Ben Cheung

When I led Partners for MongoDB for 10 years, we closed hundreds of millions of dollars in deals through the Amazon Web Services (AWS) Marketplace—an incredible testament to its power for ISVs. I was fortunate to build strong friendships with Ruba Borno and her team led by Chris Grusz. They are exceptional leaders who have made the AWS Marketplace the premier platform for independent software vendors to serve their customers. That’s why I’m thrilled that Cerebras Systems is launching a pay-as-you-go offering on Amazon Web Services (AWS) Marketplace. Now, customers can leverage their existing AWS payer accounts to access the world’s fastest AI inference—with usage-based, consolidated billing, zero upfront costs, and no long-term commitments. At Cerebras Systems, we deliver frontier AI models in the cloud up to 20x faster than leading GPUs, enabling groundbreaking applications that weren’t possible before: real-time code assistants, instant deep research, lifelike virtual agents, and more. Big thanks to the entire Amazon Web Services (AWS) team. This announcement is just the next milestone in our growing partnership—more to come! https://lnkd.in/e7_hgDyP

277

22 Comments

The demand for efficient, open, and accelerated data processing is only growing, and it’s great to see Voltron Data making big strides with its Accenture partnership. Their commitment to Apache Arrow, open-source analytics, and scalable AI infrastructure continues to drive innovation in the ecosystem. Normally the team at Voltron Data would help people deploy on GPU servers they own, but starting today you can access and test running SQL queries on GPUs using Voltron Data's Theseus: https://lnkd.in/gdhZREWy

127

4 Comments

NVIDIA is rewriting the infrastructure playbook, not just in compute architecture, but in how it sells, who gets access, and who ultimately owns the stack. For years, the big public cloud was an engine of three: Amazon, Microsoft, and Google. Not anymore, and Jensen is making it clear. He is driving a broader, deeper, and more distributed infrastructure platform moat. The future of AI infrastructure won’t be built by just three hyperscalers. It will be nation-states, enterprises, neoclouds, and eventually AI factories, because NVIDIA killed the cloud (You can read my article here: https://lnkd.in/gxZnJ4zT) At Hydra Host, we’ve been predicting this shift for years. The real effect of NVIDIA GPUs isn’t just AI and its tokens. The secondary effect? Infrastructure that’s entirely agnostic. The demand is too global, strategic, and fragmented to be served by a closed, centralized model. Public cloud was built for CPUs. It wasn’t built for this. This is about platform control. The next era of AI infrastructure will be sovereign, distributed, and vertically integrated. NVIDIA is positioned to be the trusted computing layer underneath it all. And Jensen knows the math. He doesn’t get to $8 trillion by selling to three companies who are now building their own chips anyway. The public cloud isn’t going away. But its dominance is fading. The center of gravity is shifting. Jensen is building the next computing platform that looks much more like the "Apple of semiconductors". He has the playbook and resources to do it. https://lnkd.in/gjEjuWyM

70

3 Comments

🚀 Sharing BigQuery's Momentum: 5X Customers, Unified AI & Data, and What's Next! It was a pleasure to sit down with Savannah Peterson and David Vellante on SiliconANGLE & theCUBE at Google Cloud Next recently! ✨ We had a fantastic discussion diving into the incredible scale and innovation happening with our data platform, BigQuery. I was excited to share some key updates reflecting the momentum we're seeing: ✅ 𝟱𝗫 𝘁𝗵𝗲 𝗖𝘂𝘀𝘁𝗼𝗺𝗲𝗿𝘀: For the first time, we shared that BigQuery now serves 5 times the number of customers compared to key competitors in the data warehouse/AI platform space – a testament to the trust and value customers are finding. ✅ 𝗠𝗶𝗴𝗿𝗮𝘁𝗶𝗼𝗻 𝗦𝘂𝗿𝗴𝗲: We've seen a 300% increase in the use of our migration services over the past year as organizations increasingly choose Google Cloud to modernize their data estates and leverage our AI capabilities. ✅ 𝗠𝗮𝘀𝘀𝗶𝘃𝗲 𝗗𝗮𝘁𝗮 𝗚𝗿𝗼𝘄𝘁𝗵: We're now managing exabytes of data, with a 30% increase in volume over the last year, reflecting the significant workloads customers are running and trusting us with. We also dove into the crucial shift towards truly unifying data and AI 🧠 – bringing AI to the data foundation, not just bolting it on. I explained how our new Autonomous Data Foundation capabilities, powered by agentic AI, are automating complex tasks. We also touched upon how this technology enables powerful real-time activation, like the fantastic example from Mattel, Inc., who used #BigQuery and #Vertex AI to identify and fix product issues within the same production run! It's an exciting time for data and AI, driven by customer needs and groundbreaking innovation. If you'd like to hear more about these trends and the forces driving BigQuery's growth, you can watch the full conversation here ▶️: https://lnkd.in/d-kMGkgR #GoogleCloudNext #BigQuery #DataAnalytics #AI #CloudComputing #DataStrategy #GenerativeAI #theCUBE Tomas Talius; Sanjay Krishnamurthi; Geeta Banda; Sean Zinsmeister; Peter Bailis; Vinay Balasubramaniam; Abhishek Kashyap; Chaitanya (Chai) Pydimukkala; Vijay Venugopal; Adam Wilson; Shashank Gandhi; Sam McVeety; Alon Halevy; Andi Gutmans; Taylor Mikolasy; Katie Camacho

168

4 Comments

GraphRAG can be a more effective way to perform RAG (Retrieval Augmented Generation) in Generative AI. Below is a good intro to how to do GraphRAG, using LangChain and Spanner Graph. https://lnkd.in/e79KXN5k

46