Microsoft – Unnecessary risk

Microsoft needs to acquire OpenAI.

• Satya Nadella’s tenure at Microsoft has been excellent but I think the one mistake he has made is to open up his company and his core business to the risk represented by the chaotic and deeply conflicted governance of Open AI.
• The solution is to acquire the company which I suspect will happen once the bubble has burst, and the valuation cratered.
• This weakness is on display once again as a result of Elon Musk’s lawsuit against Open AI which alleges that Open AI has abandoned its founding principles and is completely under Microsoft’s control.
• Elon Musk left the board of Open AI in 2018 after a disagreement with the rest of the board and has since been working on his own version of the company called xAI.
• The problem for Microsoft is that it has gone all in on OpenAI, and recent history has demonstrated that total self-immolation is well within the realm of possibility.
• This is particularly the case as the issues that caused the dismissal and rehiring of Sam Altman have not really been fixed meaning that they could easily blow up once again.
• Microsoft is infusing OpenAI’s foundation model, GPT-4 and its successors into everything that it does, and part of its long-term appeal is expected to be derived from this.
• This means that if OpenAI explodes Microsoft may no longer be able to get updates and improvements to the foundation model hampering its ability to create new services.
• RFM Research has concluded foundation models are difficult to switch as one has to retrain from scratch creating a serious problem for Microsoft should OpenAI implode.
• Microsoft has made some effort to mitigate this problem through its ownership of usage rights to all of OpenAI’s IP, but this does not guarantee that Microsoft can take over and evolve the IP successfully if OpenAI is no more.
• I think that the only way to mitigate this risk is for Microsoft to acquire OpenAI (probably without Sam Altman), and merge it with its own assets to create its version of Google DeepMind.
• Microsoft simply can not afford delays to GPT-4’s successors as there is now lots of competition which, by most objective measures, are just as good or better.
• OpenAI’s competitors have a direct line between product development and revenue and the ability to develop their foundation models exactly the way they want them.
• By comparison, Microsoft is stuck with whatever OpenAI decides to create and it does not have full control over the development direction that OpenAI takes with its models.
• In the long run, this could make Microsoft less competitive in AI highlighting the risk of infusing it into all of its revenue-generating product lines.
• The only answer is for Microsoft to take control of OpenAI by acquiring the company, its assets and its personnel so it can migrate GPT foundation models exactly as it desires.
• I see this as a huge risk for Microsoft and the one big mistake that Satya Nadella has made in an almost faultless run of success.
• Hence, I see increasing risks at Microsoft which is not what one needs when the shares are trading at 30x fiscal year 2025 (June) PER.
• I don’t own the shares and would sell them if I did.