---
title: "First, let's talk about TPU 8t, which is designed for large-scale training and inference throughput...."
source_name: "Jeff Dean(@JeffDean)"
original_url: "https://x.com/JeffDean/status/2047405389856297387"
canonical_url: "https://www.traeai.com/articles/47c14f52-4efe-408d-bd41-f2b68bee0ba1"
content_type: "tweet"
language: "中文"
score: 5
tags: []
published_at: "2026-04-23T20:00:58+00:00"
created_at: "2026-04-23T23:05:51.933216+00:00"
---

# First, let's talk about TPU 8t, which is designed for large-scale training and inference throughput....

Canonical URL: https://www.traeai.com/articles/47c14f52-4efe-408d-bd41-f2b68bee0ba1
Original source: https://x.com/JeffDean/status/2047405389856297387

## Summary

traeai 为开发者、研究员和内容团队筛选高质量 AI 技术内容，提供摘要、评分、趋势雷达与一键内容产出。

## Key Takeaways

- 
- 
- 

## Content

Title: Jeff Dean on X: "First, let's talk about TPU 8t, which is designed for large-scale training and inference throughput.

The pod size is increased slightly to 9600 chips, and provides ~3X the FP4 performance per pod vs. Ironwood (8t has 121 exaflops/pod vs. 42.5 exaflops/pod for Ironwood).  In https://t.co/XGvH54kuww" / X

URL Source: http://x.com/JeffDean/status/2047405389856297387

Markdown Content:
# Jeff Dean on X: "First, let's talk about TPU 8t, which is designed for large-scale training and inference throughput. The pod size is increased slightly to 9600 chips, and provides ~3X the FP4 performance per pod vs. Ironwood (8t has 121 exaflops/pod vs. 42.5 exaflops/pod for Ironwood). In https://t.co/XGvH54kuww" / X

Don’t miss what’s happening

People on X are the first to know.

[Log in](http://x.com/login)

[Sign up](http://x.com/i/flow/signup)

# [](http://x.com/)

## Post

See new posts

# Conversation

[![Image 2](https://pbs.twimg.com/profile_images/935325968280907776/AcBo6zJc_normal.jpg)](http://x.com/JeffDean)

[Jeff Dean](http://x.com/JeffDean)

[@JeffDean](http://x.com/JeffDean)

First, let's talk about TPU 8t, which is designed for large-scale training and inference throughput. The pod size is increased slightly to 9600 chips, and provides ~3X the FP4 performance per pod vs. Ironwood (8t has 121 exaflops/pod vs. 42.5 exaflops/pod for Ironwood). In addition, the ICI network bandwidth is 2X higher per chip and the scale out datacenter networking is 4X higher per chip. Importantly, this system also offers 2X the performance / watt, continuing the trend of significant energy effiency improvements that we've had for many generations of TPUs (8t offers ~60X the performance/watt as TPU v2).

[![Image 3: Image](https://pbs.twimg.com/media/HGnXaiEa4AAVwQ3?format=jpg&name=small)](http://x.com/JeffDean/status/2047405389856297387/photo/1)

[8:00 PM · Apr 23, 2026](http://x.com/JeffDean/status/2047405389856297387)

·

[7,714 Views](http://x.com/JeffDean/status/2047405389856297387/analytics)

2

3

75

10

## New to X?

Sign up now to get your own personalized timeline!

Sign up with Apple

[Create account](http://x.com/i/flow/signup)

By signing up, you agree to the [Terms of Service](https://x.com/tos) and [Privacy Policy](https://x.com/privacy), including [Cookie Use.](https://help.x.com/rules-and-policies/twitter-cookies)

## Relevant people

*     [![Image 4](https://pbs.twimg.com/profile_images/935325968280907776/AcBo6zJc_normal.jpg)](http://x.com/JeffDean)       [Jeff Dean](http://x.com/JeffDean) [@JeffDean](http://x.com/JeffDean)    Follow   Click to Follow JeffDean  Chief Scientist, Google DeepMind & Google Research. Gemini Lead. Opinions stated here are my own, not those of Google. TensorFlow, MapReduce, Bigtable, ...   

# Trending now

## What’s happening

Trending in United States

Intel

Sports · Trending

Tanner Scott

Politics · Trending

Lisa Murkowski

Trending with [Thom Tillis](http://x.com/search?q=Thom%20Tillis&src=trend_click&vertical=trends)

Sports · Trending

#FlyTheW

[Show more](http://x.com/explore/tabs/for-you)

[Terms of Service](https://x.com/tos)

|

[Privacy Policy](https://x.com/privacy)

|

[Cookie Policy](https://support.x.com/articles/20170514)

|

[Accessibility](https://help.x.com/resources/accessibility)

|

[Ads info](https://business.x.com/en/help/troubleshooting/how-twitter-ads-work.html?ref=web-twc-ao-gbl-adsinfo&utm_source=twc&utm_medium=web&utm_campaign=ao&utm_content=adsinfo)

|

More

© 2026 X Corp.