TensorWave

Hear from TensorWave CEO Darrick Horton, in an exclusive sit-down with TheCUBE during Cyber and AI Innovators Week at the NYSE, as he shares how the company is disrupting the AI compute landscape. Darrick spreads some #AMD #InstinctGPU love that highlights how the work between TensorWave and AMD is creating a cost-effective, open alternative to other proprietary "walled garden" offerings in the market.

Is the CUDA Moat Only 18 Months Deep? Last week, I attended a panel at a NYSE Wired and SiliconANGLE & theCUBE event featuring TensorWave and AMD, where Ramine Roane made a comment that stuck with me: "The CUDA moat is only as deep as the next chip generation." Initially, I was skeptical and even scoffed at the idea. CUDA has long been seen as NVIDIA's unassailable advantage. But like an earworm pop song, the statement kept playing in my head—and now, a week later, I find myself rethinking everything. Here’s why: NVIDIA’s dominance has been built on the leapfrogging performance of each new chip generation, driven by hardware features and tightly coupled software advancements HARD TIED to the new hardware. However, this model inherently undermines the value proposition of previous generations, especially in inference workloads, where shared memory and processing through NVLink aren’t essential. At the same time, the rise of higher-level software abstractions, like VLLM, is reshaping the landscape. These tools enable core advancements—such as flash attention, efficient batching, and optimized predictions—at a layer far removed from CUDA, ROCm, or Habana. The result? The advantages of CUDA are becoming less relevant as alternative ecosystems reach a baseline level of support for these higher-level libraries. In fact, KamiwazaAI already seen proof points of this shift set to happen 2025. This opens the door for real competition in inference workloads and the rise of silicon neutrality—just as enterprises begin procuring GPUs to implement GenAI at scale. So, was Ramine right? I think he might be. NVIDIA’s CUDA moat may still dominate today, but in inference, it seems increasingly fragile—perhaps only 18 months deep at a time. This is something enterprises and vendors alike need to pay close attention to as the GenAI market accelerates. The question isn’t whether competition is coming—it’s how ready we’ll be when it arrives. Phillip Koblence Piotr Tomasik 🌊 Darrick Horton 🌊Brian J. Baumann John Furrier David Vellante Keith Townsend Ryan Shrout

In an exclusive sit-down with The Cube Cyber and AI Innovators Week at the NYSE, TensorWave CEO Darrick Horton shared how the company is disrupting the AI compute landscape. Key insights from the interview include: - Why AMD? Darrick reveals how TensorWave’s partnership with AMD is creating a cost-effective, open alternative to NVIDIA’s monopoly. - Focused Excellence: Learn how specialization enables TensorWave to outperform hyperscalers like AWS and Google Cloud in AI-specific workloads. - Scaling AI: Discover TensorWave’s approach to building gigawatt-class data centers and embracing cutting-edge cooling technologies like liquid and hybrid solutions. - What’s Next: Scaling globally with hundreds of thousands of GPUs in 2025. Curious how TensorWave is reshaping AI infrastructure for training, fine-tuning, and inference? Watch the full interview now 👇 https://lnkd.in/gz3aY2hr

Last night, we hosted an incredible panel and happy hour in partnership with NYSE, bringing together leaders from Starbucks, the National Hockey League (NHL), Citadel, Perplexity, Meta, ServiceNow, Zoom, IBM, and more to explore the future of enterprise AI on AMD. Huge thanks to to all who attended!

🚨 AMD ROCm 6.3 is here! 🚨 Standout features include: - SGLang support for accelerated AI inferencing - Re-engineered FlashAttention-2 for optimized AI training and inference - Multi-node FFT support to supercharge your HPC workflows - Enhanced computer vision libraries - AMD Fortran compiler for bridging legacy code to GPU acceleration. Uncover all of the latest updated here 👉 https://buff.ly/3VArajM

👏👏👏

Congratulations Dr. Lisa Su 👏

Happy Thanksgiving all! 🦃

Thanks for the article Wccftech 🌊 "TensorWave Plans To Make AMD's Offering Much More Aggressive, With The Aim of Breaking NVIDIA's Monopoly, Plans On Making Gigawatt-Tier GPU Clusters With MI300X, MI325X, MI350X AI Accelerators "