Jotunn8
ENABLING AI DATACENTER INFERENCE AT SCALE
COST / QUERY <$0.002
DeepSeek, LLAMA, GPT... all parameter sizes supported
Our Generative AI Offer
Jotunn8
3.2 Petaflops (dense) tailormade for easy, CUDA-free Generative AI use. Large LLMs, like Llama3-405B, GPT-4 or DeepSeek-R1, can quickly be implemented and deployed using standard software.
Tyr4
3.2 Petaflops (sparse) for any AI application, including advanced Generative AI completely CUDA-free
Tyr2
1.6 Petaflops (sparse) for easy, CUDA-free AI, including Generative AI. All completely CUDA-free
Tyr1
800 Teraflops (sparse) for any AI application, including Generative AI. All completely CUDA-free
Autonomous Driving / ADAS
CUDA-Free High-Performance
Low-Power Computing
for the Vehicle
3.2 PetaFlops / 100 W
Any Algorithm
Any Host Processor
Fully Programmable
Our AD/ADAS Offer
Tyr4
3.2 Petaflops for any AD / ADAS application. All completely CUDA-free.
Tyr2
1.6 Petaflops for any AD/ADAS application. All completely CUDA-free.
Tyr1
800 Teraflops for any AD / ADAS application. All completely CUDA-free.
