Despite the boom in AI, the building blocks necessary to pre-train, fine tune, and leverage new foundational AI models are still shockingly limited. An enterprise that wants to use a proprietary AI model in their software applications would need to provision GPUs, clean and tag their proprietary data, structure that data into a database, and then run algorithms to pre-train or fine tune a foundational model. This route, however common, remains inflexible and often prohibitively expensive. Is there a better answer?
In the future, the enterprises that want to have the most sophisticated approach to leveraging AI will need to leverage tens if not hundreds of different AI models in concert. Companies will have to pre-train foundational models from scratch, fine tune open source models and leverage closed models like OpenAI’s GPT-4. In order to realize this vision, enterprises need best-in-class tools to create models as well as a platform to orchestrate how multiple models work together.
Today, we’re excited to announce our investment in Together.ai to provide just that – a platform that allows enterprises of all shapes and sizes to pre-train their own proprietary AI models, fine tune open-source models and then host and run inference across them. Just like we saw frontrunners emerge early in the evolution of SaaS, we here at Emergence see the same patterns repeating on the leading edge of B2B AI, and we’re thrilled to be part of the journey for such a breakout company.
Together provides the critical infrastructure for enterprises to develop their own innovative AI models that’ll do precisely what they need more affordably and with greater control and safety. They’re doing this by providing enterprises with access to GPUs, a library of open source models and datasets, and proprietary algorithms to pre-train their models (all the most vital building blocks to make development easier). Once a foundational model is ready to be pushed to production, Together provides an API for model hosting and inference. Enterprises can now create AI-powered solutions in a way that’s more transparent, flexible, and cost-effective than ever before.
The team at the helm of Together — including veteran entrepreneurs and well-regarded academics — is deeply technical and has long been at the forefront of AI research. The company is the brainchild of Chris Ré, an associate professor in the Stanford AI Lab, which researches and applies the foundations of the next generation of AI systems. Chris developed the idea in partnership with Percy Liang, an Associate Professor of Computer Science at Stanford and director of the Stanford Center for Research on Foundation Models (which Ré and Liang also founded together).
Chris and Percy serve as thought partners to Together CEO Vipul Ved Prakash, who previously founded and sold Topsy to Apple in 2013, and Together CTO Ce Zhang, a former Stanford Ph.D candidate that worked with Chris Re at the Stanford AI lab. The team is rounded out by Chief Scientist Tri Dao, Stanford Ph.D., whose seminal research on FlashAttention v2 is a key technological advantage for Together customers.
The unique confluence of academia and entrepreneurship yields exceptional founder-market fit and alignment with our perspectives on the future of AI – particularly the belief that enterprises will leverage a variety of foundational models in concert ranging from their own proprietary models all the way to open source. Together’s platform enables enterprises to rapidly and cost effectively build and use foundational AI models using Together’s cutting edge training algorithms and inference optimizations. Together helps companies to achieve their own AI vision and own a core piece of their intellectual property — without reliance on third parties like OpenAI.
At Emergence, we have conviction that AI is the future of enterprise technology. Companies that give enterprises flexibility and control over AI in a fast and cost-effective way will unlock the most value. Our investment in Together — our first in a company at the AI infrastructure layer — has reinforced these beliefs and allows us to take a full-stack approach to the burgeoning industry.
We look forward to seeing the innovative AI products enterprises will develop and are eager to support this team so that more of us can build the future of AI technology together.
Enjoying this article?
Sign up to gain access to our thought leadership and have future articles delivered directly to your email.