In a strategic move to address the ongoing AI chip scarcity, Microsoft is set to provide select businesses with complimentary access to Azure's potent supercomputing capabilities for constructing AI models. This initiative is part of Microsoft's startup program, the Microsoft for Startups Founders Hub, where it has introduced a free Azure AI infrastructure option specifically designed for high-end, Nvidia-based GPU virtual machine clusters.
The offering targets startups involved in training and running generative models, particularly large language models like ChatGPT. Initially, Microsoft plans to roll out a private preview of these GPU clusters in collaboration with Y Combinator (YC) and its startup founders. YC, a renowned startup accelerator, was chosen as the inaugural partner due to its extensive experience nurturing companies at their earliest stages.
Annie Pearl, Microsoft's VP of Growth and Ecosystems, emphasized the strategic partnership with Y Combinator, stating, "We're working closely with Y Combinator to prioritize the requests from their current cohort and then alumni, as part of our initial preview." The focus will be on tasks such as training and fine-tuning use cases that facilitate innovation.
This collaboration is not the first between Microsoft and Y Combinator. In 2015, Microsoft attempted to woo YC's Winter 2015 batch with a $500,000 offer in Azure credits to steer them away from rival cloud platforms. This time, however, the emphasis is on providing GPU clusters for AI training and inference.
While Microsoft asserts that its motivation is to make Azure the premier platform for building AI solutions, there is an undeniable self-serving aspect to the offer. Pearl clarified, "We believe that Azure is the best system for building AI solutions, and we're prioritizing those that are building on Azure."
To extend the reach of this program, Microsoft is collaborating with M12, its innovation fund, and its portfolio firms. The goal is to broaden access to GPU clusters beyond Y Combinator startups. Microsoft aims to engage with additional startup investors and accelerators to reduce barriers to training and running AI models for promising startups.
While other cloud providers like AWS and Google Cloud also offer startup programs and accelerators for early-stage AI firms, Microsoft is positioning its initiative as unique. Pearl stated, "This program is the first of its kind targeting earlier-stage startups and allowing them to use Azure credits to run AI workloads."
However, the acknowledgment that startups cannot run AI models on clusters for free indefinitely tempers any perception of altruism. Access to the GPU clusters will be "time-bound," allowing startups to test their models without an indefinite commitment.
The collaboration with Y Combinator and other startup ecosystem players positions Microsoft to compete with AWS and Google Cloud in supporting and nurturing emerging AI firms.