Google + Amazon join forces to attack Nvidia’s moat ⚔️
According to TechCrunch:
🔹 On October 20, Google open-sourced TorchTPU, allowing PyTorch applications to run on Google TPUs with a simple config change—no code rewrite needed.
🔹 The day before, Amazon released TorchNeuron, enabling PyTorch applications to run on its in-house Trainium chips. It also plans to submit code to the main PyTorch project, reducing the default reliance on Nvidia GPUs.
🔹 PyTorch is estimated to power around 85% of AI applications, while Nvidia’s CUDA ecosystem has long been its biggest barrier to competition.
The report also notes that developers won’t abandon Nvidia en masse overnight, but the cloud giants have begun tearing down the walls.
How many more years do you think Nvidia can hold onto its computing monopoly? Will intensifying competition in AI chips be bullish or bearish for compute-power tokens and AI stocks? Join the discussion in the comments 👇
#AI #英伟达 #AI芯片
For reference only; this is not investment advice.
According to TechCrunch:
🔹 On October 20, Google open-sourced TorchTPU, allowing PyTorch applications to run on Google TPUs with a simple config change—no code rewrite needed.
🔹 The day before, Amazon released TorchNeuron, enabling PyTorch applications to run on its in-house Trainium chips. It also plans to submit code to the main PyTorch project, reducing the default reliance on Nvidia GPUs.
🔹 PyTorch is estimated to power around 85% of AI applications, while Nvidia’s CUDA ecosystem has long been its biggest barrier to competition.
The report also notes that developers won’t abandon Nvidia en masse overnight, but the cloud giants have begun tearing down the walls.
How many more years do you think Nvidia can hold onto its computing monopoly? Will intensifying competition in AI chips be bullish or bearish for compute-power tokens and AI stocks? Join the discussion in the comments 👇
#AI #英伟达 #AI芯片
For reference only; this is not investment advice.