On Monday, Decagon CEO Jesse Zhang published a provocative new theory, posted under the title “Everyone is wrong about open source AI in the enterprise.” The post grapples with one of the most interesting contradictions of today’s AI economy: More mature AI deployments are switching to lighter models, he says, even at his own company. But the overall spend on expensive state-of-the-art models has barely budged.

It’s a new way to think about the relationship between frontier and open-source models. In Zhang’s telling, they aren’t competitors, and open-source models’ success isn’t coming at the expense of frontier labs. Instead, they’re two phases of the same lifecycle, with expensive frontier models being used to prove out use cases that can be passed along to cheaper open-source alternatives as they mature.

As more mature use cases switch to lighter models, new use cases keep arising — and the overall spend on frontier models barely goes down.

Zhang doesn’t give much data to support the point, but the data isn’t hard to find. Vercel’s AI gateway dashboard shows that, in just the past week, DeepSeek has surged into the lead for token volumes, now processing just over a third of the tokens passing through the company’s infrastructure. Z.ai — the lab behind the popular GLM-5.2 model — jumped into a respectable fourth place over the same period. 

But if you scroll down to overall token spend, you’ll see Anthropic still accounts for more than half of the overall AI spend on the platform. Given that much of the recent change comes from Anthropic’s own rising prices, the share has dropped slightly over the past month, but not significantly.

OpenRouter tells a similar story, capturing a much larger (but slightly less enterprise-y) segment of the market. Deepseek V4Flash is the main winner on overall usage, processing 5.3 trillion tokens weekly. The most popular frontier model, Opus 4.8, handles just over 2 trillion. OpenRouter doesn’t rank models by total spend, but it registers the average token cost for Opus 4.8 as roughly 23x higher than V4Flash ($1.37 per million tokens, compared to just 6 cents), which would mean Opus was still probably capturing the lion’s share of spending.

Those figures don’t even capture the newest arrival, Nvidia’s Nemotron, which is poised to leap to the front of the pack by virtue of Nvidia’s strong connections and the model’s own extreme adaptability.

Those figures don’t fully prove Zhang’s point about the AI lifecycles, but they do show frontier labs like Anthropic aren’t suffering too much from the rise of open source — at least not yet. One explanation is that the market of AI-addressable tasks is growing so fast that the top models are able to maintain their position just by dominating early-stage deployments. As Zhang puts it, “The frontier labs will keep owning discovery. Open source will increasingly own production.” Another explanation might be that, even as clients move to open-source, many use cases are so difficult that they can’t be entirely replaced with cheaper alternatives.

Either way, this two-tiered economy of models may become a relatively stable feature of the AI economy.

As recently as last September, I was writing about the possibility that foundation labs would end up selling coffee beans to Starbucks — that is, serving as commodity inputs while the application layer reaped the benefits. Some parts of that prediction came true: vertical AI plays switched to lighter models, for one, and the economics of “GPT wrapper” startups have remained mostly stable. 

But we’re also seeing that, token for token, frontier providers have been able to hold on to the most desirable part of the marketplace. the premium token price. And that doesn’t seem likely to change any time soon.

As AI costs continue to rise, companies are looking for ways to cut back. The most recent example is Microsoft, which has reportedly begun to deploy a cost-savings strategy by relying less on software from OpenAI and Anthropic and instead deploying its own in-house models.

Indeed, when it comes to two of its most widely used programs — Excel and Word — Microsoft has begun to use its homemade MAI models to respond to a certain percentage of user prompts, Bloomberg reported Tuesday. In the past, the company had advertised the fact that large parts of Office 365 are powered by models from both OpenAI and Anthropic.

While Microsoft still relies on those third-party models, it has also increasingly sought to stand up its own AI agents. Last month, at its annual Build conference, the company announced the launch of seven new MAI models, including an agentic coder and a text-to-image generator.

When reached for comment by TechCrunch, Microsoft said that it had nothing further to share.

Microsoft’s apparent cutbacks are part of a broader trend. After a brief blitz of “tokenmaxxing” earlier this year, the last few months have seen a news cycle awash in stories about tech companies acting significantly more thrifty. Other large companies — like Amazon, Uber, Meta, and Accenture — have also reportedly made moves to curb spending.

The immense cost of providing and buying AI services has become a controversial part of the industry. The sticker shock has gotten so bad in some parts of Silicon Valley that some companies are reportedly looking to Chinese models for more affordable agentic solutions — despite some concerns over potential security issues.