Amazon challenges NVIDIA with cheaper AI chips that use less power. The company developed Trainium processors to compete against expensive Blackwell graphics cards. NVIDIA sells each Blackwell chip for 30,000 to 40,000 dollars. Most tech companies cannot afford large quantities of these costly processors. Amazon believes its chips deliver better value for the money spent.
Trainium2 processors perform four times faster than the original version. The new chips cost 40 percent less than competing AI hardware from other makers. Amazon plans to release Trainium3 chips that will cut energy use by half. These advanced processors will double the speed of current Trainium2 models. The company already uses its chips to train Claude Opus 4 artificial intelligence systems.
NVIDIA maintains the performance crown despite Amazon's cost advantages. Jensen Huang dismisses custom chip threats to his company's market leadership. Amazon Web Services promotes Trainium processors to cloud computing customers. Other companies like Alphabet and Marvell also create specialized AI hardware. The chip race continues as demand for artificial intelligence computing grows rapidly.
Trainium2 processors perform four times faster than the original version. The new chips cost 40 percent less than competing AI hardware from other makers. Amazon plans to release Trainium3 chips that will cut energy use by half. These advanced processors will double the speed of current Trainium2 models. The company already uses its chips to train Claude Opus 4 artificial intelligence systems.
NVIDIA maintains the performance crown despite Amazon's cost advantages. Jensen Huang dismisses custom chip threats to his company's market leadership. Amazon Web Services promotes Trainium processors to cloud computing customers. Other companies like Alphabet and Marvell also create specialized AI hardware. The chip race continues as demand for artificial intelligence computing grows rapidly.
