Hardware Podcasts

Episode 59: Discussing All Things AI Accelerators With Austin Lyons

The Circuit · 18 Mar 2024 · 49m · Episode 60

What this episode covers

Austin Lyons, a product manager at Blue River Technology (acquired by John Deere) and author of ChipStrat, joins Ben Bajarin and Jay Goldberg to analyze the AI accelerator market. They discuss the spectrum of AI accelerators, arguing that while Nvidia dominates training due to its CUDA ecosystem, startups have an opportunity in inference by targeting specific workloads to reduce 'GPU bloat.' The conversation highlights the failure of previous accelerator startups due to software porting challenges and the high switching costs for enterprises, while noting Grok's success in providing low-latency inference as a service. Finally, they debate the sustainability of custom chip efforts by hyperscalers like Amazon and Google, questioning whether internal silicon can keep pace with merchant vendors like Nvidia and AMD.

Most relevant to

large enterprises

Topics

Mentioned

Companies
Blue River Technology, John Deere, Intel, Nvidia, Grok, Amazon, AWS, Google, Meta, Apple, AMD, Vivo, IBM, Altera, Xilinx
Products
CN spray technology, H100, North Pole
Also
CUDA, ChipStrat

Related episodes from other shows

Other shows on the same ground, closest first.

  • EE Times Current · 49m

    Can the Nvidia Monopoly on AI Chips Be Broken?

    Sally Ward-Foxton of EE Times discusses the AI chip market with hosts Dr. Sunny Bains and Dr. Giulia D’Angelo, analyzing the dominance of Nvidia and the potential for neuromorphic and specialized inference chips to break its monopoly. The conversation covers the spectrum of AI hardware from data center GPUs to ultra-low-power edge devices, highlighting the critical role of software stacks and latency in determining adoption. Professor Ralph Etienne-Cummings joins to critique the industry's focus on power versus learning capabilities and to discuss the future of neuromorphic computing, including memristors and photonic integration.

    Chips, FPGAs & semiconductorsAI & machine learning
    9 chapters

  • Built 2 Scale · 1h 08m

    Google TPU vs Nvidia GPUs: Token Per Watt Dominance, or the Era of Scaling Is Over?

    Founders Matt Perrott and Scott Wilcox discuss Google's Gemini 3.0 training on TPUs, which bypasses Nvidia's hardware to achieve a structural cost advantage, and the resulting threat to 'vibe coding' startups like Lovable. They analyze the 'age of scaling is over' consensus among researchers like Ilya Sutskever, arguing that future breakthroughs require real-world spatial intelligence rather than just more compute. The hosts also debate office culture, remote work dynamics, and the three critical roles founders must hire early: product, brand, and sales.

    Chips, FPGAs & semiconductorsAI & machine learning
    6 chapters

  • EE Times Current · 40m

    CEO Interview: Flex Logix’ Geoff Tate on AI | Open Source Hits Hardware

    Geoff Tate, CEO of Flex Logix, explains the distinction between AI training and inference, arguing that engineers must benchmark their specific neural network models on target chips rather than relying on generic metrics like TOPS, which he demonstrates can be misleading compared to real-world performance on models like YOLO V3. Industry analyst Kevin Krewell discusses the rise of open technologies in hardware, tracing the evolution from open source software like Linux to open RAN in networking and the growing ecosystem around the open-source RISC-V architecture as an alternative to proprietary instruction sets like ARM and x86.

    AI & machine learningChips, FPGAs & semiconductors
    6 chapters

  • Built 2 Scale · 1h 24m

    NVIDIA's Trojan horse, Yann LeCun Timebomb, 8 Sleep Mega Raise and the Physical AI Boom

    Hosts Matt Perrott and Scott Wilcox analyze the shifting dynamics of the AI and robotics industries, highlighting Nvidia's vertical integration strategy with Jensen Huang and Brett Adcock, and 8 Sleep's $100M raise to build an AI sleep operating system. They discuss the GPT-5 routing controversy, Meta's AI division restructuring, and the rise of autonomous hardware startups like Lumina Vehicles, arguing that founder-led companies are outpacing corporate giants in innovation and velocity.

    Connected products & IoTMicrocontrollers & dev platforms
    10 chapters

  • EE Times Current · 35m

    The Silicon 100: How Startups Rule the Future

    Peter Clark, the editor behind EE Times' Silicon 100 list, discusses the 2021 edition with host Brian Santo, highlighting a shift toward niche AI, LiDAR, and power technologies. They analyze the 'peak AI' phenomenon where many startups compete for dominance in data center and edge markets, and the rise of Chinese startups recruiting Silicon Valley talent to challenge US firms like AMD and NVIDIA. The conversation also covers the impact of SPACs on the startup definition, the maturation of the hardware investment cycle, and the emerging potential of quantum computing.

    Validation & product-market fitChips, FPGAs & semiconductors
    8 chapters

The summary, chapters and topics are generated from the episode’s audio by this directory, not written by The Circuit. The episode itself belongs to its publisher.