Microsoft is aiming to provide a select number of standout startups with free access to high-performance computing resources from Azure’s cloud for developing artificial intelligence models.
The company announced an update to its Microsoft for Startups Founders Hub program, now including free access to the AI infrastructure from Azure, specifically advanced GPU virtual machine sets based on Nvidia processors for training and running generative models, including large language models like ChatGPT.
Accelerator company Y Combinator and its community of startup founders are among those granted access to the private preview sets.
Microsoft referred to Y Combinator as the ideal initial partner due to its successful track record with early-stage startups.
The company stated, “We work closely with Y Combinator to prioritize requests from the current cohort as part of our initial preview, focusing on tasks such as training and use case tuning that drive innovation.”
This isn’t Microsoft’s first attempt to garner support from startups under the Y Combinator umbrella.
In 2015, the company provided $500,000 in Azure credits to Y Combinator startups, seen as an effort to distance these startups from competitive offerings.
Microsoft acknowledges that the GPU virtual machine sets used for training AI and inference follow the same pattern that serves self-interest.
Microsoft stated, “We believe that Azure is the best system for building AI solutions, and we prioritize solutions that rely on Azure.”
The offer is exclusive to startups already relying on Azure, aligning with the vision of making Azure the preferred cloud for building AI solutions.
Microsoft collaborates with its venture capital subsidiary M12 and its investment fund, as well as startups in the M12 portfolio, to expand access to the preview sets.
Over time, Microsoft plans to partner with additional startup accelerators to reduce barriers for promising startups in training and operating AI models.
Microsoft’s approach aims to meet the diverse needs of the startup community by enabling early-stage startups to use credits for training and tuning their AI models.
However, startups won’t be able to run their AI models on the sets indefinitely for free but will receive limited access for a specific period, intended to assist startups in testing their operations and experiments, rather than full-scale deployment.
