By Christophe André, Manuel Bétin, Peter Gal and Paul Peltier.
The release of Deepseek’s R1 model on January 20th stunned the world. This “sputnik moment” in AI showed that an almost unknown Chinese company could develop an AI model at the very top of AI capabilities at a fraction of the development costs of other leading models, release its parameter set (“weights”) for open use and offer ten times cheaper access to users.
Our recent OECD paper, “Developments in Artificial Intelligence markets: New indicators based on model characteristics, prices and providers” (André, Bétin, Gal and Peltier, 2025), shows that while important risks for competition in digital markets persist, the strong position of digital incumbents in the supply of AI has not curbed innovation and prevented potential AI-users from accessing better and cheaper AI models, which provides strong preconditions for adoption across many sectors of the economy.
New data and indicators to monitor AI markets
After the skyrocketing popularity of OpenAI’s GPT models in late 2022, concerns emerged that AI may further entrench dominant positions in digital markets, with incumbents gaining a definitive advantage by controlling the three key AI inputs to AI development: access to data, computing capacity and top AI talents (OECD, 2024). However, emerging empirical evidence offers some nuance regarding such concerns.
The paper relies on an extensive data collection on AI foundation models on the market and shows that, so far, there have been several signs indicating dynamism in three segments of the AI value chain (AI model development, AI model provision from the cloud and AI downstream applications). First, the number of available AI foundation models has been rising exponentially (Figure 1), developed by an increasing number of companies and offering several interaction modalities.
Second, using common industry benchmarks to evaluate AI models’ performances and collecting prices of AIfrom cloud providers, we construct an AI Economic Frontier by identifying, each month, the best models in terms of the price-performance trade-off (Figure 2). Results suggest that in the last two years, the positions at this AI Economic Frontier have shifted continuously towards lower prices and higher quality. Moreover, the developers and models that make it to the frontier have been changing, with five to six players alternating at the frontier (OpenAI, Meta, DeepSeek, Anthropic, etc.) and around ten others following closely.
Figure 2. The AI Economic frontier shows the continuous improvements of AI
Note: Performance is defined by a normalised weighted performance index on industry benchmarks. Each dot represents the model with the best available price-performance trade-off within Text-to-Text models.
Source: André, Betin, Gal and Peltier, 2025.
This variety of models at the frontier is important from an economic perspective. Many users may not always need the best available models and would rather pay an order of magnitude less to access “good enough” models specialised for specific tasks or preferences. In addition to the offer of closed models directly from the cloud, open-weight models offer an option for cheaper (with no license fee), transparent and easily customable (fine-tuned) models used outside of the public cloud environment. This option provides opportunities for better tailored performance and greater control in specific business applications and enhanced data privacy.
AI is getting better, cheaper and more accessible
Figure 2 illustrates the upward shift of the AI economic frontier, implying that AI has become more efficient and cheaper. Indeed, our quality-adjusted AI price index has fallen by on average 80% in two years (Figure 3) and, on average, 30% of models at the frontier have been replaced every month by cheaper and better models.
AI-adopting firms have benefited from greater access to AI models via a widespread offer accessible through several cloud providers (for business use) and an increasing number of AI-powered consumer services (consumer-facing applications). According to our data collection, around 60 cloud providers offer access to AI models, on average from five different AI developers. Downstream, in consumer facing applications, we recorded more than 12 000 AI tools ranging from chatbots to image editing software, customer support applications or domain specific services. While this offer is large and growing, only a few of them (like ChatGPT) attract most users.
AI market developments have been favourable for AI users, but risks for competition exist
Dynamic AI markets are a necessary condition for the diffusion of AI across the economy via widespread AI adoption in various sectors, a central determinant of long-term productivity gains from AI (Filippucci et al., 2024). Our evidence so far suggests that the supply of AI has been more open than initially expected in various segments of the AI value chain, driving innovation and generating the optimal conditions for broad AI adoption (lower price, better quality, broader accessibility). If current trends persist, dynamic AI markets can foster adoption and boost innovation which in turn are preconditions for widespread economic and welfare benefits.
Nonetheless, several uncertainties and risks persist about the future dynamism of AI markets. For instance, the capacity of digital incumbents to leverage existing compute infrastructure and user base in adjacent markets is high. Furthermore, the high concentration of the necessary inputs for AI development — data, compute, and talent — creates additional risks for long-term competition.
References
André, C. et al. (2025), “Developments in Artificial Intelligence markets: New indicators based on model characteristics, prices and providers”, OECD Artificial Intelligence Papers, No. 37, OECD Publishing, Paris, https://doi.org/10.1787/9302bf46-en.
Filippucci, F., P. Gal and M. Schief (2024), “Miracle or Myth? Assessing the macroeconomic productivity gains from Artificial Intelligence”, OECD Artificial Intelligence Papers, No. 29, OECD Publishing, Paris, https://doi.org/10.1787/b524a072-en.
OECD (2024), “Artificial intelligence, data and competition”, OECD Artificial Intelligence Papers, No. 18, OECD Publishing, Paris, https://doi.org/10.1787/e7e88884-en.
Discover more from ECOSCOPE
Subscribe to get the latest posts sent to your email.