Meet 7 startups and labs that are capitalizing on the kind of ‘generative’ A.I. that powers ChatGPT

Cohere AI

This company founded by ex-Googlers is dedicated to making it easier for companies to use large language models; it competes with OpenAI’s GPT products. Cohere secured $125 million in a Series B round last year. It also maintains close ties to Google, which has donated Cohere training time on specialized A.I. Computer chips in Google’s data centers.

Stability AI

The company made news with Stable Diffusion, an open-source text-to-image generator that competes with Midjourney and OpenAI’s DALL-E 2. It raised a $101 million seed round—among the largest first financing rounds in history—that valued the less-than-year-old company at over $1 billion. It also recently inked a partnership to make it easier to deploy Stable Diffusion through Amazon’s AWS cloud service.

Jasper

This Austin-based startup uses OpenAI’s technology to power copywriting software for marketing departments. It raised a $125 million investment round last year that valued it at over $1.5 billion. Its founders were reportedly miffed that OpenAI made ChatGPT freely available, since that tool could eat into its own business.

Tome

This company, incubated by Greylock Partners and backed by $32.3 million from them and venture capital firm Coatue, aims to help users create narratives ranging from children’s books to corporate sales pitch decks. A new feature, based on OpenAI’s GPT and DALL-E, allows a person to create an illustrated narrative from simple text prompts.

Prisma Labs

Creator of the viral Lensa app, which takes selfies and turns them into cool digital avatars using Stable Diffusion.

Midjourney

This A.I. Research lab was founded by David Holz, who previously cofounded the gesture control company Leap Motion. Its popular text-to-image generator, which is available on a freemium basis, competes with OpenAI’s DALL-E and Stable Diffusion.

ProFluent Bio

Taking the same underlying language-modeling techniques that power ChatGPT, this small San Francisco–based startup creates entirely new proteins from just a text-based description of the protein’s function. The technique could revolutionize drug discovery and design and possibly create new catalysts for agriculture and manufacturing.

This article appears in the February/March 2023 issue of Coins2Day as a sidebar to the cover story with the header, “Hype generators.” Michal Lev-Ram and Jessica Mathews contributed reporting to the story.