New Technology / Ai Development
Track AI development, model progress, product releases, infrastructure shifts and strategic technology signals across the artificial intelligence sector.
Inside the US-China AI Chip Trade War
Topic
US-China AI Chip Trade War
Key insights
- AMD reported its quarterly results
- Revenue grew 34% in the fourth quarter
- The company is expecting top line to grow about 32% at the midpoint of its range this quarter
- AMD faces competition from Nvidia and ASIC projects from Meta, Microsoft, and Google
- Googles TPUs are gaining traction and have impressive specs
- Chinese customers want AMD chips but lack domestic alternatives
Perspectives
Discussion on the competitive landscape of the chip industry.
Chris Miller
- Highlights AMDs revenue growth of 34% indicating strong market demand
- Warns of significant competition from Nvidia and Googles TPUs
- Claims China lacks domestic manufacturing capacity for necessary chips
- Argues that TSMC is rapidly advancing its production capabilities in the US
- Proposes that Chinese tech companies will continue to seek imports for chips
- Questions the long-term strategy of the Chinese government regarding GPU production
Neutral / Shared
- Notes ongoing debates in Washington about chip sales to China
- Acknowledges the complexity of TSMCs expansion in the US
Metrics
revenue_growth
34%
quarterly revenue growth for AMD
Indicates strong demand and performance in a competitive market.
Revenue grew 34% in the fourth quarter.
expected_growth
32%
expected revenue growth for the next quarter
Reflects continued optimism about AMD's market position.
The company is expecting top line to grow about 32% at the midpoint of its range this quarter.
production_gap
30 times
comparison of AI accelerator production between the US plus Taiwan and China
Highlights the significant technological gap that China faces.
US plus Taiwan are going to produce something like 30 times the number of AI accelerators quality adjusted as China is.
production_percentage
25 or 30%
TSMC's advanced production in the US by the mid-2030s
This indicates TSMC's commitment to US manufacturing and its strategic shift.
they're going to have something like 25 or 30% of their advanced production in the US by the mid-2030s.
gap_ratio
30 to one ratio
The technology gap between TSMC and China in chip manufacturing
A significant gap that highlights the challenges China faces in catching up.
And that's how you get this 30 to one gap, which is really pretty extraordinary.
Key entities
Timeline highlights
00:00–05:00
AMD's revenue growth of 34% indicates strong market demand, but competition from Nvidia and Google's TPUs poses significant challenges.
- AMD reported its quarterly results
- Revenue grew 34% in the fourth quarter
- The company is expecting top line to grow about 32% at the midpoint of its range this quarter
- AMD faces competition from Nvidia and ASIC projects from Meta, Microsoft, and Google
- Googles TPUs are gaining traction and have impressive specs
- Chinese customers want AMD chips but lack domestic alternatives
05:00–10:00
TSMC is rapidly advancing its chip production capabilities in the US, aiming for 25-30% of its advanced production by the mid-2030s, while China struggles with significant technology gaps.
- China is trying to make advanced ships using second rate equipment
- TSMC is scaling up faster, building more fabs, and advancing its process technology
- Chinas gap in chip technology is significant and not easily solvable in a few years
- China has been competitive in sectors like electric vehicles, but chips require higher quality
- TSMC aims to have 25 or 30% of its advanced production in the US by the mid-2030s
- TSMCs expansion in the US has faced challenges, including higher costs and workforce development
10:00–15:00
Investors are funding new inference specific architectures, indicating confidence in the market's capacity for both training and inference chips.
- There are acquisitions happening, but they are not fully traditional acquisitions
- New startups in the space are raising substantial sums of money
- The venture space is willing to fund new inference specific architectures
- Investors believe there is a pathway for inference specific chips
- The market is vast enough for both training specific and inference specific chips