Startups receive free AI chips from Microsoft and VCs

Microsoft wants to give a few startups free Azure cloud “supercomputing” resources to develop AI models during an AI chip shortage.

Microsoft added a free Azure AI infrastructure option for “high-end,” Nvidia-based GPU virtual machine clusters to train and run generative models, including large language models like ChatGPT, to its startup program, Microsoft for Startups Founders Hub, today.

A private preview of the clusters will begin with Y Combinator and its startup founders. Y Combinator—why? Microsoft VP of Growth and Ecosystems Annie Pearl called YC the “ideal initial partner” for its experience working with startups “at the earliest stages.”

“We’re working closely with Y Combinator to prioritize requests from their current cohort and then alumni as part of our initial preview,” Pearl said. The focus will be on training and fine-tuning use cases to unlock innovation.

Microsoft has tried to win over Y Combinator startups before. The company offered $500,000 in Azure credits to YC’s Winter 2015 batch in 2015 to lure them away from rival clouds. GPU clusters for AI training and inference may be self-serving.

This may be confirmed by Microsoft

Pearl said, “We believe Azure is the best system for building AI solutions, and we’re prioritizing those building on Azure.” Our goal is to make Microsoft the best cloud for AI solutions, so this offer is for Azure-based startups.

This time, more than Y Combinator startups will benefit.

Microsoft is expanding cluster access with M12, its venture fund, and its startups. Microsoft plans to work with more startup investors and accelerators to “lower the barrier to training and running AI models for any promising startup” (and familiarize them with Azure).

Pearl said that while all cloud providers offer credits to startups, our approach addresses the broader needs of this community by allowing credits to be used for training and fine-tuning for earlier-stage startups.

Microsoft now states that it is a business, not a charity. Startups can’t run AI models on clusters for free forever. Pearl said startups will have “time-bound” access to test and trial rather than run their operations.

Microsoft still markets the offering as unique in the AI ecosystem.

Pearl said, “This program is the first of its kind targeting earlier-stage startups and allowing them to run AI workloads using Azure credits.” That means free GPUs for early-stage startups in this program that would normally be behind larger customers to train their AI models and drive AI innovation.

AWS and Google Cloud offer startup programs and accelerators for early-stage AI companies, so they would likely dispute that “first of its kind” claim. By partnering with investors and their networks, Microsoft may surpass the competition.