The Growing Market for Open AI Models
The market for openly available artificial intelligence models appears to be expanding quickly, even though the available measurements are still rough. Usage data is spread across many platforms, and each source tracks activity in a different way. This makes it difficult to know the exact size of the market or compare it fairly with closed AI systems.
A Free Dashboard Tracks the Open Model Economy
An ongoing research project collects data about the use and performance of open AI models. The dashboard brings together information from across the internet and is available for free. It is designed to help people understand how widely these models are being used and how well they perform.
The numbers should be viewed as estimates rather than exact measurements. Public data can be incomplete, and many companies share information that is shaped by marketing goals. Platform activity, startup announcements, and public testing results each show only part of the picture.
Understanding the market requires careful judgment. A high token count may show strong demand, but it does not always reveal how many customers are involved, what they are building, or how much revenue the activity creates.
The Main Question: How Large Is Inference?
The central issue is the size of the market for inference using open models. Inference is the process of running a trained model to answer questions, create text, analyze information, or complete tasks.
The key questions include:
- Is the use of open models growing quickly?
- How much computing power is being used to run them?
- Is this market still small compared with leading closed-model companies?
- Are open models becoming important for business customers and developers?
Several companies that provide hosting and computing services for open models offer useful clues. Their activity can serve as a rough measure of demand because their business depends heavily on customers choosing open systems.
Token Use Shows Rapid Growth
Public information from these providers points to strong growth in the number of tokens processed. Tokens are small pieces of text used by AI systems. A larger token total generally means that models are handling more requests, producing more answers, or supporting larger applications.
The growth curve appears close to exponential. That means usage has been increasing by larger amounts over time rather than rising at a steady, modest pace.
The trend became even stronger after a major new model release helped improve AI agents. These agents can carry out several steps, use tools, and complete longer tasks with less direct human control. As agents become more capable, they often need many more model calls, which can sharply increase token use.
One Provider Reports More Than 40 Trillion Tokens Daily
One open-model infrastructure company announced that it was processing more than 40 trillion tokens each day for customers using customized models. This figure covers only one part of its inference business.
The company also said that this part of the business had grown by more than two times since the summer. Because customized models represent only one segment of total usage, the provider’s full daily token count is expected to be much higher when it releases another update.
The estimate could rise above 100 trillion tokens per day. That would include activity from other types of models and services handled by the same provider.
These figures are large, but they should not be treated as a complete count of the open-model market. Other providers may have significant activity that is not publicly reported. Cloud companies also host open models through their own services, but they generally do not separate usage by model category.
Comparing Open Models With Closed Systems
A useful way to understand the market is to compare open-model token volume with the volume handled by large closed-model and cloud platforms.
One major technology company publicly shares token totals for its cloud platform and application programming interface. Those figures provide a reference point for comparing overall AI usage with the activity reported by open-model providers.
Still, the comparison is imperfect. The large cloud platform includes many kinds of AI services, while open-model companies may focus mainly on model inference. The groups may also count tokens differently. Some may include input and output tokens together, while others may report only selected products or customer categories.
The Market Is Growing, but the Exact Size Is Unclear
The available evidence suggests that open-model inference is no longer a minor experiment. It is becoming a meaningful part of the AI economy, supported by startups, cloud platforms, developers, and businesses that want more control over their models.
Even so, the market remains difficult to measure. Public announcements provide helpful signals, but they do not offer a complete industry total. The strongest conclusion is that demand is rising quickly, while the exact distance between open and closed AI systems is still uncertain.