Introducing Discovery and Lux: New AI systems with DOE and AMD

AMD and HPE have partnered with the US Department of Energy to develop high-performance AI supercomputers. This $1 billion initiative aims to boost American computational research, support AI advances, and enhance national infrastructure. Check it out here: https://lnkd.in/eirmyfdJ What’s your take on this topic? Share your thoughts below! #AI, #TechNews, #ArtificialIntelligence, #Automation, #Innovation

I think this is a very conservative estimate of how much impact the AI revolution has. One key point here is that this revolution depends on how much computing power we can generate. So, any AI infrastructure related businesses, chip makers, datacenter builders and operators, electricity generators, etc. will benefit greatly in the next 5-10 years. Think how much compute we will need when everyone from our gramma to every level of government are using AI.

The Wall Street Journal’s coverage of the OpenAI-AMD partnership captures something bigger than a single deal: AMD, OpenAI Announce Massive AI-Chip Deal   A 6GW commitment at this scale shows just how central compute capacity has become to global innovation. But beyond the numbers, it marks a shift toward deeper collaboration between hardware and AI companies - where product roadmaps, incentives, and ecosystems are increasingly built in sync.   And the lines between AI, semiconductors, and energy are basically gone. You can’t scale one without solving for the others. The next phase of AI isn’t about just adding more compute. It’s about doing it efficiently, sustainably, and intelligently.   Partnerships like this show how tightly materials, power systems, and thermal design are now intertwined with the future of AI. The industry is starting to operate as a connected system, one that spans physics, infrastructure, and intelligence itself.

Nvidia Blackwell B200 Sets New AI Benchmark The Nvidia Blackwell B200 has emerged as a leader in AI performance as of 2025, boasting a remarkable 15x ROI promise. It excels in InferenceMAX benchmarks, establishing new standards for AI workloads. This chip features advanced parallelization that enhances computational efficiency and AI-powered predictive management for improved system resilience. As a result, organizations leveraging the B200 can expect significant gains in performance and reliability. The advancements in this technology mark a pivotal moment for the future of AI infrastructure.

What is The Power Behind the AI Revolution? In my Master Class presentation to investors and experts in AI data centers, I recently spoke about what I believe is the real foundation of the AI revolution: power. Just as the California Gold Rush was fueled by the picks and shovels that made it possible for miners to explore a new frontier, today’s AI boom equally depends on an essential but very scarce commodity: energy. As NVIDIA CEO Jensen Huang said in his recent keynote address: AI data centers are the new factories. Each one converting energy into tokens, the raw units of intelligence, that are then processed by Large Language Models like ChatGPT OpenA.I - The future of technology and Grok. Power is so essential that NVIDIA’s latest Rubin rack now draws 375 kilowatts, some 75X more power than your typical Data Center rack consumed just a few years ago. That type of scale, orders of magnitude higher, should tell us how important securing strategic energized land is in this new era of AI. At RHPC, we’re scaling to meet that challenge head-on. Our development pipeline spans 4,800 megawatts across Texas, Virginia, and Europe—designed to power the next generation of sustainable, efficient AI infrastructure. We’re currently raising our Series A round, to accelerate our mission of creating power-ready AI sites that hyperscalers, colocators, innovation startups, and enterprises can deploy quickly and sustainably. We aim to deliver the right kind of power, in the right place, at the right time and enable the next wave of intelligence infrastructure. If you share our vision or want to learn more, I invite you to connect or reach out directly @ https://www.rhpc.ai The AI gold rush has begun. The real winners will be those who deliver the electrons that power the intelligence of the future. #AI #DataCenters #EnergyTransition #Infrastructure #Investment #Texas #Sustainability #NVIDIA #Hyperscale #PowerTheFuture

AI is accelerating demand for faster, more efficient silicon, and it’s changing the way we think about compute. On stage at the All-In Summit, Rene Haas spoke about how this shift is reshaping the industry. He reflected on how the surge in AI-first workloads is driving a new approach to chip design, why the rise of physical AI will require power-efficient compute everywhere, and what it will take to keep pace with innovation that moves at unprecedented speed. Loved this insightful look at why the conversation around AI has to be as much about semiconductors as it is about software. https://okt.to/U9Tjno

🚀 Scaling AI Smarter: Why Efficiency Beats Excess As enterprises race to scale AI, many are investing heavily in GPUs—but most are missing the mark. According to recent findings, over 75% of organizations report GPU utilization below 70%, even at peak load. That’s billions in idle infrastructure, wasted energy, and delayed innovation. Fujitsu’s new AI Computing Broker (ACB) flips the script. Instead of static GPU allocation, ACB uses runtime-aware orchestration to dynamically assign GPUs where they’re needed most—unlocking latent capacity, accelerating throughput, and slashing energy consumption. 💡 In life sciences, ACB boosted AlphaFold2 throughput by 270% per GPU. In financial services, it’s speeding up risk model training. In retail, it’s scaling AI vision systems without extra hardware. And across industries, it’s helping leaders do more with less. 🌱 The environmental impact? Significant. By reducing idle time and reallocating resources, ACB helps enterprises shrink their carbon footprint while meeting sustainability goals. 🔑 As Nick Cowell, Fujitsu Distinguished Engineer, puts it: "In the race to scale AI, the winners will not be those who merely acquire the most GPUs. They will be those who orchestrate them best." Explore how Fujitsu’s AI Computing Broker can transform your AI infrastructure: Read the full article: https://lnkd.in/gnTCJ6Dz

AI is accelerating demand for faster, more efficient silicon, and it’s changing the way we think about compute. On stage at the All-In Summit, Rene Haas spoke about how this shift is reshaping the industry. He reflected on how the surge in AI-first workloads is driving a new approach to chip design, why the rise of physical AI will require power-efficient compute everywhere, and what it will take to keep pace with innovation that moves at unprecedented speed. It’s an insightful look at why the conversation around AI has to be as much about semiconductors as it is about software. https://okt.to/Th4yO7

𝙏𝙝𝙚 𝘼𝙄 𝙘𝙝𝙞𝙥 𝙢𝙖𝙧𝙠𝙚𝙩 𝙞𝙨 𝙨𝙚𝙩 𝙩𝙤 𝙚𝙭𝙥𝙡𝙤𝙙𝙚. 𝙁𝙧𝙤𝙢 𝙟𝙪𝙨𝙩 56.8 𝙗𝙞𝙡𝙡𝙞𝙤𝙣 𝙐𝙎𝘿 𝙞𝙣 2023, 𝙞𝙩 𝙞𝙨 𝙥𝙧𝙤𝙟𝙚𝙘𝙩𝙚𝙙 𝙩𝙤 𝙜𝙧𝙤𝙬 𝙩𝙤 927.7 𝙗𝙞𝙡𝙡𝙞𝙤𝙣 𝙐𝙎𝘿 𝙗𝙮 2034. This growth will be fueled by demand for advanced chips powering generative AI, autonomous systems, and enterprise AI adoption. 𝐊𝐞𝐲 𝐛𝐞𝐧𝐞𝐟𝐢𝐜𝐢𝐚𝐫𝐢𝐞𝐬: • 𝐍𝐯𝐢𝐝𝐢𝐚, already leading with GPUs optimized for AI training and inference • 𝐀𝐌𝐃, expanding its footprint with AI-focused accelerators • 𝐈𝐧𝐭𝐞𝐥, betting big on Gaudi and AI-focused server chips • 𝐓𝐒𝐌𝐂, the foundry behind most of the world’s advanced AI chips • 𝐁𝐫𝐨𝐚𝐝𝐜𝐨𝐦, 𝐐𝐮𝐚𝐥𝐜𝐨𝐦𝐦, 𝐚𝐧𝐝 𝐌𝐚𝐫𝐯𝐞𝐥𝐥, building specialized chips for AI workloads This is not just a technology shift. It is a trillion-dollar market in the making. Companies with strong chip design and manufacturing capabilities will shape the AI economy of the next decade. #smartinvesting #stock