Hardware Podcasts

Ep 149: TSMC Q4 25 Earnings, OpenAI Needs more Compute and Monetization

The Circuit · 19 Jan 2026 · 54m · Episode 151

What this episode covers

Ben Bajarin and Jay Goldberg analyze TSMC's Q4 25 earnings, highlighting the massive $54 billion CapEx commitment and the critical bottleneck of CoWoS advanced packaging, which currently contributes less than 10% of revenue despite its complexity. They discuss OpenAI's strategic partnerships with Cerebras and other chip vendors to secure compute, alongside the introduction of ads and lower-tier subscriptions to monetize their user base. Finally, the hosts break down GPU pricing trends, noting the significant premium hyperscalers charge compared to NeoClouds and the difficulty in tracking pricing for Nvidia's new Grace Blackwell chips due to long-term contract restrictions.

Most relevant to

large enterprises

Topics

Mentioned

Companies
TSMC, OpenAI, Cerebras, Nvidia, AWS
Products
Grace Blackwell, H100
Also
CoWoS, NeoClouds

Related episodes from other shows

Other shows on the same ground, closest first.

  • EE Times Current · 7m

    TSMC Scraps a Batch of Wafers | AI Limitations | ISSCC

    Alan Patterson reports that TSMC scrapped a batch of 12 and 16 nanometer wafers at Fab 14B in Tainan due to substandard photo resist, cutting Q1 revenue forecasts by $550 million. Junko Yoshida highlights Noriko Arai's book 'AI versus Chutolin,' emphasizing that current AI lacks true understanding and cannot read textbooks. Dylan McGrath and Rick Merritt cover the 2019 ISSCC, noting Intel's production-ready embedded MRAM at the 22nm node and discussions on unsupervised learning.

    Chips, FPGAs & semiconductorsAI & machine learning
    5 chapters

  • EE Times Current · 18m

    ON Semi’s Acquisition | TSMC’s Finer Nodes | Tesla’s AI Chip | China’s Fabless Companies

    Rick Merritt reports on ON Semiconductor's acquisition of GlobalFoundries' East Fishkill fab to transition to 300mm wafers and TSMC's roadmap updates on 7nm, 5nm, and advanced packaging. Junko Yoshida analyzes Tesla's new Full Self-Driving computer and criticizes the company's marketing of its autonomous capabilities. Echo Zhao discusses the 2019 China fabless survey, highlighting rapid growth, fragmentation, and the rise of IoT applications in the region.

    Chips, FPGAs & semiconductorsPolicy, trade & tariffs
    5 chapters

  • Oxide and Friends · 1h 31m

    AI Disruption: DeepSeek and Cerebras

    Andy Hock and James Wang of Cerebras discuss how their wafer-scale silicon executes open-weight models like DeepSeek R1 Distill Llama 70B faster than any number of GPUs, achieving 57x speedup by keeping all 44GB of model weights in on-chip SRAM. They explain that the release of efficient, open models like DeepSeek and Llama transformed Cerebras from an exotic supercomputer vendor into a provider of commodity inference APIs, standardizing the unit of sale as tokens. The guests highlight that hardware-software codesign and the shift toward reasoning models allow smaller teams to compete with well-funded incumbents by optimizing algorithms for specific architectures rather than relying on brute-force scaling.

    AI & machine learningChips, FPGAs & semiconductors
    10 chapters

  • Built 2 Scale · 1h 25m

    2025 Finale: Yann LeCun Raises $3B, Domain Addiction Confessions & Built 2 Scale Year Wrap

    Hosts Matt Perrott and Scott Wilcox review the year's tech news, focusing on Yann LeCun raising $3 billion for a spatial intelligence startup and the implications for autonomous home company Elysium. They discuss Sergey Brin returning to coding at Google, the competitive landscape between Google's TPUs and Nvidia's GPUs, and Brett Adcock's Figure AI. The episode concludes with a 'Receipts or Regrets' segment where they evaluate predictions like robots in homes, the Guzman y Gomez IPO, and OpenAI's hardware ambitions.

    AI & machine learningChips, FPGAs & semiconductors
    9 chapters

  • The Amp Hour · 1h 09m

    Bonkers Birthday Badinage

    Hosts Dave Jones and Chris Gammell celebrate the 100th episode of The Amp Hour, discussing the show's growth to 5,000 listeners and the launch of a new discussion server based on the Hacker News engine. They analyze the semiconductor industry, noting TSMC's $8-10 billion investment in a 450mm wafer fab and Intel's dismissal of ARM as a competitor. The episode also covers practical engineering topics, including the use of gallium nitride (GaN) FETs, the TI LM5113 driver, and the feasibility of sourcing cheap Chinese electronics components for product design.

    Chips, FPGAs & semiconductorsPCB design & layout
    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.