Amazon Releases TorchNeuron to Challenge Nvidia in AI Chips
· Technology · TechCrunch
Amazon Web Services has released a software tool called TorchNeuron to help AI startups run their applications on Amazon's Trainium chip without rewriting code. The tool uses PyTorch, an open-source framework used by roughly 85% of AI applications, and allows developers to switch configurations with a single line of code. By offering TorchNeuron, Amazon aims to counter the software advantage of Nvidia's CUDA infrastructure, which currently makes Nvidia GPUs the default choice for AI developers. AWS executives argue that porting applications to Trainium gives startups direct access to Amazon's broader cloud infrastructure rather than relying on single-purpose compute providers. The company says Trainium handles both model training and inference on a single platform, helping startups reduce their overall computing costs.
Why it matters
AI startups and developers face lower infrastructure costs and an alternative to Nvidia's dominant ecosystem when building and scaling applications on cloud platforms.
Read the original report — TechCrunch
Join us on Telegram
Breaking news the moment it lands. At 10,000 members we ship the Android app.