we evaluated gpt-oss, Gemma 3, and Llama 4 on Radiant Inference Endpoints. Rather than looking for a theoretical winner, we focused on saturation points and scaling behavior. The data revealed three clear insights: 📊 Throughput numbers can be deceiving: Models can appear highly performant at low concurrency, yet collapse once saturation is reached. 🚦 Concurrency has a breakpoint: Beyond this point, adding users just leads to request loss rather than higher throughput. 🏗️ Architecture dictates how you scale: Single‑GPU models deliver excellent efficiency, until they don’t. Multi‑GPU MoE models trade that peak efficiency for much higher concurrency tolerance. Model selection is only part of the equation; your serving stack is critical. We built the Radiant Inference Delivery Network (IDN) to solve these exact challenges, using global routing and automatic scaling to maintain stability as demand grows. Read the full benchmark data here: hubs.la/Q045__KJ0

Superb post by Rod Gilmore of @YahooFinance about @Radiant_Infra President Mahdi Yahya's journey from teenage coder to telecom startup then from theater school to Ori. It is a great read and well worth your time: uk.finance.yahoo.com/news/ai-uk-com…

As AI factories approach gigawatt capacities, there is no room for error when designing and building them. The traditional “move fast and break things” approach does not work when we're talking about one of the biggest infrastructure build-outs in history. This is precisely why NVIDIA’s Omniverse DSX is emerging as a game changer for the next generation of AI factories. Omniverse DSX provides an infrastructure blueprint that leverages digital-twin technology to eliminate design flaws and de-risk billion-dollar capital investments. It also brings economic value to the forefront by optimizing for tokens-per-watt rather than relying on abstract power-efficiency metrics. Read our deep-dive on Omniverse DSX to see why we think it will transform how AI factories will be designed in the coming months: hubs.ly/Q045FZsW0

Excited to share that SemiAnalysis led the technical due diligence for Brookfield's acquisition of Ori, now merged into Radiant, Brookfield's vertically integrated AI infrastructure platform. Sovereign AI compute is becoming a key theme in global infrastructure investment. Governments and enterprises are increasingly seeking controlled computing environments where data remains within national borders, and the capital flowing into this space reflects that. Brookfield's $100B AI infrastructure investment program is one of the largest vehicles targeting this opportunity. Congrats to Mahdi Yahya and the Ori team, and to our partners at @Brookfield! bloomberg.com/news/articles/…

We are excited to announce that portfolio company, Ori, the distributed AI infrastructure and edge computing company, has been acquired by @Brookfield to be merged with @Radiant_Infra, as they continue to expand in Saudi Arabia. This acquisition and merger highlights the growing strategic demand for AI infrastructure solutions and the innovation in this segment and the region. Congratulations to Ori team! يسعدنا أن نعلن أن شركتنا ضمن المحفظة، “أوري”، التي تقدم حلول البنية التحتية السحابية للذكاء الاصطناعي، قد تم الاستحواذ عليها من قبل بروكفيلد لإدارة الأصول، لدمجها مع شركة راديانت. معززًا توسّع الشركة بالمملكة العربية السعودية. يسلط هذا الاستحواذ والاندماج الضوء على الطلب الاستراتيجي المتزايد على حلول البنية التحتية السحابية للذكاء الاصطناعي بالمنطقة، والابتكار في هذا القطاع. تهانينا لفريق أوري!

AI has been defined by scarcity. Compute, power, data center space and capital. Every company building in AI is fighting for at least one of those. Radiant was built to fix that. Not by solving one constraint but by collapsing all four under a single entity. Brookfield-backed capital. A massive powered-land bank. Ori's compute and software stack. Each one is a competitive advantage on its own. Together they compound. When you eliminate scarcity, AI stops looking like alchemy and starts looking like a utility. That's the shift. That's what we're building. Go deeper: radiant.co/blog/the-age-o…

Brookfield's Radiant acquires Ori Industries for combined AI cloud platform dlvr.it/TR7m3x

I'm so proud to announce that my first (ever) investment has been acquired by @Brookfield @Radiant_Infra !! When I first met Mahdi Yahya, he was in motion with his pivot from a distributed cloud computing platform to an AI-native GPU cloud, and I was one of the first five investors to back him as he made this pivot at the same time I was launching Banyan Ventures. The timing couldn't have been more perfect for both of us. We were both scrappy and looking to build a huge outcome. Over several rounds, Banyan invested seven figures, starting on seed terms. Shortly after investing, Mahdi Yahya asked me to join as his fractional COO, and I hired Scott Collins, Deepak Manoor, Christian Magorrian, Ben Whetzel, as well as help closing deals like Together AI, San Francisco Compute Company, and more. It was an extremely rewarding experience and became one of the early cornerstones of investments at Banyan Ventures. Now my first exit! I couldn't be more proud of the team at Ori for fighting through so much adversity and ultimately coming out on top. What I'm so excited about is to see the Ori team support Brookfield Radiant, a $100 billion GPU build-out, as they get ready for the IPO!! Banyan Ventures Fund I and other LLC's Banyan manages is now a proud shareholder of Brookfield Radiant. Congrats to the entire team!!

Nice piece by @tech_eu's @Cate_Lawrence on the merger of Ori with @Brookfield backed @Radiant_Infra. Worth your time: tech.eu/2026/02/24/rad…

@SebMoss over at Data Center Dynamics (@dcdnews ) nails the story on the creation of Radiant and why @Brookfield added Ori to the mix. Hint - its about the full stack: datacenterdynamics.com/en/news/brookf…

Brookfield challenges cloud giants with new offering. Radiant cloud platform will be tied to Brookfield's $10bn AI fund - DCD buff.ly/gElWXIN

We announced today the launch of Radiant, a vertically integrated AI infrastructure company that will help deliver sovereign compute capacity for governments and enterprises worldwide. Radiant is backed by Brookfield’s Artificial Intelligence Infrastructure Fund, which will anchor a program to deploy up to $100 billion across the AI value chain backed by long-term contracts. brkfld.co/yzd2m13p