Can OpenAI stay on top?
“FAANG.” “MAANG.” What’s the latest acronym? Well, if you ask around some might say it should be “MAANGO.” Yum. OpenAI took the world by storm by introducing ChatGPT in November 2022. Within two months of this launch, 100 million users tried ChatGPT and OpenAI, as recently as October 2023, was reportedly seeking a valuation of $80 billion. As the dust settles and AI fervor cools ever so slightly, the question remains if OpenAI will be able to keep a commanding market position indefinitely. The answer is likely no.
To begin, OpenAI’s fundamental “business model” is at odds with any such desire for global tech dominance by making fundraising difficult. A surprise to many, Open-AI is a non-profit founded in 2015 with a goal of building artificial intelligence that would equal or surpass human intelligence. In order to deliver upon this goal, OpenAI has specific and evolving investment criteria such as “profit-participation units” that have scared off investors in previous rounds. As other LLMs become more advanced, investors are likely to chase more traditional term sheets. Brad Lightcap, OpenAI’s COO, has even admitted that investment in OpenAI is more akin to a “donation” which is not the most comforting words to investors chasing a profit.
Training LLMs is incredibly resource intensive, and other tech giants can likely outspend OpenAI. Sam Altman has said that OpenAI will be “the most capital-intensive startup in Silicon Valley history.” Development of GPT-4 cost $100 million, a figure said to be several times more than GPT-3. Currently, OpenAI is working on GPT-4.5 as opposed to GPT-5 due to these exorbitant costs. GPT-4.5 will function similarly to GPT-4, but be more cost effective. Currently, OpenAI has raised $11.3 billion from 28 investors. Google and Meta, per their most recent September 2023 10-Qs, have $30.7 billion and $36.9 billion, respectively. Other tech giants are certainly capable of outspending OpenAI in this arms race, and may even be able to do so more wisely with decades of experience with working on a variety of complex, expensive problems.
OpenAI’s models operate in a black box, and the company refuses to share paramters publically. Experts such as Baldur Bjarnason argue that AI models can be poisoned, and users none the wiser. OpenAI’s refusal to disclose information about its language and diffusion models, validation processes, and training data sets could raise doubts amongst enterprise security, engineering, and governance teams about the safe management of technologies like ChatGPT. Secrecy can very well undermine trust and collaboration in the research community, hindering OpenAI’s potential dominance in the tech industry. Corporations, where large B2B contracts are at play, may prefer open-source models like Meta’s LLAMA. Other corporations may even choose to build their own models. Open source is also more likely to create a community of developers.
Finally, OpenAI’s has a strong niche in the generative content market with ChatGPT and DALL-E, but AI is far more generative models. Tools like RapidMiner, Alteryx, DataRobot, and Dataiku are established companies with strong footholds in the enterprise market. Generative AI from OpenAI or any competitor can be a plugin or complement, but capabilities within these tools speak far more to what enterprise customers would seek out of an AI tool at large. These tools are being used to deploy AI/ML models with far more sophisticated use cases than marketing copy or digital assets. From helping assist clients predict likely cases of insurance fraud, to helping governments track the movement of armed insurgencies, OpenAI does not cover the entire scope of AI solutions like many believe.
Only time can tell what the future of OpenAI holds. OpenAI has achieved much, but tech dominance is not a guarantee. Could OpenAI one day be synonymous with AI like Meta is social media and Google is search? Possibly. OpenAI has a clear first-mover advantage, but must grapple with the company’s original goal as a nonprofit, strong competition from other tech companies, staying within ethical realms, and standing out in a very crowded AI/ML environment if OpenAI is to stay on top.
