OctoML is a machine learning acceleration platform. It is a Seattle-based startup. Their goal is to optimize businesses and deploy machine learning models. They have raised $85 million in a Series C investment, headed by Tiger Global Management. This round included previous investors. Additionally, Madrona Venture Group and Amplify Partners joined them. They brought up the company’s total funding to $132 million. It includes a $28 million Series B round disclosed earlier this year.
OctoML : The Founders
The founding team includes Luis Ceze, Tianqi Chen, Jason Knight, Jared Roesch, and Thierry Moreau. They co-founded the company. They established the Apache TVM open-source machine learning compiler framework. Amazon, Microsoft, and Facebook are among the companies that use TVM. TVM’s ability to automatically optimize machine learning models. Moreover, it runs them on any hardware that is included in OctoML.
Hardware Partners for Machine Learning Platform
Ceze has signed up several hardware partners. These include Qualcomm, AMD, and Arm, since raising its Series A financing. Besides, the company has collaborated with Microsoft on a project. It involves large-scale video content moderation deployment. After using its service, the startup claims that its subscribers. These include several Global 100 companies like Toyota. Furthermore, Toyota reports a 2 to 10 times improvement in their ML model performance.
The company had just started onboarding some early adopters to its SaaS platform. When it raised its Series B earlier this year. That service hasn’t yet achieved wide availability. Still, OctoML is now working with a more significant number of customers. Mainly, they are focusing on helping them succeed on its platform.
Strategies and Opportunites for OctoML
As models become more complex, installing them on the cloud becomes more expensive. According to Ceze, a system that can optimize these models immediately saves money for clients. It’s not just a cost issue. it’s also a sustainability issue. It has an impact at scale if you create something twice as fast on the same hardware while using half the energy.
Large cloud providers are increasingly hitting capacity restrictions for installing high-end GPUs. So, transferring their models to a different GPU or even a CPU is an advantage.
Ceze stated that the company didn’t need to raise more funds. However, management chose to be opportunistic even though the runway was still healthy. They focused on their hardware enablement, SaaS business acceleration, and cloud enablement prospects.
The new money will be used to snowball. Moreover, this growth is in various engineering and sales areas. And, they are capitalizing on this opportunity as it adds new clients. OctoML also intends to expand its ecosystem of partners.
Future Approach and Forecasting of Machine Learning by OctoML
OctoML is causing a significant shift in how companies construct next-generation AI models. Moreover, they are aiming for applications. OctoML is providing a consistent deployment lifecycle for customers. Moreover, it is available all over the ML hardware suppliers of ML. They rely on the basic fact of growth. And that is making machine learning development more cost-effective and accessible to developers.