On July 22nd, Microsoft & OpenAI announced a partnership to develop and commercialize what OpenAI calls beneficial AI. This announcement provides substantial indications to Microsoft’s AI strategy as a whole. Given OpenAI’s original philosophy and the spin-off later to being for-profit, there have been many interesting reactions to this news. From the OpenAI’s point-of-view, it is probably easier to understand that they get capital & infrastructure for AGI (Artificial General Intelligence) development out of this deal. Microsoft gets help to establish Azure as a leading AI development & deployment platform against AWS and GCP, and research capabilities where it is lagging to Google.
What’s more interesting is to look at how this fits into Microsoft’s overall AI strategy. From a business perspective, any business would need three things;
- Research to come up with robust solutions to real-world problems
- Infrastructure to develop & deploy solutions
- Skilled people to implement & extend the solutions in specific client contexts
Microsoft is lagging on the first and second points. Google is an undisputed leader when it comes to AI research. Azure, while doing good, is still not the default alternative when it comes to cloud, it is AWS. Google and Amazon both have taken the lead in exposing AI to the world by making the research, technology & solutions available. These companies surged ahead when Microsoft was struggling with its problems. However, Microsoft has taken a very different approach to the AI business.
With the OpenAI partnership, Microsoft is addressing the first two points. Consider another deal announced a few months back with OpenClassrooms. With that collaboration, Microsoft is solving the third problem. It is fair to assume that there will be at least some part of that education, focusing on Microsoft’s AI tools and solutions. All these put together, Microsoft seems to be adopting an outside-in approach towards its AI business against Google and AWS’s inside-out execution.
There is another dimension to Microsoft’s strategy as well. It is also leveraging its prominent position in IDE (Visual Studio) to provide tightly integrated ML capabilities for developers through ML.Net libraries. This strategy has worked in the past.
Whether this strategy will pay dividends for Microsoft or not, only time will tell. As an observer, it will be interesting to see how it goes.
