AWS Pushes Trainium as a Cost-Effective Alternative to Nvidia GPUs
Amazon Web Services (AWS) is actively encouraging its customers to adopt its in-house Trainium chips over Nvidia’s GPUs, positioning them as a more affordable alternative. According to AWS, Trainium offers comparable performance while reducing costs by 25%.
This push coincided with Nvidia’s GTC 2025 event, where the company was showcasing its latest hardware. Trainium is part of Amazon’s broader silicon strategy, alongside Graviton and Inferentia, designed to support machine learning workloads in the AWS cloud. While Trainium isn’t a direct replacement for Nvidia’s high-end GPUs, it doesn’t have to be—it offers a viable, lower-cost solution for many AI training needs.
AWS’s move reflects a wider industry trend, with cloud providers like Amazon and Google developing their own custom chips to mitigate the high costs and limited availability of Nvidia GPUs.
The key advantage AWS offers is accessibility. By promoting Trainium, the company enables customers to experiment with AI training and inference workloads without long wait times or premium pricing for Nvidia’s in-demand GPUs.
Enterprises used to working with Nvidia’s compute unified device architecture (CUDA) need to think about the cost of switching to a whole new platform like Trainium, but once the switching cost issue has been overcome - and the discounted chip price ought to help make that happen - one needs to ask, "Is Amazon about to disrupt another industry?"
Amazon is shaking up the car market once again, expanding its Amazon Autos platform to include used cars across 68 markets. This bold move is designed to help dealers clear out inventory while making the car-buying process smoother for customers. But it’s also a direct challenge to traditional auto retailers, potentially accelerating the disruption of the industry—especially as used car prices continue to drop.
Originally, Amazon Autos launched with new Hyundai models, allowing customers to browse, finance, and schedule pick-ups directly with local dealers. Now, by adding used cars to the mix, Amazon is doubling down on its mission to simplify car shopping.
The timing couldn’t be more strategic. The used car market has been struggling with price declines due to cyclical challenges, forcing many traditional dealerships to cut back on spending. But while others scale down, Amazon is scaling up - Capital Cycle Theory in action!
Amazon has a history of disrupting industries, and this latest expansion could reshape how people buy cars. As always, the e-commerce giant seems to have perfect timing.
Amazon has secured its first known UK military contract through its satellite venture, Project Kuiper, which won a £670,000 deal with the Ministry of Defence (MoD) and UK Space Command. It will undertake a study on advanced space-based communications systems. The consultancy deal involves Kuiper exploring the use of “translator” satellites for British military purposes.
Kuiper’s research will focus on “translator” satellites, which aim to improve communications between military, government, and private networks, an area also being explored by the US Defence Advanced Research Projects Agency (DARPA).
This move aligns with Western governments' efforts to find alternatives to Musk’s Starlink, as geopolitical concerns over his business ties gain attention. Amazon has already established itself as a key UK government supplier through its Amazon Web Services (AWS) division, and it recently received approval from Ofcom (The Office of Communications - the regulatory and competition authority for the broadcasting, telecommunications and postal industries of the United Kingdom) to provide satellite-based broadband in Britain.
Amazon has been actively engaging with defence officials worldwide seeking additional contracts for its growing satellite business.
The UK is advancing its military satellite programs, including the £5bn GBP Skynet-6 project, while also exploring intelligence-gathering satellite networks. The EU is funding its own sovereign satellite constellation, Iris2, and Taiwan has already engaged with Amazon over Kuiper amid concerns about Elon Musk's Chinese allegiance (China is a huge market for Tesla, so there is a conflict of interest that concerns Taiwan).
These developments reflect growing interest in securing independent and diversified space-based communication networks, all of which is great news for Amazon's project Kuiper.
Very good article, thanks for sharing. I share your view about AMZN being the best option of the mag 7, but also hold a smaller position in GOOG.
You do have a few typos in the valuation part - the values should be $170 (not $1.70), $392 (not $3.92), etc. I assume this may have something to do with UK companies being denominated in pence :)
Having read my analysis of Amazon, and my humble opinion that it is a better choice than Alphabet or Apple, you suggest that you are rethinking your Apple and Alphabet positions.
To clarify, I am not suggesting that Apple and Alphabet shares won't continue to rise. However, the market is currently broken (see: https://rockandturner.substack.com/p/is-the-stock-market-broken) which means share prices often deviate significantly from intrinsic value. For instance, Apple's sales and earnings declined in 2023 and, as of 2024, remained below 2022 levels. Yet, during this same two-year period, its share price doubled. Is this sustainable? In a broken market, who can say?
The core issue lies in the distinction between investing and speculation. Intelligent investing is an academic game of skill and wit, while speculation is a game of chance. I identify as an investor, not a speculator. My point in this article is that, in my view, Amazon represents an intelligent investment for the reasons I’ve outlined. Conversely, I currently see Apple and Alphabet as far more speculative.
Please note that my writing is for informational purposes only. It’s intended to encourage constructive thinking, not to guide your investment decisions directly.
Consider this analogy: betting on red at a casino roulette table gives you less than a 50/50 chance of winning because of the green spaces, yet the payout is only even money. It's a bad bet. Fundamentally, my advice against taking such a bet would be sound. But if you were to bet on red anyway and win, doubling your money, does that make my reasoning flawed? Does it mean your decision was correct?
I trust you understand the point I’m making. Whatever choices you make, I hope your investment decisions work out well for you.
April 10, 2025 - Annual Shareholder Letter
I commend everyone to read this letter from Andy Jassy, Amazon CEO. It says everything you need to know about Amazon, its culture and its future.
https://www.aboutamazon.com/news/company-news/amazon-ceo-andy-jassy-2024-letter-to-shareholders?utm_source=substack&utm_medium=email
AWS Pushes Trainium as a Cost-Effective Alternative to Nvidia GPUs
Amazon Web Services (AWS) is actively encouraging its customers to adopt its in-house Trainium chips over Nvidia’s GPUs, positioning them as a more affordable alternative. According to AWS, Trainium offers comparable performance while reducing costs by 25%.
This push coincided with Nvidia’s GTC 2025 event, where the company was showcasing its latest hardware. Trainium is part of Amazon’s broader silicon strategy, alongside Graviton and Inferentia, designed to support machine learning workloads in the AWS cloud. While Trainium isn’t a direct replacement for Nvidia’s high-end GPUs, it doesn’t have to be—it offers a viable, lower-cost solution for many AI training needs.
AWS’s move reflects a wider industry trend, with cloud providers like Amazon and Google developing their own custom chips to mitigate the high costs and limited availability of Nvidia GPUs.
The key advantage AWS offers is accessibility. By promoting Trainium, the company enables customers to experiment with AI training and inference workloads without long wait times or premium pricing for Nvidia’s in-demand GPUs.
Enterprises used to working with Nvidia’s compute unified device architecture (CUDA) need to think about the cost of switching to a whole new platform like Trainium, but once the switching cost issue has been overcome - and the discounted chip price ought to help make that happen - one needs to ask, "Is Amazon about to disrupt another industry?"
I wouldn't bet against them.
CAPITAL CYCLE THEORY IN ACTION
Amazon is shaking up the car market once again, expanding its Amazon Autos platform to include used cars across 68 markets. This bold move is designed to help dealers clear out inventory while making the car-buying process smoother for customers. But it’s also a direct challenge to traditional auto retailers, potentially accelerating the disruption of the industry—especially as used car prices continue to drop.
Originally, Amazon Autos launched with new Hyundai models, allowing customers to browse, finance, and schedule pick-ups directly with local dealers. Now, by adding used cars to the mix, Amazon is doubling down on its mission to simplify car shopping.
The timing couldn’t be more strategic. The used car market has been struggling with price declines due to cyclical challenges, forcing many traditional dealerships to cut back on spending. But while others scale down, Amazon is scaling up - Capital Cycle Theory in action!
Amazon has a history of disrupting industries, and this latest expansion could reshape how people buy cars. As always, the e-commerce giant seems to have perfect timing.
AMAZON BREAKING NEWS
Amazon has secured its first known UK military contract through its satellite venture, Project Kuiper, which won a £670,000 deal with the Ministry of Defence (MoD) and UK Space Command. It will undertake a study on advanced space-based communications systems. The consultancy deal involves Kuiper exploring the use of “translator” satellites for British military purposes.
Kuiper’s research will focus on “translator” satellites, which aim to improve communications between military, government, and private networks, an area also being explored by the US Defence Advanced Research Projects Agency (DARPA).
This move aligns with Western governments' efforts to find alternatives to Musk’s Starlink, as geopolitical concerns over his business ties gain attention. Amazon has already established itself as a key UK government supplier through its Amazon Web Services (AWS) division, and it recently received approval from Ofcom (The Office of Communications - the regulatory and competition authority for the broadcasting, telecommunications and postal industries of the United Kingdom) to provide satellite-based broadband in Britain.
Amazon has been actively engaging with defence officials worldwide seeking additional contracts for its growing satellite business.
The UK is advancing its military satellite programs, including the £5bn GBP Skynet-6 project, while also exploring intelligence-gathering satellite networks. The EU is funding its own sovereign satellite constellation, Iris2, and Taiwan has already engaged with Amazon over Kuiper amid concerns about Elon Musk's Chinese allegiance (China is a huge market for Tesla, so there is a conflict of interest that concerns Taiwan).
These developments reflect growing interest in securing independent and diversified space-based communication networks, all of which is great news for Amazon's project Kuiper.
Very good article, thanks for sharing. I share your view about AMZN being the best option of the mag 7, but also hold a smaller position in GOOG.
You do have a few typos in the valuation part - the values should be $170 (not $1.70), $392 (not $3.92), etc. I assume this may have something to do with UK companies being denominated in pence :)
Thank you for pointing out the formatting of the money values. It has been corrected now.
Thanks for the report, makes me think about my existing Apple and Alphabet position :-/
Having read my analysis of Amazon, and my humble opinion that it is a better choice than Alphabet or Apple, you suggest that you are rethinking your Apple and Alphabet positions.
To clarify, I am not suggesting that Apple and Alphabet shares won't continue to rise. However, the market is currently broken (see: https://rockandturner.substack.com/p/is-the-stock-market-broken) which means share prices often deviate significantly from intrinsic value. For instance, Apple's sales and earnings declined in 2023 and, as of 2024, remained below 2022 levels. Yet, during this same two-year period, its share price doubled. Is this sustainable? In a broken market, who can say?
The core issue lies in the distinction between investing and speculation. Intelligent investing is an academic game of skill and wit, while speculation is a game of chance. I identify as an investor, not a speculator. My point in this article is that, in my view, Amazon represents an intelligent investment for the reasons I’ve outlined. Conversely, I currently see Apple and Alphabet as far more speculative.
Please note that my writing is for informational purposes only. It’s intended to encourage constructive thinking, not to guide your investment decisions directly.
Consider this analogy: betting on red at a casino roulette table gives you less than a 50/50 chance of winning because of the green spaces, yet the payout is only even money. It's a bad bet. Fundamentally, my advice against taking such a bet would be sound. But if you were to bet on red anyway and win, doubling your money, does that make my reasoning flawed? Does it mean your decision was correct?
I trust you understand the point I’m making. Whatever choices you make, I hope your investment decisions work out well for you.