The clock is ticking.

The combination of a rapid cash burn, the lack of a sustainable “moat”, and slow enterprise adoption of LLMs are causing people to question the mid-term viability of OpenAI. OpenAI’s many public controversies and leadership defections have only served to catalyze the concern for its future.

Cash is king.

OpenAI plans to spend $7 billion this year (2024) on AI training and inference and $1.5 billion on staffing while generating approximately $3.5 billion in revenue. That’s a $5 billion loss. In one year. That loss represents almost half of the $13B OpenAI raised to date. While OpenAI’s 2025 revenues should grow relative to 2024, it’s doubtful that those revenues will grow by enough to plug OpenAI’s cash drain by the end of 2025. Could OpenAI raise more money? Sure. But its biggest investor (by far) is Microsoft and Microsoft already gets access to OpenAI’s pre-AGI assets as part of its existing funding agreements. Could they be acquired? Perhaps. But their $80 billion valuation would top Microsoft’s $68.7 billion acquisition of Activision Blizzard (which was the largest tech acquisition in US history).

An increasingly shallow, narrow moat.

An economic moat, or “moat” for short, is a business concept that describes a company’s ability to maintain a competitive advantage over its competitors. In June 2020 when OpenAI released ChatGPT-3 it seemed to have gained a huge jump on the competition. When ChatGPT-4 was released in March 2023, the moat seemed almost impregnable. The improvements from one release to the next were jaw dropping. The model’s MMLU score between ChatGPT-3 and ChatGPT-4 doubled from 43 to 86. That extreme level of improvement has stopped. Seven months after ChatGPT-4, OpenAI released ChatGPT-4o. The MMLU score stayed the same. Today, 17 months after the eye-popping release of ChatGPT4, OpenAI’s biggest competitors have largely caught up. In fact, the HELM MMLU leaderboard currently shows three models from two vendors (Anthropic and Meta) ahead of ChatGPT-4o. While there can be considerable debate about the best way to measure the capabilities of a large language model, there is no debate that the large advantage once held by OpenAI has now vanished.

Build it and they will come … eventually.

The optimism around GenerativeAI’s ability to transform workers, companies, and industries has cooled in 2024. U.S. Census Bureau data released in March found only an estimated 5.4% of businesses use AI of any type to produce goods or services with businesses in the information sector and large enterprises more likely to adopt. The real culprit, as usual, is the hype. ChatGPT amassed 100 million monthly users just two months after its launch in November 2022, making it the application with the fastest-growing user base in history. Compare that to search engines, a technology that is ubiquitous and indispensable today. In 2002, four years after Google incorporated, only 29% of internet users used any search engine on a typical day, according to the Pew Research Center.

Timing is everything.

In a horse race, it’s rarely the first horse out of the gate that wins. OpenAI, with its aggressive launch and fast cash burn might be the horse that draws all the initial attention by immediately taking the lead only to end up well behind at the finish line. Who can forget Webvan, the grocery delivery company that was launched in 1996 only to go bankrupt in 2001? Today, there are dozens of successful grocery delivery operations available to consumers. Webvan came out of the gate too fast and didn’t have the staying power to persevere.

Strawberry fields forever.

OpenAI is still the leader in the GenerativeAI space. Its product wins more of the benchmarks than anybody else’s LLMs. Despite controversies and defections, OpenAI still maintains a top tier employee base of AI experts. Moreover, it is rumored to be close to releasing a new version of ChatGPT that incorporates reasoning into the existing chatbot functionality. Dubbed “Project Strawberry”, the reasoning release got considerable attention when Sam Altman, OpenAI’s CEO, recently posted a mysterious photo of strawberries growing in his garden on X. If OpenAI releases that capability before the end of the year, they will have once again turned the industry on its head.

For what it’s worth.

Infinitive wishes OpenAI nothing but success and prosperity. They have already changed the world, and we hope they will continue to do so for a long time to come.