The Information reported yesterday that OpenAI is in talks with Microsoft about a new round of investment. You may recall that Microsoft invested $1 billion in OpenAI in 2019. The expressed goal of that investment was to support OpenAI’s quest to develop artificial general intelligence (AGI). In 2020, Microsoft scored an exclusive license to use GPT-3 solutions beyond the API though it was never clear what that exactly meant.
“OpenAI, whose text- and image-generating artificial intelligence has become a mainstream hit, is in advanced talks to raise more funding from Microsoft, which previously backed the startup with capital that includes credits to use Microsoft’s Azure cloud computing services to develop its technology, according to a person with knowledge of the discussions,” reported The Information.
$20 Billion Valuation?
The article also revealed that existing shareholders of OpenAI equity sold shares in 2021 to new investors Sequoia Capital, Tiger Global Management, and Andreessen Horowitz at a $20 billion implied valuation. Given the expansion of OpenAI’s business over the past year and the initial momentum behind the DALL-E text-to-image generator, the true valuation may be higher. With that said, OpenAI has an unusual cap on investor returns which almost certainly reduces the overall valuation.
In addition, it is surprising that OpenAI’s revenue stands at only “the low tens of millions of dollars this year,” given that Jasper AI, a company that built an application suite that uses OpenAI’s GTP-3 and DALL-E, is expecting to do $80 million this year.
Automated Synthetic Media Generation
The story here is about automating media generation and the use of cloud computing resources. GPT-3, DALL-E, and other text-to-x solutions have shown that AI can dramatically increase synthetic media production. There are already “killer apps” driving rapid adoption. And the AI models draw a lot of computing power that cloud providers covet.
The demand for these services may turn out to be insatiable in the near term because they have succeeded in automating higher-order media creation tasks such as writing and art. When humans take on these tasks, they are slow and expensive. Much of the easy automation that does not require human creativity has already been implemented. AI is helping automate those more complex tasks that don’t follow discrete, predictable processes. That is appealing to both company executives and consumers and even more so when facing a recession.
AI and Productivity Gains
Productivity improvements began to taper materially during the recession that began in 2008, and they have not returned to pre-recession levels. The technology-fueled automation gains of the 1980s and 1990s transferred rote tasks from humans to systems. AI may fuel another productivity rise as more complex tasks begin to shift from humans to software. It doesn’t show up in the data today, but that may soon change.
That is the value of OpenAI much more than the quest for AGI. The tasks that OpenAI solutions are automating can lead to near-term productivity gains, while AGI is likely to remain elusive, and there is no guarantee it is even possible. It may be a less inspiring outcome, but there appear to be clear benefits to capture right now.