OpenAI's Enterprise AI Strategy in a Nutshell
OpenAI is getting serious about winning enterprise AI market.
Until recently, OpenAI’s enterprise strategy revolved around selling APIs ($1 Bn+ ARR) and ChatGPT enterprise plans ($800M ARR *) with mostly product-led tactics and inbound leads (“it sold itself”). As a result, OpenAI was able to scale without a massive sales team (<300 head count as of September, 2024).
All this is about to change dramatically in 2025 as OpenAI looks to become an AI agent company. Starting with the Deep Research enterprise version, OpenAI is set to release enterprise versions of coding and sales agents in the coming months.
But little was revealed about how OpenAI plans to distribute these AI agents to enterprise. Would it be through Microsoft partnership? Or will they poach a bunch of enterprise salespeople, á la AWS and GCP in 2021? After all, OpenAI is hiring for the new Forward Deployed Engineering and Enterprise Sales Strategist orgs.
It turns out OpenAI will sell enterprise AI with a refreshingly different, and more powerful, approach.
On February 2nd, OpenAI announced a joint venture with Softbank called “SB OpenAI Japan”. The structure of this partnership foreshadows OpenAI’s enterprise AI strategy globally, and its path to becoming an AI agent company, competing in the application layer. This will have profound impact on startups and competition.
In this post, I will unpack these moves and paint a cohesive narrative of OpenAI’s enterprise AI strategy. Specifically, there’s 3 important themes:
Build out of independent distribution channels
Delivery of highly customized models (enterprise grade models), separate from consumer grade versions
Mass delivery of AI agents built on top of these enterprise grade models
(*) As of April 2024, ChatGPT enterprise had about 600K customers, and 1M users by October 2024. Extrapolating, we have about ~$800M ARR in February 2025. This makes OpenAI a top 20 productivity software company already.
