WEKA

✨ Malaysia vLLM Day is bringing together AI builders, innovators, and business leaders from across ASEAN next week to discuss all things sovereign, open, and production-grade AI. If you care about real-world deployments, hands-on learning, or building AI that’s truly inclusive and sustainable, this is the event to be at. We’re thrilled that Ronald Pereira will take the stage to share “How WEKA AMG Powers Next-Gen AI Deployments” — diving into how Augmented Memory Grid is helping teams push past memory limits and unlock faster, more efficient, production-ready AI. If you're in the region (or just want to tap into some serious AI momentum), don’t miss it. 🔗 https://weka.ly/48ncYjI

🦃 From all of us at WEKA: Happy Thanksgiving! Today, we’re taking a moment to slow down, grab a second helping of our favorite sides, and reflect on what we’re most grateful for — and that list is long. We’re thankful for: 💜 A team of brilliant, curious, genuinely awesome humans 🙌 Customers who push us to innovate and inspire us every single day 🤝 Partners who make the impossible possible right alongside us You’re the reason WEKA keeps reaching new heights — and why we love the work we get to do. Wishing our WEKA families, friends, and colleagues a restful, joyful, and delicious Thanksgiving. May your plates (and your hearts) be full.

💬 AI is moving fast — but for enterprises, speed isn’t the only metric that matters. WEKA’s Val Bercovici recently sat down with Steve McDowell of NAND Research for a candid conversation about something AI teams don’t talk about enough: 💰 the real economics of AI, 📈 how to keep AI initiatives cash-flow positive (not cash-flow negative), and 💡 why smart infrastructure choices can determine whether you scale — or end up throttling your own users. Explore more in this report: https://weka.ly/48zWy8F

How do you build a sovereign #AI cloud that can serve 70 million people — and compete on the global stage? For Siam AI, the answer started with a bold vision: create an intelligent, culturally aware AI platform designed to understand and support Thailand's people and businesses. To make that vision a reality, they needed an infrastructure foundation that could scale effortlessly, eliminate latency, and keep GPUs running at full throttle — and they found it with WEKA's NeuralMesh. Together, Siam AI and WEKA stood up a production-ready AI storage platform that could also support cutting-edge NVIDIA accelerated computing in just four days — delivering: ⚡ ~3 TB/s reads and ~1.5 TB/s writes ⚡ <0.5 ms latency across millions of IOPS ⚡ Seamless access for 140 GPU clients With NeuralMesh, Siam AI has built Thailand's first sovereign AI cloud — redefining what's possible for national-scale innovation. See how they did it. 🔗 https://weka.ly/3XOa9D9

💥 ICYMI: Last week, we announced that Augmented Memory Grid is now commercially available on NeuralMesh and Oracle Cloud Infrastructure, making it easier than ever for teams to break through the AI memory wall and run agentic, long-context inference at scale. Why is this a big deal? As models get bigger and smarter, inference is hitting its limit. GPUs keep running out of memory, cache gets evicted, and systems waste time re-generating the same tokens again and again — costing money, power, and performance. If you want the full story and details, this blog is a must-read: https://weka.ly/48xFO1U

🦃🍂 Nothing brings a team together like good food — and our Campbell crew delivered! Some of our team recently gathered for a Thanksgiving celebration, complete with full plates, full hearts, and more shared laughs than we could count. From classic comfort dishes to secret family recipes, the spread did not disappoint. Because we have to ask… If you could only pick ONE Thanksgiving side dish for the rest of time, what are you choosing? 👀

🎬 And just like that… #SC25 is a wrap – and what a week it’s been! From engaging conversations to insightful sessions with SiliconANGLE & theCUBE, one thing has been crystal clear: ➡️ The AI world is done accepting bottlenecks as normal. We heard from customers, partners, and innovators who are building AI factories that demand faster inference, smarter infrastructure, and efficiency at massive scale. And seeing the excitement around NeuralMesh and Augmented Memory Grid? Absolutely amazing. A huge thank you to our partners and everyone who stopped by and connected with the team this week. The conversations, collaboration, and momentum were next-level. More to come! #Supercomputing25 #HPCignites

🎤 WEKA's Callan Fox took a deep dive into Augmented Memory Grid at #SC25 and broke down how it’s solving one of the most painful challenges in AI today. As models get larger and more agentic, inference gets expensive fast. Recomputing the same tokens again and again slows pipelines, wastes GPU cycles, and drives up cost. Take a closer look at how Augmented Memory Grid works — and why it matters. 🔗 https://weka.ly/4ppQkhw #Supercomputing25 #HPCignites

🏭 AI factories are the new industrial infrastructure—turning raw data into insights at massive scale. WEKA’s Shimon Ben-David and NVIDIA’s Dion Harris joined SiliconANGLE & theCUBE at #SC25 to discuss how building efficient AI factories—combining high-performance compute and storage—can maximize token output and drive real ROI. It was an insightful conversation on architecting #AIInfrastructure that multiplies performance and boosts profitability. Special thank you to NVIDIA for sharing the stage with us! #Supercomputing25 #HPCignites

📺 WEKA's Callan Fox, Principal AI Product Manager, and Betsy Chernoff, Principal AI Product Marketing Manager, joined SiliconANGLE & theCUBE's hosts David Vellante and Jackie McGuire at #SC25 to dive into the hottest topics in #AIInfrastructure. From breaking the AI memory wall to making inference faster and more efficient, they explored the challenges and innovations shaping the future of AI. 👉 Watch the full session: https://weka.ly/47QUI35 #Supercomputing25 #HPCignites