Is Microsoft About to Buy OpenAI and ChatGPT?
Semafor says talks are underway for a monster investment that looks more like an M&A deal than an equity investment
Semafor reported yesterday on the rumored Microsoft talks around a new investment in OpenAI. The talks weren’t exactly new information as articles from the Wall Street Journal had indicated the deal size, and other news outlets had suggested talks were underway. However, the Samafor reporting offered some interesting new details.
The funding, which would also include other venture firms, would value OpenAI, the firm behind ChatGPT, at $29 billion, including the new investment, the people said. It’s unclear if the deal has been finalized but documents sent to prospective investors in recent weeks outlining its terms indicated a targeted close by the end of 2022.
Microsoft’s infusion would be part of a complicated deal in which the company would get 75% of OpenAI’s profits until it recoups its investment, the people said. (It’s not clear whether money that OpenAI spends on Microsoft’s cloud-computing arm would count toward evening its account.)
After that threshold is reached, it would revert to a structure that reflects ownership of OpenAI, with Microsoft having a 49% stake, other investors taking another 49% and OpenAI’s nonprofit parent getting 2%.
What?
There is a lot to break down here. First, note that it says the deal “would value OpenAI, the firm behind ChatGPT, at $29 billion, including the new investment.” This is what is called a post-money valuation. It is the implied value of a private company after it takes in the new investment. That means the pre-money valuation (i.e. before new cash is added to the company) is around $19 billion. The $10 billion is added to the valuation after the investment funds are received.
We discussed this rumor previously in Synthedia. A previous funding round for OpenAI was at $14 billion before the public availability of DALL-E and ChatGPT. These figures suggest that those two products, along with other research and development advances and customer revenue, have helped raise OpenAI’s implied valuation by $5 billion. To get to $29 billion, you add the new investment.
Then we have the comment about Microsoft’s right to 75% of all profits until it has recouped its investment. That means this is an equity stake that includes a claim on profits. If OpenAI ever generates $10 billion in profits, Microsoft will walk away with 49% of the company for a net cost of zero. That is if we ignore the pesky details about the time value of money.
We also have the question of what constitutes an investment. Is this a cash deal, or is it completely or partially paid for with Azure cloud computing credits? Might Microsoft be making available a lot of computing resources now with the promise of payment in the future and a substantial ownership stake as compensation for the risk they are taking on by providing services in advance? Computing resources are known to be the largest cost line item for OpenAI, so it may not matter as much as it would for other companies. However, there is no substitute for cash when building a company. Credits are only good for something specific and only when they are needed.
Then we have the ownership structure. Microsoft will end up owning 49%, other investors 49%, and the foundation that created OpenAI will retain 2%. This is not what your typical venture deal looks like.
Investment or Acquisition
“I think the way to think about this is actually M&A. They are going to end up owning about half of this thing,” said Liz Hoffman, Semafor’s business and finance editor, on CNBC’s Squaw Box, about Microsoft. I see her point. If the outlines of this deal are true, and the ownership is at 49%, Microsoft will be by far the largest shareholder and effectively have control of the company.
“I think the way to think about this is actually M&A. They are going to end up owning about half of this thing,” said Liz Hoffman of Semafor.
What happens when you buy a company? You gain control, and you have access to the cash flow being produced. Microsoft’s move has both elements based on the deal terms described. Granted, they would not fully own the intellectual property outright, which would be of interest to Microsoft.
This would be an interesting way for Microsoft to gain control without technically acquiring the company. It is not even clear that Microsoft could acquire OpenAI. Regulators would surely take a long hard look at what it would mean for a tech giant to own OpenAI and whether that may stifle competition. In addition, ChatGPT has been so popular that many politicians and their constituents will know about the solution and may be wary of Microsoft’s intentions.
These factors would surely drag out any formal acquisition process. They would also leave both companies in limbo and hampered from confidently making concrete plans given the uncertainty of regulatory approval. However, making a deal like this now allows Microsoft and OpenAI to move forward as if an acquisition were completed and then go through a more formal process later.
Is Microsoft Crazy?
There is another question about whether a deal with this structure is good for Microsoft. It is a lot of money to invest in a single company, and after the deal is final, you still only have 49% ownership of a company generating only tens of millions of dollars in revenue last year. Rumors suggest OpenAI expects to generate $1 billion in revenue in 2024, but revenue is not profit. There is also cost to consider, so the cash flows coming from OpenAI are extremely uncertain.
However, Microsoft falls into the strategic investor category. They are not looking strictly at the financial investment but also at what it can do with preferred access to the technology. Is there any company that has more users than Microsoft that could immediately employ a GPT-3 writing assistant and DALL-E image generator?
OpenAI’s products are near perfect matches for Word, Powerpoint, Outlook, and Teams. In fact, several competing products have already implemented similar features. The AI writing and image generation features might also enable Microsoft to charge higher fees for its Office application suite, Microsoft 365.
There is also the expected integration of OpenAI’s GPT-4 into the languishing Bing search engine. Bing has never mounted a credible threat to Google’s search dominance. There is a chance that the introduction of conversational search powered by a large language model could provide an opening to reshuffle the search industry market share along with consumer behavior.
And don’t overlook GitHub copilot, which is built on top of OpenAI’s Codex AI model. The Microsoft software helps developers write code faster and added 400,000 paying subscribers in its first 30 days after launch. The paid plans for Copilot range from around $50 to $250 per user per year. It also keeps users embedded in the GitHub ecosystem. This could very quickly become a business with hundreds of millions of dollars in annual revenue.
But wait! There’s more. Microsoft’s biggest rival for office productivity software is Google which is known to have several large language models and features similar to GPT-3 and DALL-E. The tie-up with OpenAI can ensure Microsoft’s long-term access to what are expected to be important productivity features such as text, image, and video generation. The move also ensures OpenAI’s patent portfolio doesn’t wind up being owned by a competitor.
If this deal transpires as described, Microsoft may be overpaying. That is hard to say because we know so little about the size of the market today, how fast it will grow, and how OpenAI will perform competitively over time. However, there is some rationale for Microsoft to make a big move now when the market still has some uncertainty. Next year may be too late. The investment may turn out to be a bargain.
Is this Good for the Industry?
A deal like this is definitely good for OpenAI’s investors, and it may be good for Microsoft. What about the rest of us? I’m not so sure. On the one hand, OpenAI may need the financial backing of a company like Microsoft to ensure its long-term success. This type of deal will probably be welcomed by companies built on OpenAI technology since it will provide more assurance of long-term financial stability.
Like most of those startups, I am not particularly interested in OpenAI’s pursuit of general AI, but I do think their technology is valuable today for a variety of use cases that I would like to see supported going forward. And, OpenAI is a formidable competitor in a market that could be dominated one day by Google. It is generally best to keep more options in a market.
On the other hand, concentrating more AI capabilities and intellectual property in a tech giant and reducing the number of competitors driving the industry forward may not be a good thing. Microsoft has its own AI capabilities and was always seen as a company that could develop competitive offerings to OpenAI. It may also be warier about conducting basic research or taking on risky new projects required to move the industry forward than OpenAI’s founders.
We will have to wait and see how this plays out. Microsoft may not even make the investment. Other investors may show up and try to wrest control from Microsoft. However, the specific details of the deal suggest there may be merit to the earlier rumors. It also provides an opportunity to discuss some larger themes around the generative AI market and how different scenarios may play out.
Let me know what you think.