Claude Mythos shutdown, let me see the true cost of renting AI

Author: Lin Qiao

Compiled by: Deep Tide TechFlow

Deep Tide Brief: Mythos was suddenly shut down this week, directly exposing a fatal risk that most founders ignore: when your core capabilities depend entirely on someone else’s platform, your fate is no longer in your own hands. Who truly owns the intelligence your product runs on?

Mythos was shut down this week. Whether you agree with this decision or not, that’s almost beside the point.

A company built on intelligence it can’t control suddenly finds itself exposed to decisions it can’t influence. Many founders, seeing this, ask themselves the same question: which parts of my business are actually just being rented?

Over the past few years, discussions about open-source models have mainly focused on cost. Can they really get the job done? If they can, how much cheaper are they than calling frontier APIs?

Now we have a fairly clear answer. We’ve partnered with companies like Ramp, Cursor, and Harvey, using the same basic approach: start with a powerful open-source model, fine-tune it for the work in your business that truly matters, and rigorously evaluate it against frontier models.

The results have been surprising. On the tasks they care about most, a tuned open-source model can reach frontier-quality at an extremely low cost. What happened this week made one thing clear: cost was never the most important issue.

The deeper problem is control. Who owns the intelligence your product depends on?

Recently, a lot of discussion has been framed as renting versus owning. It’s not a perfect analogy, but it’s useful.

Renting Intelligence

Renting has worked well—until problems arise. Move-in ready apartments. The lights turn on. The water runs. Someone is responsible for maintenance. That’s why most companies start here.

Frontier APIs are incredible products. They let startups build things that would have seemed impossible just a few years ago.

But renting comes with limitations. Landlords can raise the rent. They can decide what changes you’re allowed to make. They can change the rules. Occasionally—because of reasons unrelated to you—they’ll tell you to move out.

You did nothing wrong. You’re just operating on someone else’s turf. That’s why Mythos’s story resonates with so many people. When your core capabilities depend entirely on someone else’s platform, you expose yourself to decisions you can’t control.

Most of the time, this doesn’t matter. Sometimes, it suddenly becomes very important.

Owning Intelligence

The lesson isn’t that companies should stop using frontier models. On the contrary. Frontier Labs has built extraordinary technology. Most products should use it. We do too. In many ways, frontier models are becoming infrastructure. But infrastructure and ownership are two different things.

You can use public infrastructure while still owning the things that create value for your business. In AI, ownership means starting from the most advanced open-source models and shaping them around your company’s uniqueness.

Your data.

Your workflows.

Your domain expertise.

Your edge cases.

Your evaluation standards.

Your definition of “good.”

Over time, models become less general and more reflective of what your company does every day. That’s where value is created.

Think of a house. Moving furniture is easy. Painting walls is easy. But if your future depends on the layout itself, eventually you’ll want the ability to move walls. Intelligence is the same.

When intelligence belongs to you, no one can quietly pull the foundation out from under your product.

That’s why we build Fireworks this way.

Training and inference under one roof, so companies can adopt the best open-source models, shape them around the questions that matter most to their business, and deploy them reliably in production.

Not just consuming intelligence. Owning it.

No Single Frontier

A hopeful conclusion this week is that the future of AI doesn’t depend on the victory of any single model.

There is no single frontier. There are many frontiers.

A frontier model is one kind.

Models fine-tuned based on years of proprietary company knowledge are another.

Specialized models that better solve specific narrow problems are yet another.

A router that maps requests to a set of models—and together outperform any single model on many tasks—is also a kind.

The most interesting thing in AI isn’t that one model gets smarter. It’s that intelligence is becoming increasingly customizable. The winning companies aren’t necessarily the ones with the largest models. They’re the ones that turn intelligence into unique proprietary assets.

Looking Ahead

While everyone is reacting to the news this week, we’re busy shipping products—Kimi Moonshot K2.7 Code, MiniMax M3, and Alibaba Qwen 3.7 Plus.

The future I’m looking forward to isn’t that some model quietly devours everything it sees. It’s that many teams own the frontier parts that matter most to them.

If Mythos’s shutdown has made you think differently about these trade-offs, we’d be happy to talk.