TensorWave's fresh $43 million funding will allow it to enhance its datacenter with AMD's MI300X accelerators, launching a new inference platform, Manifest, by Q4.
Recent successes of other cloud providers motivated TensorWave to leverage AMD's Instinct accelerators over Nvidia's, showcasing a commitment to a different direction in generative AI.
AMD's MI300X offers higher performance and greater memory capacity than Nvidia's H100, making it ideal for running larger models without compression, leading to widespread adoption.
With next-gen MI325X accelerators on the horizon, AMD is set to dominate the market in terms of memory capacity, further increasing its appeal to cloud providers.
Collection
[
|
...
]