Episode

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

Podcast
The Circuit
Published
Jan 19, 2026
Duration seconds
3282
Processing state
not_requested
Canonical source
https://share.transistor.fm/s/f38cc9dc
Audio
https://media.transistor.fm/f38cc9dc/cd29ffa1.mp3
JSON
/v1/public/podcasts/the-circuit-5928182/episodes/ep-149-tsmc-q4-25-earnings-openai-needs-more-compute-and-monetization
Markdown
/podcast/the-circuit-5928182/ep-149-tsmc-q4-25-earnings-openai-needs-more-compute-and-monetization.md

Actions

  • POST https://stenobird.com/v1/public/podcasts/the-circuit-5928182/episodes/ep-149-tsmc-q4-25-earnings-openai-needs-more-compute-and-monetization/transcription-requests
    Idempotently request low-priority transcript generation for this episode.
  • GET https://stenobird.com/podcast/the-circuit-5928182/ep-149-tsmc-q4-25-earnings-openai-needs-more-compute-and-monetization.md
    Read the agent-friendly Markdown representation of this episode resource.

Summary

In this episode of The Circuit , Ben Bajarin and Jay Goldberg discuss the launch of Ben's new publication, "The Diligent Stack." The duo then performs a deep dive into TSMC’s recent earnings, analyzing the risks of semiconductor cyclicality, the massive CapEx requirements for the future, and the specific bottlenecks in advanced packaging (CoWoS). Later, they shift focus to OpenAI’s partnership with Cerebras and the introduction of ads to fund massive compute needs. Finally, they break down the latest data on GPU pricing, highlighting the significant premiums hyperscalers charge compared to NeoClouds and the difficulty of tracking pricing for Nvidia's new Grace Blackwell chips.