Cohere President, Martin Kon believes that many of the leading artificial intelligence startups in today’s market are akin to high-performance sports cars. In contrast, he describes his company’s product as more akin to a heavy-duty truck.
In an interview with CNBC, Kon explained, “If you’re looking for vehicles for your field technical service department, and I take you for a test drive in a Bugatti, you’re going to be impressed by how fast and how well it performs.” However, he emphasized that the price, space limitations, and lack of a trunk associated with such cars could pose problems.
“What you actually need is a fleet of F-150 pickup trucks,” Kon asserted. “We make F-150s”. Founded by former Google AI researchers and backed by Nvidia, Cohere is placing its bets on generative AI for enterprise applications rather than consumer-oriented chatbots, which have garnered significant attention in the tech industry since the release of ChatGPT by OpenAI in late 2022.
In June, Cohere secured $270 million in funding at a valuation of $2.2 billion, with participation from Salesforce and Oracle. Executives from the company have been active participants in AI forums at the White House, and reports are indicating that Cohere is in discussions to raise $1 billion in additional capital. “We don’t comment on rumors,” Kon stated to CNBC. “But someone once told me startups are always raising.”
The field of generative AI has experienced explosive growth over the past year, with a record $29.1 billion invested across nearly 700 deals in 2023, marking a more than 260% increase in deal value from the previous year, according to PitchBook. It has become a prominent topic on corporate earnings calls, and various forms of technology are being utilized across numerous industries, including financial services, biomedical research, logistics, online travel, and utilities.
While Cohere is often mentioned alongside AI giants like OpenAI, Anthropic, Google, and Microsoft, its focus solely on enterprise chatbots sets it apart from its competitors.
Kon, who also serves as the company’s operating chief, highlighted that by concentrating solely on the enterprise sector, Cohere can operate efficiently and manage costs effectively, even in the face of challenges such as a chip shortage, escalating GPU costs, and fluctuating licensing fees for AI models. “I’ve rarely seen, in my career, many companies that can successfully be consumer and enterprise at the same time, let alone a startup,” Kon remarked.
“We don’t have to raise billions of dollars to run a free consumer service,” he added. Current clients of Cohere include Notion, Oracle, and Bamboo HR, as listed on the company’s website. Kon noted that many customers are within the banking, financial services, and insurance sectors. In November, Cohere reported an increase in customer interest following OpenAI’s sudden and temporary removal of CEO Sam Altman.
Kon acknowledged the persistent challenges posed by shifting dynamics in the hardware industry. He mentioned that the company has maintained a reserve of Google chips for over two years, acquired during Cohere’s early stages to facilitate the retraining of its models. Currently, Cohere is transitioning towards greater utilization of Nvidia’s H100 GPUs, which power most of today’s large language models.
According to Kon, Cohere’s relationships with strategic investors differentiate it from its generative AI competitors. While many companies have secured funding from entities like Nvidia and Microsoft, Cohere’s agreements with such investors are subject to certain conditions tied to the use of their software or chips.
Kon firmly asserts that Cohere has never accepted a conditional investment, emphasizing that every check the company has received, including those from Nvidia, came with no strings attached. “In our last round, we had multiple checks the same size; we had no conditions associated with any one of them,” Kon emphasized. “We explicitly made that decision so we could say we’re not beholden to anyone.”
Cohere’s strategic choice to concentrate solely on enterprise chatbots could potentially shield the company from the contentious realm of misinformation concerns, especially with the approach of election season. In January, the Federal Trade Commission initiated an AI inquiry targeting Amazon, Alphabet, Microsoft, OpenAI, and Anthropic. FTC Chair Lina Khan characterized it as a “market inquiry into the investments and partnerships being formed between AI developers and major cloud service providers.”
Notably, Cohere was not included in the investigation. According to Kon, the company’s growth trajectory has predominantly revolved around areas such as search and retrieval, which necessitate dedicated AI models. He refers to this as “tool use,” involving training models on how, where, and when to locate information required by enterprise clients, even if the model wasn’t initially trained on that data.
Kon highlights search as a crucial aspect of generative AI that often receives less attention compared to other areas. “That’s certainly, for enterprise, going to be the real unlock,” he remarked. Regarding expansion plans, Kon characterizes 2023 as “the year of the proof of concept.” “We think 2024 is turning into the year of deployment at scale,” he added.
