When OpenAI unveiled its new generative AI model, o3, just before Christmas 2024, it sparked a wave of excitement and skepticism. The model’s enhanced reasoning capabilities raised hopes of advancing AI closer to superhuman intelligence, but its absence from public release fueled doubts about OpenAI’s motivations and business strategy, especially as the firm faced increasing pressure from rivals and industry players.
According to The Economist, one clear takeaway has emerged from this development: as AI models become more sophisticated, they also become significantly more costly to run. Unlike previous models like ChatGPT, which benefitted from low marginal costs, o3’s processing demands drive up the costs of querying—marking a shift away from the zero-marginal cost era that once dominated the digital economy. The higher computing needs per query point to a transformation in the economics of AI, where scaling models will require much greater investment in computing power and infrastructure.
OpenAI’s valuation has soared to $157 billion, but with rising costs associated with advanced models like o3, experts predict that the firm could face stiff competition, especially from the likes of Google’s Gemini 2.0 Flash, Anthropic, and others in the AI sector. The The Economist notes that AI may no longer be the monopoly-creating force it once seemed to be, as costs rise and competition increases, pointing to a future where the AI industry could resemble an oligopoly.
The report also highlights the success of OpenAI’s o3 model in the “Abstraction and Reasoning Corpus” (ARC) challenge. This breakthrough, which showcased the model’s enhanced ability to tackle novel reasoning tasks, underscored how models like o3 are capable of improving inference (the process of answering queries) rather than simply relying on bigger training datasets. This shift, however, comes with the caveat of increasing operational costs, making access to these models more exclusive.
OpenAI’s o3 signals a dramatic shift in AI’s economics, moving away from the earlier, cheaper model of AI distribution to a future where the price of advanced capabilities may limit access, create barriers to entry, and necessitate significant innovation to justify premium pricing. The Economist stresses that while OpenAI’s technical advancements are clear, the sustainability of its business model remains unproven.
