Dive Brief:
- Amazon Web Services has expanded the initial cohort of its generative AI accelerator program, the company announced Wednesday. The planned group of 10 start-ups ballooned to 21 after AWS received more than 1,200 applications.
- The 10-week program began Wednesday in San Francisco and will continue remotely until it concludes at the end of July with an in-person demonstration day.
- Of the 21 start-ups, at least one-third could have an enterprise impact, from content authenticators for deepfakes to end-to-end AI platforms.
Dive Insight:
Enthusiasm for the program exceeded expectations, according to Rob Ferguson, AWS Startups’ Global Head of AI/ML.
“When we talked about setting up this accelerator, we were thinking it’d be wonderful if we had 500 applications,” Ferguson said. “Then to have 1,200 absolutely blew away our expectations.”
The call for applicants came at a time when venture capital spending had tightened. The total number of late-stage deals fell for the seventh straight quarter and the volume of angel and seed-stage deals hit a 10-quarter low, during the first three months of the year, according to PitchBook data.
But there seems to be hope for generative AI start-ups.
Anthropic raised $450 million in a Series C funding round, receiving support from Google, Salesforce and Zoom, the company announced Tuesday.
Median pre-money valuations for early-stage rounds of generative AI companies have increased by 16% so far this year compared to last year, according to PitchBook. Prices for all other start-ups raising a Series A or Series B have fallen by nearly 24%.
Participating start-ups in the accelerator will receive $300,000 in credits to build products and services on AWS’ tech stack, according to the company.
“Certainly, the economic situation was a factor in us trying to figure out how to be as supportive with a credit package as possible,” Ferguson said.
While some of the start-ups have a niche focus, a few have broader enterprise connections, including Protopia AI, which is working on software to enable users to extract ML data insights without exposing sensitive information in identifiable form. Wand AI has an end-to-end platform in development that connects data from multiple sources to pipelines, builds AI solutions and scales up to support additional data and solutions.
Stack AI is a no-code tool to quickly design, test and deploy AI workflows with models, like ChatGPT. It could eventually be deployed to build chatbots, answer questions on databases, generate content and process documents.
“Hopefully, we can continue these kinds of programs in the future, because I think that generative AI is only going to expand from here,” Ferguson said.
