Microsoft is looking to establish an open relationship with OpenAI

Microsoft and OpenAI have had a symbiotic relationship, with the former providing billions in capital to the fledgling AI lab and in return getting early access to cutting-edge models that are now integrated into Microsoft’s productivity software suite. However, the two companies have gone in different directions Reuters Today I reported that Microsoft is looking to add more models to its 365 Copilot product that aren’t created by OpenAI.

The reason, according to the report, is that Microsoft sees OpenAI’s cutting-edge GPT-4 model as too expensive and not fast enough to satisfy its enterprise customers. Copilot 365 is an AI-powered assistant built into Microsoft’s suite of productivity applications including Word and PowerPoint. The tool is supposed to ingest all of a company’s data and do a myriad of things, like giving users the ability to find information quickly without having to search through different apps; Quickly create a list of the company’s most profitable business units; Or summarize meetings and emails instantly.

that it supposed To do these things, but customers and insiders alike still can’t use Copilot 365, which costs an additional $30 per month per user on the team. Recently Business insider storyMicrosoft employees, speaking anonymously, described the tools as “awful” and “gimmicky,” and that they don’t work well 75% of the time. On the customer front, Business insider He cited a survey of 123 IT leaders published by management consulting firm Gartner, which found that only four said Copilot provided significant value to their companies. It should be noted that some other stories have reported on companies finding value in using large language models, such as by Simplify customer support.

Some of the customers I spoke to Business insider I specifically noted that 365 Copilot is very expensive.

OpenAI’s ChatGPT is a parametric and general model, which means it is trained on large amounts of data and can be more expensive and slow to run; This is why most models are presented in “lite” versions that perform less intense reasoning or “reasoning.” Microsoft trained its own smaller models in-house such as one called Phi-4 and Reuters Sources speaking to the outlet reportedly said the company is looking to “customize other open-weight models to make the 365 Copilot faster and more efficient.”

On the one hand, it makes sense that Microsoft would want to reduce its reliance on OpenAI. If the company is right and AI is the next generational change in computing, then relying on an independent company for the underlying technology is not a great idea.

Microsoft has invested billions of dollars in OpenAI and will receive 75% of its profits Until it can break even on its investments, and even then it will still own a significant stake in the startup. In fact, the company has to hedge its bets, building its own internal models while holding onto OpenAI’s lottery ticket in case it continues on its current skyward trajectory.

Despite being ahead of the curve today, some OpenAI skeptics say we may not know a real winner in the AI ​​race yet (if these technologies are as revolutionary as we’re led to believe). In the same way that there were many search engines that appeared on the Internet in the 1990s, only to be quickly defeated when Google made its late appearance. Microsoft will likely be wise to hedge.