TensorWave partners to bring fast AI inference to AMD cloud
TensorWave, a leading innovator in cloud computing, has announced a groundbreaking partnership with MK1, a startup founded by early Neuralink engineers, including co-founder Paul Merolla.
This strategic alliance aims to revolutionize the AI industry by democratizing access to high-performance AI inference using AMD hardware.
Neuralink co-founding team member, Paul Merolla, and his new startup are on a mission — to help enterprises get the most performance out of their valuable compute resources.
Recognizing the limitations posed by a potential GPU monopoly in the market, MK1 set its sights on AMD’s emerging capabilities. AMD’s software system, although primarily consumer-focused, presented a unique opportunity for MK1 to develop software tailored specifically for AMD’s cloud-native hardware.
“Seeing the potential of AMD hardware to catch up to and even surpass existing solutions was a game-changer for us,” said Paul Merolla. “This partnership with TensorWave aligns perfectly with our goals. Together, we are building the first competitive alternative to run GenAI at scale in the cloud, independent of NVIDIA hardware.”
Bringing choice in AI inference
This partnership is set to make a significant impact in the AI industry by introducing a strong alternative to NVIDIA’s dominance. “For the first time, AMD hardware is a formidable competitor for AI-accelerated workloads,” Merolla explains. “Many in the industry have yet to grasp the benefits this will bring to the ecosystem, simply by giving customers a choice.”
With 70% of AI workloads dedicated to inference, optimizing performance is crucial. Merolla believes that this is the moment for AMD to compete effectively. “Anyone who sees the trajectory and is willing to make the bet, this is the time,” he asserts. “With MK1 software running on TensorWave’s cloud, we are offering a solution that makes the most of every dollar spent.”
Darrick Horton, CEO of TensorWave, shares Merolla’s enthusiasm for the partnership. “We are getting closer to outperforming existing solutions every day,” said Horton. “Our goal is to democratize access to high-performance AI inference, making it available and affordable for everyone.”
This collaboration brings together MK1’s cutting-edge software with TensorWave’s robust cloud infrastructure, providing a seamless and powerful option for customers. “Our combined efforts are showing great signs of success,” added Horton. “We are ready to tackle real-world workloads and deliver a service that is both simple and highly effective.”
The partnership between MK1 and TensorWave is set to disrupt the AI industry, offering a more user-friendly and efficient alternative to existing offerings. With a focus on delivering competitive performance for large language models (LLMs) and other inference tasks, this collaboration promises to bring significant advancements to the field.
MK1 is a highly-technical team dedicated to making AI inference efficient. Our current product is optimized inference software that is natively cross-platform, user-friendly, and delivers high performance for a variety of real-world workloads. MK1 software operates on over a thousand GPUs in the cloud, currently serving over 15 trillion tokens per month for clients.
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
