TensorWave

The open AI ecosystem isn't slowing down, and Kimi K3 is the latest example. For developers, this release isn't just about benchmark scores. It's about having more choice, greater flexibility, and the freedom to build with the model that best fits the workload. The ecosystem is already moving fast to make Kimi K3 production-ready on AMD. Ecosystem partners like RadixArk have achieved 423 TPS with DSPARK and passed the official Kimi Vendor Verifier, while support for AMD Instinct™ MI350X and MI355X GPUs is already available through SGLang. At TensorWave, we're excited to see the open ecosystem continue to grow. Powered exclusively by AMD Instinct GPUs, our cloud is built to help developers evaluate and deploy the latest open-weight models as soon as they're ready. Read the full blog: https://lnkd.in/gViNwZjt

Signal65's latest report evaluated AMD Instinct™ MI355X and NVIDIA HGX B200 across production AI inference workloads. Here are a few findings that stood out: • MI355X led in raw performance on 2 of the 3 models tested. • Even where B200 delivered higher throughput, MI355X achieved up to 1.42X more tokens per dollar. • Long-context and high-concurrency workloads benefited from MI355X's larger HBM memory, delivering stronger performance and lower latency under load. • Across every workload tested, infrastructure economics translated into measurable business advantages. Read the full report attached...

At RAISE Summit, TensorWave President & Co-Founder Piotr Tomasik 🌊 joined Val Bercovici (WEKA), Carmen Li (Silicon Data), Parimal Pandya (Akamai Technologies), Sam V. Tabar (WhiteFiber) to discuss how compute is reshaping the future of AI, from power and capital to infrastructure and long-term growth. Key takeaways: ⚡ Power is becoming one of AI's biggest constraints. 🏗️ AI infrastructure is evolving into a competitive advantage. 📈 The AI buildout is a long-term structural shift, not a short-lived trend. Watch the full discussion: https://lnkd.in/gwcDbQ3g

We've signed the "Open Weights and American AI Leadership" letter. We believe the future of AI should be built on openness, innovation, and strong leadership in AI infrastructure. As an AMD-first AI cloud, we're committed to giving developers and enterprises the freedom to build with open models on high-performance infrastructure without vendor lock-in. We're excited to stand alongside the organizations supporting an open, competitive AI ecosystem and helping shape the next generation of AI innovation.

So... last week's afterparty got a little out of hand (in the best way). TensorWave and VAST Data are doing it again at AI4: craft cocktails, good company, and real talk about deploying AI at scale. Come see what everyone's been talking about. 📍 Electra Cocktail Club, The Palazzo at The Venetian 📅 During AI4 Week 🍸 21+ | RSVP required, space is limited Register to Attend: https://luma.com/bk1seglv

Leading AI labs and cloud providers are choosing AMD Helios for its open, full-stack performance. As seen at #AdvancingAI 2026, AMD launched the next-generation AI infrastructure and physical AI portfolio, led by AMD Helios rackscale solutions, now in production to be deployed by leading AI companies at gigawatt scale. Delivering frontier AI requires a fully integrated rack architecture, with every part of the stack pushing the boundaries of performance. Systems will be available from leading OEMs, including Bull, Hewlett Packard Enterprise, Lenovo and Supermicro, as well as infrastructure partners Sanmina and Wiwynn. Learn more: https://lnkd.in/eBy_cDfV

Thanks to everyone who joined us for ROCm & Roll! It was a fantastic evening of great conversations, new connections, and a chance to unwind with the AI community after an exciting day at AMD Advancing AI. A huge thank you to our incredible co-hosts and sponsors for helping make the event possible. We couldn't have done it without you: Trace3 | Wafer | AMD | Makora | Luma | Moreh | ai& | Edgecore Networks Corporation | VAST Data | Featherless AI | dstack And to everyone who stopped by, grabbed a waffle, sang a karaoke song (or cheered someone else on), and shared ideas with fellow builders... thank you for making it such a fun and memorable night. Can't wait to see you at the next one!

Watch TensorWave Co-Founder & CEO Darrick Horton 🌊 join SiliconANGLE & theCUBE at the NYSE to discuss: • Why TensorWave bet on AMD from day one • The future of open AI infrastructure • What it takes to deliver a seamless AI cloud experience • Where AI infrastructure is headed next Whether you're building, deploying, or investing in AI, this is a conversation you won't want to miss. Watch the full interview: https://lnkd.in/gzGAVye7

Join TensorWave, Featherless AI, and AMD Silo AI for "AMD for Powering Open-Model Inference at Scale", where we'll explore what it really takes to serve open-weight models in production. Learn how TensorWave built an AMD-first AI cloud on AMD Instinct™ GPUs and ROCm, and hear real-world lessons from Featherless AI on serving hundreds of models at scale... from throughput and cost to production readiness. July 23, 2026 1:00–1:45 PM PDT Featuring Darrick Horton 🌊 (TensorWave), Eugene Cheah (Featherless AI), and Markus Hartikainen (Silo AI | AMD) If you're building or deploying AI at scale, we'd love to see you there. https://lnkd.in/gXzc953E

Read Luma's latest blog , "Luma Runs Production Inference on AMD and TensorWave". https://lnkd.in/g_WpJySp Make sure to find the TensorWave team at AMD #AdvancingAI - Booth 935