The competition among AI companies that for years revolved around having the most powerful and sometimes most dangerous models has now entered a new phase: competition over price.
According to ITNA, the biggest risk facing the AI boom has always been demand, and lowering model prices without reducing their intelligence is seen as a positive sign for the industry's future.
With price reductions for top models from companies like OpenAI and Anthropic, their usage has increased markedly; a trend that seems necessary to justify the multi-trillion-dollar costs of this industry.
A crowded week of new model launches
This week can be seen as the peak of AI companies' competition; OpenAI, Anthropic, SpaceX and several Chinese makers of language models have all launched new products.
Most of these models offer a level of intelligence at prices that were unimaginable just a few months ago.
OpenAI on Tuesday unveiled the GPT-6 Sol and GPT-6 Luna models which, the company says, reduce the cost for large commercial customers by up to 50 percent compared to earlier versions of the same models.
Anthropic simultaneously released Opus 5.5 which, the company announced, runs at about 40 percent lower cost than Opus 5, without reducing the model's intelligence.
Meanwhile, X.AI, owned by Elon Musk, introduced Grok 4.7 on Monday with a focus on price-to-performance; although competition in the industry is so intense that within 24 hours rivals neutralized much of that price advantage.
Pressure on margins, but a positive outlook for the industry's boom
The intensity and speed of this price war is a serious challenge for OpenAI and Anthropic; two companies that need to retain customers and grow revenue to justify their multi-trillion-dollar valuations.
However, experts believe this trend is overall positive for the AI industry's boom, because lower prices make investing in the future of this technology easier.
According to Axios, growing evidence also shows that cost reductions can benefit all AI providers.
Citadel Securities told its clients in a report that lower cost per token has increased consumption and total spending in AI is rising; a trend that could bring greater profitability for AI labs and providers of computing infrastructure.
Morgan Stanley also views competition from Chinese companies as a positive sign for the industry, because that competition, by attracting more customers who can now access AI at lower cost, increases overall demand for compute.
This phenomenon is known in classical economic theory as the "Jones paradox"; a theory according to which lowering the cost of a technological resource increases its consumption.
Leading open-source Chinese models
Meanwhile, open-source model makers, mostly Chinese, have made significant progress in recent weeks.
DeepSeek this month released V4.1 Flash which, the company says, outperformed its previous flagship on several benchmark tests and whose technical improvements enabled price reductions. This model now holds the top spot on the OpenRouter ranking with a 172 percent surge in usage this week.
Price cuts to gain market share are not new in the AI industry. Companies in this field have a long history of launching new models and then cutting prices of earlier versions, or even releasing smaller, cheaper "flash" versions of their larger models.
However, increased competition from open-source model providers and their sharp rise in rankings has put broader focus on price among leading AI labs.
Given the extensive efforts by AI labs to release cheaper models, it seems previous concerns about the exorbitant prices of this technology have now become a distant memory.