Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
The AI lab called Important came out of the closet on Thursday, presenting a new model for solving an old problem: how to get information from the many things that are created by businesses. By combining old predictive AI systems with today’s tools, the company believes it can revolutionize the way large enterprises analyze their data.
“While LLMs have been good at working with unstructured data, like text, audio, video, and code, they don’t work well with structured data like tables,” CEO Jeremy Fraenkel told TechCrunch. “With our Nexus model, we’ve created a great foundation for using this kind of data.”
The idea has already attracted a lot of interest from investors. The company is raising $255 million at a valuation of $1.2 billion. Most of it comes from a $225 million Series A round led by Oak HC/FT, Valor Equity Partners, Battery Ventures, and Salesforce Ventures; Hetz Ventures also participated in Series A, with angel support from Perplexity CEO Aravind Srinivas, co-founder Brex Henrique Dubugras, and Datadog CEO Olivier Pomel.
Called a large tabular model (LTM) rather than a large language model (LLM), Fundamental’s Nexus differs from current AI systems in several important ways. The model is deterministic – that is, it will give the same answer every time a question is asked – and it is independent. transformer design which refers to examples from modern AI labs. Essential calls it a basic model because it goes through the standard training and planning process, but the results are very different from what a customer can get when interacting with OpenAI or Anthropic.
This difference is important because the key is to accelerate the application process where current AI models often fall short. Because the Transformer-based AI models can only change the content within their window, they often have trouble thinking on very large pages – analyzing a spreadsheet with billions of rows, for example. But this kind of large-scale model is common in large enterprises, which creates a great opportunity for models that can handle scale.
As Fraenkel sees it, this is a huge opportunity for Fundamental. Using Nexus, the company is able to bring modern methods to big data analysis, providing something more powerful and flexible than the algorithms currently used.
“Now you can have one brand for all your applications, so you can significantly increase the number of applications you use,” he told TechCrunch. “And for every use case, you do a better job than you could with a team of data scientists.”
That promise has already resulted in several high-profile contracts, including seven contracts with Fortune 100 customers. The company has also entered into a strategic partnership with AWS that will allow AWS users to deploy Nexus directly from existing infrastructure.
