Anthropic's prospectus discloses mixed use of four types of chips, potentially weakening Nvidia's compute pricing power
Anthropic disclosed in its S-1 prospectus that it uses four types of chips—Amazon Trainium, Google TPU, Nvidia GPU, and AMD GPU—to train Claude and provide services, and integrates compute from multiple parties into a unified, flexibly allocatable pool through software. Analysts believe that customers able to shift tasks across the four platforms have stronger bargaining power, which could weaken Nvidia's pricing power; however, Moorhead noted that Anthropic is still continuing to invest heavily in Nvidia compute, and Brad Gastwirth of Circular Technology also believes it does not constitute a major threat in the short term.
Body · AI translation
Cailianshe Financial News Agency
Open the APP
A matter of computing power pricing power! Nvidia's biggest nightmare is hidden in Anthropic's prospectus
Cailianshe Bian Chun | Editor: Lin Qi 2026-10-28 11:19
① Top US AI lab Anthropic is using multiple chips to develop its business; ② Analysts point out that this flexibility may gradually reduce large AI companies' dependence on Nvidia and strengthen their bargaining power, thereby weakening Nvidia's pricing power.
Cailianshe, October 28 (Editor: Bian Chun) — Top US AI lab Anthropic is using multiple chips to develop its business — a diversified strategy that could ultimately weaken Nvidia's dominance in the AI computing power sector.
In the S-1 prospectus published on Monday, Anthropic said it "integrates computing resources from multiple strategic partners into a unified, flexibly allocatable pool," and uses software technology to flexibly allocate tasks based on the performance advantages of each chip in different scenarios.
Anthropic said it uses four types of chips to train Claude and provide related services: Amazon Trainium, Google TPU, Nvidia GPU and AMD GPU.
Nvidia's pricing power may be weakened
"What could really be affected is Nvidia's pricing power, because a customer that can shift tasks across four platforms is in a completely different negotiating position from one that cannot," said Patrick Moorhead, founder and CEO of Moor Insights & Strategy.
IDC analyst Ashish Nadkarni said the multi-chip strategy can also make Anthropic less susceptible to fluctuations in Nvidia's chip supply.
Anthropic is not the only leading AI company diversifying its computing power supply. According to the latest media reports, OpenAI plans to purchase Google AI chips. The financial terms and other details of the purchase are not yet clear. OpenAI has long been one of Nvidia's most important customers, and Nvidia has also invested tens of billions of dollars in the startup.
Breaking Nvidia's dominance is not easy
Moorhead said that although Anthropic's mixing of multiple chips is not new in itself, the company is now pitching it to investors as a core advantage.
Moorhead also noted that Anthropic is still continuously investing heavily in Nvidia's computing power.
Gil Luria, head of technology research at D.A. Davidson, said that while major companies like OpenAI may adopt a multi-chip strategy, for smaller companies, devoting so many resources to this is not worth it.
Brad Gastwirth, global head of research and market intelligence at Circular Technology, said Anthropic's strategy will not pose a major threat in the short term. He pointed out that having more chip options does not eliminate Nvidia's advantages in performance, networking and software.
Citadel CEO Ken Griffin has warned that if supply exceeds demand, it could lead to a decline in computing power rental prices. "The marginal cost of running a GPU is very low, just like airlines. Once there is overcapacity, prices will go into free fall," Griffin said last week.
Foreign institutional views - US
US stock market developments
Cailianshe statement: The content of this article is for reference only and does not constitute investment advice. Investors who act on it do so at their own risk.
Business cooperation
Popular unlocks
[Gold-digging industry leader] Robotics + industrial mother machines, the micro ball screws used in dexterous hands have obtained customer designation, and the company is customizing joint module products for humanoid robots for domestic enterprises. This company's second-generation machine tool production capacity will exceed 40 units/month
Robotics + industrial mother machines, the micro ball screws used in dexterous hands have obtained customer designation, and the company is customizing joint module products for humanoid robots for domestic enterprises. The second-generation machine tool production capacity will exceed 40 units/month, and this company's aircraft electric drive products have completed delivery.
Positioned at the entry point for physical AI tactile data, this company has built a dedicated robot electronic skin R&D platform, completed iterative upgrades of core process equipment, and launched flexible electronic skin products across multiple technology routes, sending samples to multiple downstream customers for testing, with some customers already placing sample orders.
Related reading
No rate hike tonight? Fed decision day arrives: Wall Street closely watches Warsh's remarks!
Cailianshe Xiaoxiang 2026-10-28 09:02
① According to the schedule, the Federal Reserve will announce its October interest rate decision at 2 a.m. Beijing time on Thursday, and Fed Chair Warsh will hold a regular press conference at 2:30. ② The market currently widely expects that after raising rates by 25 basis points in September, the Fed will likely keep rates unchanged at this meeting. The CME FedWatch tool shows that the probability of a Fed rate hike tonight is only about 7%.
① In the second half of the year, market sentiment shifted, and share prices of AI component manufacturers such as semiconductors and memory began to come under pressure, with investors doubting whether massive spending commitments can be sustained; ② One positive factor for memory stocks is that valuations are not expensive. Micron's forward price-to-earnings ratio for the next 12 months is 5 times, the second-lowest P/E ratio among S&P 500 stocks.
Cailianshe Liu Rui 2026-10-27 17:16
① Over the past three weeks, changes in the semiconductor sector's performance can explain about 54% of the performance differences across various assets, far exceeding the explanatory power of oil prices; ② The South Korean stock market is particularly sensitive to the AI trading trend, and memory chip stocks now account for nearly half the weight of the KOSPI.