Hardware Podcasts

How Much is Enough for Hyperscale Companies?

EE Times Current · 26 Jul 2019 · 29m

What this episode covers

Brian Santo hosts a discussion on how hyperscale companies like Google, Facebook, Amazon, and Microsoft are reshaping the semiconductor industry by demanding extreme power efficiency and designing custom AI accelerators. Experts Rick Merritt, Gary Hilson, and Dylan McGrath explain that these giants are pushing beyond standard roadmaps, collaborating directly with equipment vendors like Applied Materials, and leveraging their vast capital to drive innovation in memory and process technology. The episode concludes with Wally Rhines' insight that these companies' ultimate goal is owning the data and systems rather than just selling silicon, raising concerns about their growing societal power.

Most relevant to

large enterprises

Topics

Mentioned

Companies
Google, Facebook, Amazon, Microsoft, Apple, Applied Materials, Broadcom, Verasilicon, AMD
Products
Intel Optane, Tensor
Also
Semicon West, EDA

Related episodes from other shows

Other shows on the same ground, closest first.

  • The Circuit · 43m

    Episode 15: Growth Drivers for Semiconductors in the Datacenter

    Ben Bajarin and Jay Goldberg discuss the shifting economics of data center semiconductors, highlighting that while total silicon content is growing, CPU share is declining as GPU, FPGA, and accelerator workloads expand. They analyze how hyperscalers like Google, Amazon, and Microsoft are pursuing custom silicon to control roadmaps and solve specific pain points, such as Google's TPU for YouTube, while noting the high risks for startups entering the accelerator market. The conversation emphasizes that heterogeneous compute is driving value, with high-intensity workloads like LLM training justifying expensive chips like Nvidia's H100, whereas commodity workloads remain on cheaper, parallelizable infrastructure.

    Chips, FPGAs & semiconductorsPricing, margins & unit economics
    7 chapters

  • The Circuit · 47m

    EP 136: WE DON'T HAVE THE POWER!

    Ben Bajarin and Jay Goldberg discuss the definition of hyperscalers versus Neo clouds, noting that companies like CoreWeave and Oracle have pivoted from Bitcoin mining to AI infrastructure by leveraging locked-in power contracts. They analyze the storage bottleneck driven by OpenAI's Sora app and Meta's acquisition of Rivos for custom silicon development, highlighting the intense competition for HBM and NAND capacity. The episode concludes with a tour of Intel's Fab 52, emphasizing the technological marvel of 18A process manufacturing and the strategic importance of keeping such production in the US.

    Chips, FPGAs & semiconductorsSourcing & supply chain
    7 chapters

  • The Circuit · 46m

    EP 152: CAPEXXXXXX!!!! The Trillion Dollar Datacenter Race and More

    Ben Bajarin and Jay Goldberg discuss the surge in hyperscaler capital expenditure, dubbed 'CapEx Palooza,' driven by massive backlogs from Microsoft, Google, and Amazon, and the resulting investor skepticism regarding returns on this spend. They analyze the competitive dynamics between AWS, Azure, and GCP, noting that spending is largely driven by market share battles rather than immediate AI revenue, while also covering the impact of rising memory prices on consumer electronics from Qualcomm and MediaTek. The hosts highlight strategic M&A activity, specifically Texas Instruments' acquisition of Silicon Labs and SiTime's purchase of Renesas's timing division, as key shifts in the semiconductor landscape.

    Policy, trade & tariffsPricing, margins & unit economics
    7 chapters

  • The Circuit · 53m

    EP 140: NVIDIA's $300B chart, CAPEX to the MOON FOREVER!

    Ben Bajarin and Jay Goldberg discuss the semiconductor industry's massive capital expenditure surge, focusing on NVIDIA's projected $300B revenue and the strategic shift of hyperscalers like Amazon and Google toward custom silicon (Tranium, TPUs) to reduce reliance on NVIDIA. They analyze the emergence of Substrate, a startup challenging ASML with new lithography tools, and the consolidation in the RF component market with Skyworks acquiring Qorvo. The hosts also touch on the fragmentation of the cloud market into hundreds of 'neo-clouds' and the uncertain future of AI infrastructure demand.

    Chips, FPGAs & semiconductorsScaling production
    6 chapters

  • The Circuit · 40m

    Episode 28: A Whole Lot of Custom Going On

    Ben Bajarin and Jay Goldberg discuss the hidden growth of custom silicon, highlighting Amazon's aggressive push with Graviton, Inferentia, and Trainium chips to achieve performance parity and energy efficiency. They explore how hyperscalers like Google, Microsoft, and AWS are developing semi-custom networking chips (e.g., Nitro) to reduce latency for AI workloads, while noting that automotive OEMs in China are increasingly licensing RISC-V IP for critical systems like ADAS and sensor fusion. The hosts analyze why these custom efforts remain under-reported and how consumer electronics marketing strategies may eventually drive automakers to brand their own silicon.

    Chips, FPGAs & semiconductorsAI & machine learning
    6 chapters

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