Subscribe to our daily and weekly newsletters to stay updated on the latest and most exclusive industry-leading AI news.
AI21, a competitor of OpenAI in the generative AI market for enterprises, has announced a timely new round of funding. While OpenAI is mired in internal turmoil, with much of its workforce urging the board to step down, AI21 has secured an additional $53 million for its series C funding, bringing the total to $208 million at a valuation of $1.4 billion. The company’s total raised funds now amount to $336 million.
Similar to Cohere and Anthropic, AI21 provides large language model (LLM)-driven AI systems to enterprises, which are not open-sourced like OpenAI’s. Given OpenAI’s current instability, businesses may be rethinking their reliance on its products, searching for safer alternatives.
Yoav Shoham, co-founder and co-CEO of AI21, emphasized the importance of choice, robustness, and safety for businesses. During an interview with VentureBeat, he highlighted how AI21 uniquely offers a “white-glove” service, closely collaborating with enterprise clients, which distinguishes it from OpenAI.
AI21, with its 250 employees, is headquartered in Tel Aviv. Its latest investors include Intel Capital and Comcast Ventures, along with Nvidia and Google. Although Shoham did not comment on OpenAI’s recent chaos, he had previously voiced confidence in AI21’s market position.
AI21 has appointed Tom Nides, the former U.S. Ambassador to Israel and former leader at Morgan Stanley, to its board of directors. Shoham mentioned that while OpenAI is its primary competitor, AI21 holds its ground effectively in competitive scenarios. He acknowledged OpenAI’s success with general-purpose LLMs like ChatGPT, popular for casual use, but pointed out that chat interfaces are not ideal for many enterprise needs.
Shoham noted that while OpenAI focuses on general-purpose models, AI21 aims to excel at specific tasks, which is often what enterprises require. He critiqued OpenAI’s approach, suggesting it isn’t well-suited for close collaboration with businesses, given its large-scale service model.
He explained that enterprise-grade AI demands more than basic APIs; it needs comprehensive applications, suggesting that the current trend of using tools like LangChain with OpenAI’s LLMs is just a step in the right direction. Shoham predicted that in two years, discussions would shift from LLMs to more complete AI systems.
Shoham also commented on OpenAI’s recent price cuts, interpreting them as a strategic move to gain market share, which he doesn’t necessarily see as negative.
AI21 is building both general-purpose and task-specific large language models, backed by robust algorithms designed for specific tasks. Shoham highlighted that AI21’s summarization model outperformed GPT-4, ChatGPT, and Claude significantly in tests conducted by a major financial institution.
Although enterprises have been slow in adopting generative AI, interest has surged in the past six months, moving from sporadic trials to widespread experimentation. Shoham believes that within a year, generative AI will become a significant financial force in the enterprise world.