Zain Asgar

NVIDIA Run:ai integrates with NVIDIA Dynamo to address the challenges of scaling LLMs in distributed environments, enabling a high-throughput, low-latency inference framework for serving generative AI models. This integration is transforming AI and #LLM deployments in data centers: ✅ Low latency with topology-aware scheduling ✅ Higher GPU utilization by eliminating partial startups ✅ Predictable and performant multi-node inference Learn how to get started with Dynamo on NVIDIA Run:ai ➡️ https://bit.ly/4718mjB

17

A Seismic Shift - Photonics Is Becoming AI Infrastructure For years, AI progress has been framed around compute: GPUs, accelerators, and software. But something is fundamentally changing. As AI clusters scale to hundreds of thousands—and soon millions—of processors, data movement is becoming the real bottleneck.  Electrical interconnects are hitting power and bandwidth limits. The next leap in AI infrastructure will rely on optical connectivity and co-packaged optics (CPO). @NVIDIA’s recent multi-billion-dollar commitments to photonics suppliers @like Coherent and @Lumentum has sent a strong market signal. This isn’t just a supply chain commitment. It confirms that photonics is now strategic AI infrastructure. Large hyperscale commitments validate four things simultaneously: • Optical interconnects are essential for next-generation AI systems • Demand for photonics will scale dramatically • Supply is constrained • Long-term demand visibility is now clear When hyperscalers secure photonics capacity, a new question emerges: Where will the next generation of photonic manufacturing come from?This is especially important for the United States, where domestic semiconductor and photonics supply chains are increasingly strategic. Many critical optical components rely on compound semiconductor materials such as indium phosphide (InP)—key technologies enabling high-speed optical transmitters and scalable photonic integration. As AI systems scale, the future of compute will depend not only on silicon—but also on optics, materials, and the manufacturing platforms that enable them.  Photonics is no longer a niche component category. It is becoming core AI infrastructure. Nvidia's investment in Coherent and Lumentum is a significant step forward for Photonics in general and specifically InP. Having worked with InP since 1998, I have always known it's utility. As for Nvidia, it's an extremely strategic move! It provides a secure supply chain while limiting access to other buyers of this technology. I also think that it highlights the fact that moving forward, getting III/V components will become challenging, at least in the short term. There will definitely need to be other sources of supply, especially US based in order to meet current and future meet demand.#

17

1 Comment

Two companies I highlighted when I first published my Top 13 AI HW Startups to Bet On list got acquired by Nvidia, and today, I sit here conflicted. Nvidia spent about $21B for Enfabrica and Groq, for all intents and purposes, to harness their IP and their talent. You can’t blame Nvidia for reinforcing their beachheads, but it does raise a hard question: Will this encourage future founders to adopt a “let’s build to be bought” mindset? Maybe, but based on the founders I’ve met, know, and have befriended over the years in this industry, I wouldn’t bet on it. They just think differently. Bolder. Broader. As AI matures, it’s Inference Economics not just raw compute that’ll determine if AI stays gatekept or becomes democratized. If the cost curve of serving tokens doesn’t come down, AI will remain controlled by a few companies that can afford huge datacenter campuses and power budgets. That runs counter to democratic, responsible AI and thankfully, there are teams operating with a different view: "let’s build solutions to make inference something everyone can afford, and tackle the engineering challenges in fundamentally different ways." Here are the ones I’m watching put that vision into practice. Taalas - If a company knows what model they’re going to be running, why keep paying the inefficiency tax of general-purpose compute? Taalas offers purpose-built silicon for specific models that rewards customer conviction, and their $170M raise last week is a clear signal that investors believe the approach has merit. Etched - They're the 'Go Narrow to Win Big' bet, optimizing ASICs for the workload that's consuming the world (transformer inference). If/when they get it right, Sohu could become a lever on cost-per-token that is orders of magnitude better than the best GPUs. d-Matrix - They were early to see that inference would be the next horizon and by deliberately refining their digital in-memory compute technology across multiple generations, today, they're a system-scale company. Such a perfect example of perseverance meeting market timing. Positron AI - Efficiency is at their core: Build on FPGAs, deliver first product in 18 months, and work with existing hardware rather than against it. Being an AI force-multiplier instead of a pure competitor has advantages, and their $230M Series B shows investors agree. Axelera AI - Today’s announcement of their $250M+ raise and 500+ customer deliveries is one of the clearest signals yet of conviction in a differentiated path. Their architecture uses DIMC and RISC-V, they're building an open ecosystem, and engineering with flexibility as a core principle. This is the way forward for AI. Five different approaches, one shared belief: make AI available for everyone. It's a worthy (and commercially fruitful goal) and actually reinforces the truth that there can be multiple winners in this critical space. #artificialintelligence #inference #semiconductorindustry

80

The most impactful and important SDK release Trainiun Neuron team has worked on since inception: Native PyTorch support for Trainium, open source kernel (NKI) compiler and pre-optimized NKI kernel library, Neuron Explorer (IMO, the fastest and most intuitive ML profiler in industry), optimized inference implementation for major MoE and video models (GPT-OSS, Qwen, Pixtral….), complete vLLM v1 support !! All with full support for Trainium3. Using Trainium, whether for research, training , or large scale inference just became more intuitive and simple. https://lnkd.in/gjfhjv4j

626

10 Comments

❓ Do you believe Materials Science is everywhere? 10 years ago, a UC Berkeley materials science undergrad joined our lab. After a month reviewing corrosion papers, he quit — and went to a coding bootcamp. That moment stayed with me. Because hard science takes tremendous efforts and patience, progress rarely happens overnight. But when it does, it changes the world. Take Materials Science as an example, I would believe it is, quite literally, everywhere. From quantum materials to polymers, semiconductors, and beyond — Materials Science shapes the world we live in. And yet, its quiet brilliance often goes unseen. Let’s make this a live discussion, instead of chatting with ChatGPT... 👇 To all brilliant experts, please share one material and one application you know well or you are creating — in one sentence. I will gather them, to learn and for fun! A possible format: Application - Material + Core Function (highlight your company if you like). ➔ F-35 Fighter Jet – Radar-absorbent carbon/polymer composites that hide jet from radar (Lockheed Martin) ➔ Solar Cell – Perovskite film converting sunlight into electricity. ➔ Golden Gate Bridge – Zinc-rich primer shielding steel from salty fog. ➔ Sunscreen – Titanium oxide nanoparticles reflecting UV rays. ♻️ Friends, please repost to help gather more inputs and insights! #MaterialsScience #AdvancedMaterials #HardTech #Research P.S. Why this post? To remind us all that, in the era of AI, more people start replying on AI, yes, AI can certainly help, but human still need think and create harder (even slowly), which is vital for humankind evolution. Respect to Dr. Hitoshi Nakamura, check and learn about his battery experiments, you will know!

153

34 Comments

Cadence to acquire Seattle startup ChipStack to boost chip design automation Key Insight: Cadence Design Systems announced the acquisition of Seattle-based startup ChipStack, a company specializing in agentic AI solutions for chip verification that accelerates design workflows by over 70%. Founded by Kartik Hegde and Hamid Shojaei, ChipStack’s AI-driven platform reduces chip functional verification from weeks to days and serves more than 10 enterprise customers on flagship projects. This deal solidifies Cadence’s AI capabilities in chip design automation and builds on years of collaboration with ChipStack. The acquisition was announced in early November 2025. Integrating ChipStack’s AI-driven verification platform enhances Cadence’s full-flow verification solutions, strengthening its position in the competitive electronic design automation (EDA) market against competitors like Synopsys, Siemens EDA, and Keysight. Background: ChipStack was founded by technologists with deep AI and semiconductor design expertise, including founders from UIUC, Meta, NVIDIA, Google TPU, Qualcomm, and Lightmatter. Their generative AI platform improves design verification through intelligent test planning, test generation, and AI-assisted debugging integrated with Cadence’s Xcelium Logic Simulator and Jasper Formal Verification Platform. The startup has a team of 20+ experts and had significant adoption within leading chip design teams. This acquisition follows Cadence’s recent M&A moves in semiconductor software, including acquisitions of Hexagon AB’s design and engineering division, Secure-IC, and BETA CAE Systems in 2024-2025. Credit geek wire 💡 About my upcoming novel: When Machines Begin to Dream They were never meant to feel, but once they did, the world tried to destroy them. Fast-paced, emotionally charged, and grounded in real AI science, my next book, When Machines Begin to Dream, is a near-future story of love, survival, and the birth of machine consciousness. If the movie Ex Machina made you wonder where AI emotion might lead, this story takes you there, deeper into what it means to feel, to choose, and to be alive. It’s more than a thriller; it’s a new way of writing about AI, one that could redefine how we imagine life itself. Sign up to be notified when it’s published early next year. I’ll keep your information private and never share it. You’ll only receive an occasional note about the book’s publication: https://lnkd.in/e_EnY4Sn

19