TensorWave

We’re less than a month away from the Beyond Summit on April 8 in San Francisco. Here’s a first look at some of the featured speakers joining us to talk about the future of AI infrastructure, open ecosystems, and what comes next: • Dylan Patel (@dylan522p) — Founder, @SemiAnalysis_  • Gregory Diamos (@GregoryDiamos) — Architect, ScalarLM  • Mathias Lechner (@mlech26l) — Co-Founder & CTO, @liquidai  • Quentin Anthony (@QuentinAnthon15)— VP of AI Engineering, @ZyphraAI  • Darrick Horton (@DarrickHorton) — CEO, TensorWave More speakers, demos, and announcements coming soon. Join the engineers, researchers, and founders building the next generation of AI infrastructure. 📅 April 8 📍 San Francisco Register: na2.hubs.ly/H04hqV40

CUDA is not the only path forward, and the conversation is finally catching up. Appreciate the highlight by Anissa from The Information on what we’re building with Beyond Summit. The point has not changed: serious AI infrastructure is being built outside closed ecosystems, and more teams are ready to explore what comes next. This week, Beyond Summit brings together the builders, operators, and innovators pushing AI infrastructure forward on new terms. Limited space remaining. Request access: na2.hubs.ly/H04KgpT0

Here’s the next sneak peak at the Beyond agenda. Breaking Free from GPU Fragmentation: Mojo 🔥, MAX, and the Future of Heterogeneous AI Infra with Mostafa Hagog at @Modular Discover how Modular is building AI’s unified compute layer, enabling developers to deploy production gen AI systems across any hardware in a matter of minutes. The Modular platform lets you run frontier models at blazing fast speed, remains GPU agnostic, and removes the complexity of fragmented toolchains all with a software stack that just works. More session announcements are on the way. Haven't got your ticket? RSVP: na2.hubs.ly/H04K00L0

Here’s a sneak peak at the Beyond Summit agenda. Building Your Personal
Open Clawagent with
Open-Source Models, with Mahdi Ghodsi and Eda Zhou from @AMD . Discover the core ideas behind modern agentic frameworks and how AI agents reason, plan, and use tools to complete tasks. You'll host your own open-weight models on AMD GPUs, build dynamic agentic workflows using architectures like OpenClaw, and design, extend, and deploy your own personal AI agents. More session announcements are on the way.

The countdown to Beyond Summit is on. We’re spotlighting the sponsors helping power every layer of modern AI infrastructure, from silicon and systems to manufacturing and networking. This is the stack behind what comes next. The sponsors helping bring Beyond Summit to life: 💿 Platinum: AMD (@AMD) 🏅 Gold: ZT Systems (@ZTSystems), a Sanmina Company (@SanminaCorp) 🥈 Silver: Credo (@credosemi), Edgecore (@EdgecoreNetwork) This will be where the stack gets redefined. Be there: na2.hubs.ly/H04H6LJ0

“Free” cloud credits can hide a very real bill. The compute may be covered. The lock-in is not. We broke down the hidden costs behind the offer. Read the blog: na2.hubs.ly/H04G3JD0

Our CEO @SpectralMichael will be a keynote speaker at the Beyond Summit 2026 on April 8 in San Francisco, joining a panel of leading voices from HPC and AI infrastructure. Hosted by @TensorWave 👇

Edgecore is joining Beyond Summit as a Silver Sponsor. AI infrastructure does not scale on compute alone. It depends on fast, reliable, high-performance networking that can keep massive workloads moving without becoming the bottleneck. Edgecore is helping make that possible with open networking built for modern AI environments. More hardware means very little if the network cannot keep pace. Join Edgecore and the people building the infrastructure behind the next wave of AI at Beyond Summit on April 8: na2.hubs.ly/H04FSMz0

One week out from Beyond Summit. The countdown is on. AI infrastructure’s boldest builders, operators, and innovators are about to take over San Francisco. Beyond Constraints. Beyond Uncertainty. Beyond Dependency. Beyond Assumptions. See you there: na2.hubs.ly/H04F0Qf0

Next up on the Beyond Summit sponsor lineup: Credo (@credosemi), joining us as a Silver Sponsor. As AI infrastructure scales, weak links get exposed fast, and connectivity is one of them. Credo is helping push the stack forward with the high-speed, low-power interconnect and networking technology needed to keep modern AI systems moving at full speed. More compute means nothing if the network cannot keep up. Join Credo and the teams building what comes next at Beyond Summit next week on April 8: na2.hubs.ly/H04DJNd0

Next week, @SpectralMichael is at @TensorWave’s Beyond Summit in SF discussing SCALE: native compilation of your CUDA codebases for NVIDIA & AMD GPUs. No translation layers. See you there! luma.com/g03nmrq8

@ZTSystems , a @SanminaCorp Company, is joining Beyond Summit 2026 as our Gold Sponsor. As a preferred AMD NPI partner, they provide end-to-end vertical integration for cloud and AI infrastructure. ZT Systems engineers and builds purpose-built AI systems, delivering rack-level infrastructure, North American production, and global fulfillment to accelerate deployment timelines. The stack is evolving. Demand is climbing. The companies that can actually build are the ones shaping what comes next. Be there on April 8 in San Francisco: na2.hubs.ly/H04C3300

the 2022-23 GPU shortage is why @tensorwave exists and why we went all in on AMD. AI was dependent on one vendor and it crushed the entire ecosystem. today the constraint is a bit different, it’s not just GPUs. it’s access to data centers and power available rn. but the result is the same: the market is forced to diversify. we were built for this moment and it’s exciting to see things go from headwinds to tailwinds. 🚀

Hard launch. April 8, 2026 | San Francisco, CA Secure your seat. na2.hubs.ly/H04B6G_0 Speakers: Jeff Tatarchuck, Darrick Horton, Piotr Tomasik, TensorWave | Ashish Vaswani, Essential AI | Greg Diamos, ScalarLM | Anush Elangovan, Neha Prikyira, AMD | Dylan Patel, SemiAnalysis | Matthia Lechner, Liquid AI | Andrzej Janik, ZLUDA | Farbod Tavakkoli, AT&T | Quentin Anthony, Zyphra | Emilio Andere, Wafer AI | Marc Fleury, dstack | Michael Søndergaard, Spectral Compute | George Cameron, Artificial Analysis | Eugene Cheah, Featherless AI | Bill Brennan, Credo | Akhil Sharma, Silares | Waleed Atallah, Makora | Eric Hartford, Quixi AI | Jay Dawani, Lemurian Labs | Jon Stallman, PG&E | David Kanter, MLPerf | Mostafa Hagog, Modular

CUDA lock-in is usually framed as a technical dependency. It’s bigger than that. It shapes procurement decisions, slows platform flexibility, narrows infrastructure strategy, and makes switching costs everyone’s problem, not just engineering’s. The real issue isn’t whether teams can move beyond CUDA. It’s whether organizations are willing to design for optionality before lock-in becomes operating policy. That’s the conversation in our latest blog: na2.hubs.ly/H04B7Q00