Silicon Valley genius investor Peter Thiel's hedge fund Thiel Macro has recently made a big move - completely liquidating its holdings in NVIDIA ( NVDA ) for Q3, fully reallocating to Microsoft ( MSFT ). This action hides the undercurrents in the AI chip market.
How long can Nvidia dominate?
On the surface, Nvidia holds over 80% of the revenue share in the AI accelerator market, which can be considered a monopoly position. However, the fact that Thiel chose to exit indicates that he has seen potential concerns.
Threats come from two directions: First, the performance gap of AMD's MI series chips is narrowing. Second, tech giants are developing their own chips. Google, Amazon, and Meta are all working on custom AI chips to reduce reliance on NVIDIA, and OpenAI even plans to deploy MI450 by the end of 2026.
Does it seem dangerous? But analysts provided an unexpected answer: the cost of self-developed chips is actually higher. The reason is simple—NVIDIA spent nearly 20 years building the CUDA ecosystem, which includes a complete toolchain of pre-trained models, development frameworks, code libraries, and more. Self-developed chips lack this set of tools, and companies have to build from scratch, making the development costs actually more expensive than purchasing NVIDIA GPUs.
Therefore, the market expects Nvidia to maintain a 70-90% share in the AI accelerator market, which is projected to grow at an annual rate of 29% until 2033. Nvidia's revenue is expected to grow by 37% annually over the next three years, and its current valuation at 44 times PE is relatively moderate.
Teal may have withdrawn too early.
Why turn to Microsoft?
Microsoft is the largest enterprise software company in the world and the second-largest public cloud player, monetizing AI through these two advantages.
The action is very clear: integrate generative AI Copilot into popular products like Microsoft 365 to provide comprehensive cloud AI services. CEO Nadella revealed that the adoption rate of Microsoft 365 Copilot by users exceeds that of any new product, with 90% of Fortune 500 companies using it.
In the cloud computing sector, although Microsoft's sales growth has dropped to 28%, its market share remains unchanged. The key issue is that Microsoft faces capacity bottlenecks—after the data center scale doubles in the next two years, its share is expected to expand.
The market expects Microsoft to have a 14% annual revenue increase over the next three years. This estimate may be conservative, as enterprise software and cloud spending are expected to grow at annual rates of 12% and 20%, respectively, until 2030.
At the current PE valuation of 34 times, Microsoft's PEG ratio is 2.4 (typically >2 is considered expensive). However, this is lower than the 3-year average of 2.6 and the 5-year average of 2.5, making it relatively a good entry point.
Who wins and who loses?
Tiel's reallocation essentially reflects the shift in topics in the AI era: the value transfer from chip hardware to software applications. Nvidia remains strong, but Microsoft, as an AI software and cloud service provider, is in a period of explosive growth.
However, this wave of operations also illustrates one point: even the bets of top investors do not necessarily represent absolute correctness. Nvidia has its defenses, and Microsoft has its risks. The key is how you understand the long-term landscape of the AI industry.
View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
Thiel closes all positions in Nvidia and turns to Microsoft: the AI chip war behind it
Silicon Valley genius investor Peter Thiel's hedge fund Thiel Macro has recently made a big move - completely liquidating its holdings in NVIDIA ( NVDA ) for Q3, fully reallocating to Microsoft ( MSFT ). This action hides the undercurrents in the AI chip market.
How long can Nvidia dominate?
On the surface, Nvidia holds over 80% of the revenue share in the AI accelerator market, which can be considered a monopoly position. However, the fact that Thiel chose to exit indicates that he has seen potential concerns.
Threats come from two directions: First, the performance gap of AMD's MI series chips is narrowing. Second, tech giants are developing their own chips. Google, Amazon, and Meta are all working on custom AI chips to reduce reliance on NVIDIA, and OpenAI even plans to deploy MI450 by the end of 2026.
Does it seem dangerous? But analysts provided an unexpected answer: the cost of self-developed chips is actually higher. The reason is simple—NVIDIA spent nearly 20 years building the CUDA ecosystem, which includes a complete toolchain of pre-trained models, development frameworks, code libraries, and more. Self-developed chips lack this set of tools, and companies have to build from scratch, making the development costs actually more expensive than purchasing NVIDIA GPUs.
Therefore, the market expects Nvidia to maintain a 70-90% share in the AI accelerator market, which is projected to grow at an annual rate of 29% until 2033. Nvidia's revenue is expected to grow by 37% annually over the next three years, and its current valuation at 44 times PE is relatively moderate.
Teal may have withdrawn too early.
Why turn to Microsoft?
Microsoft is the largest enterprise software company in the world and the second-largest public cloud player, monetizing AI through these two advantages.
The action is very clear: integrate generative AI Copilot into popular products like Microsoft 365 to provide comprehensive cloud AI services. CEO Nadella revealed that the adoption rate of Microsoft 365 Copilot by users exceeds that of any new product, with 90% of Fortune 500 companies using it.
In the cloud computing sector, although Microsoft's sales growth has dropped to 28%, its market share remains unchanged. The key issue is that Microsoft faces capacity bottlenecks—after the data center scale doubles in the next two years, its share is expected to expand.
The market expects Microsoft to have a 14% annual revenue increase over the next three years. This estimate may be conservative, as enterprise software and cloud spending are expected to grow at annual rates of 12% and 20%, respectively, until 2030.
At the current PE valuation of 34 times, Microsoft's PEG ratio is 2.4 (typically >2 is considered expensive). However, this is lower than the 3-year average of 2.6 and the 5-year average of 2.5, making it relatively a good entry point.
Who wins and who loses?
Tiel's reallocation essentially reflects the shift in topics in the AI era: the value transfer from chip hardware to software applications. Nvidia remains strong, but Microsoft, as an AI software and cloud service provider, is in a period of explosive growth.
However, this wave of operations also illustrates one point: even the bets of top investors do not necessarily represent absolute correctness. Nvidia has its defenses, and Microsoft has its risks. The key is how you understand the long-term landscape of the AI industry.