The most fascinating parallel here isn't AMD vs Intel, but how AMD's chiplet architecture mirrors the broader tech industry's move toward specialized, interconnected systems. Just as the first cities weren't built as single megastructures, but as interconnected neighborhoods, the future of computing might favor AMD's modular approach over monolithic designs.
Very nice read.. well researched and well written. Loved the historical background, and the analysis of the business. Thank you for sharing this article. Long AMD, entered around $80.
NVIDIA has a competitive edge because it is not just a chip designer. NVIDIA provides a suite of software tools and libraries, including CUDA, TensorRT, and the NVIDIA Deep Learning Institute, enabling developers to leverage GPU power efficiently. NVIDIA often bundles its hardware with proprietary software, optimized for its GPUs, making it difficult for customers to switch to competitors.
NVIDIA has also reinforced its position through strategic acquisitions. The Mellanox Acquisition in 2019 is a case in point. This acquisition has been crucial for NVIDIA's vertical integration strategy.
AMD doesn't have the same depth to compete. It isn't as simple as just switching out an NVIDIA chip for an AMD chip as it was in the early days of computing. NVIDIA is very much entrenched.
You should also consider that successful hardware companies often thrive by creating ecosystems that attract and support software development.
Back in the day when Intel dominated, it built an open ecosystem for software developers and established an architecture review board to gather feedback, ensuring that software being developed by third parties ran optimally on Intel architecture. If software ran best on Intel architecture, consumers preferred PC's with Intel chips, creating a competitive advantage.
More recently NVIDIA collaborated closely with EA and other game developers - NVIDIA positioned its GPUs as best in class architecture for the game industry, creating a robust competitive advantage that helped propel it to where it is today.
Does AMD do anything like this?
Then there is the question of whether silicon chips will soon be replaced entirely. Silicon has been scaled about as small as it can, which is why the chiplet or stacked design has become necessary. This creates a whole load of new issues plus there are heat and energy efficiency issues to deal with. Silicon may be coming to the end of its useful life in this industry. As alternatives, there are graphene chips coming and quantum computing which will take us to the next level. How will the kings of the silicon chips respond? This is a major risk.
Another risk is the huge dependency on TSMC. What happens if China reclaims Taiwan and uses all of its capacity to manufacture Chinese chips, maybe even refusing to service US firms - there is no love lost between the two countries and the US has sought to impede China's access to cutting edge chips, so retaliatory action can't be ruled out. If AMD can't access TSMC, what then?
As you state in your article, companies such as Coca-Cola have no R&D expense, no major ongoing CAPEX, nothing changes and so it enjoys durable earnings. The chip industry is constantly changing, requires vast amounts of capital just to remain relevant, and is always in danger of being disrupted (as Intel learned the hard way).
This industry is just such a difficult one to play in.
The most fascinating parallel here isn't AMD vs Intel, but how AMD's chiplet architecture mirrors the broader tech industry's move toward specialized, interconnected systems. Just as the first cities weren't built as single megastructures, but as interconnected neighborhoods, the future of computing might favor AMD's modular approach over monolithic designs.
Completely agree with that, thanks for the contribution!
Good work, Oguz! I was planning a Quick Look post on AMD. I will link to your article so my people can read up further.
Of course!
Very nice read.. well researched and well written. Loved the historical background, and the analysis of the business. Thank you for sharing this article. Long AMD, entered around $80.
Glad you liked it PK!
thank you very much aporta mucho valor lo agradesco mucho saludos
Brilliant analysis, I think AMD is a great buy at current prices and we will see $200+ per share at some point in the next 12-18 months...
I think so, it has a great potential, it just needs to execute.
There is something missing from this analysis.
NVIDIA has a competitive edge because it is not just a chip designer. NVIDIA provides a suite of software tools and libraries, including CUDA, TensorRT, and the NVIDIA Deep Learning Institute, enabling developers to leverage GPU power efficiently. NVIDIA often bundles its hardware with proprietary software, optimized for its GPUs, making it difficult for customers to switch to competitors.
NVIDIA has also reinforced its position through strategic acquisitions. The Mellanox Acquisition in 2019 is a case in point. This acquisition has been crucial for NVIDIA's vertical integration strategy.
AMD doesn't have the same depth to compete. It isn't as simple as just switching out an NVIDIA chip for an AMD chip as it was in the early days of computing. NVIDIA is very much entrenched.
You should also consider that successful hardware companies often thrive by creating ecosystems that attract and support software development.
Back in the day when Intel dominated, it built an open ecosystem for software developers and established an architecture review board to gather feedback, ensuring that software being developed by third parties ran optimally on Intel architecture. If software ran best on Intel architecture, consumers preferred PC's with Intel chips, creating a competitive advantage.
More recently NVIDIA collaborated closely with EA and other game developers - NVIDIA positioned its GPUs as best in class architecture for the game industry, creating a robust competitive advantage that helped propel it to where it is today.
Does AMD do anything like this?
Then there is the question of whether silicon chips will soon be replaced entirely. Silicon has been scaled about as small as it can, which is why the chiplet or stacked design has become necessary. This creates a whole load of new issues plus there are heat and energy efficiency issues to deal with. Silicon may be coming to the end of its useful life in this industry. As alternatives, there are graphene chips coming and quantum computing which will take us to the next level. How will the kings of the silicon chips respond? This is a major risk.
Another risk is the huge dependency on TSMC. What happens if China reclaims Taiwan and uses all of its capacity to manufacture Chinese chips, maybe even refusing to service US firms - there is no love lost between the two countries and the US has sought to impede China's access to cutting edge chips, so retaliatory action can't be ruled out. If AMD can't access TSMC, what then?
As you state in your article, companies such as Coca-Cola have no R&D expense, no major ongoing CAPEX, nothing changes and so it enjoys durable earnings. The chip industry is constantly changing, requires vast amounts of capital just to remain relevant, and is always in danger of being disrupted (as Intel learned the hard way).
This industry is just such a difficult one to play in.
Food for thought
Watch out the word 10x.
Do you mean P/E 10?
Thank you very much😊
But where is AMD's advantage when Alphabet, Amazon etc. will also start design their own chip?
It means their customers will be their competitors in the future.
Fantastic analysis and a wonderful read. Thank you for sharing!