ESSENTIAL AI

The origin story runs through one of the most important research papers in modern computing. In 2017, a group of researchers at Google published Attention Is All You Need.

It introduced the transformer, a new way to build neural networks that turned out to be the foundation of the entire generative AI boom. Ashish Vaswani was the paper's first author.

Niki Parmar was a co-author. Both spent years at Google Brain watching their idea reshape the industry.

In 2023 they left to build their own company in San Francisco, called Essential AI. They came out of stealth in December that year with a $56.5 million round and a simple credibility pitch.

We built the thing everyone else is using.

Essential AI wants to build what it calls the enterprise brain. The idea is software that helps big companies automate the messy knowledge work that eats up human hours, things like digging through documents, pulling data together, and handling repetitive tasks.

Companies would pay for AI agents that plug into their systems and do this work. In plain terms, it is trying to sell businesses a smart assistant that actually understands their internal information, not just a chatbot that sounds clever.

Essential AI is building large language models and AI agents aimed at enterprise work rather than consumers. The pitch centers on the enterprise brain, tools that read a company's own data and take action on it.

Alongside its products, the team has published open AI research, keeping its scientific reputation front and center. The clearest through-line is the founders' expertise.

This is a company built by the people who designed the core technology the whole field runs on.

Essential AI's biggest asset is its founders' names. When the inventors of the transformer start a company, the best engineers want to work there and the biggest investors want in.

That reputation pulled Google, Nvidia, and AMD onto the cap table at once, which is rare. The company has also leaned into open research, publishing work and datasets that keep it visible in the AI community.

The strategy is talent and credibility first, giant product splash later.

Fame is not a moat. Every serious AI lab, from OpenAI to Google itself, is chasing enterprise customers with huge budgets and armies of engineers.

Essential AI raised $56.5 million. Its rivals have raised billions.

Inventing the transformer earns respect, but it does not pay the compute bills or guarantee a product people buy. The company has also kept a low profile, which is fine early on but eventually it has to ship something that stands out in a very crowded field.

March Capital led the $56.5 million Series A in December 2023. The supporting cast is what turns heads.

Google, Nvidia, and AMD all invested, which is unusual because those three are giants with their own AI ambitions. Thrive Capital, Franklin Venture Partners, and KB Investment joined too.

When Nvidia and Google both back the same young lab, it is a strong vote of confidence in the founders. The money says the industry still wants to bet on the people who started it all.