Anthropic, Blackstone bets next trillion-dollar AI business is startups, not models

AI models are becoming increasingly viable, but what business adoption will look like is still a big question. In an effort to shape that future, labs like Anthropic and OpenAI have created separate businesses dedicated to sending AI engineers to their clients’ offices — a bet that helps businesses figure out how to use their AI models in the next billion-dollar phase.

One of those businesses now has a name: Ode with Anthropic is a $1.5 billion, AI startup launched by AI Lab in May as part of a joint venture with Blackstone, Hellman & Friedman, Goldman Sachs, and others. The move follows OpenAI’s own take on this, the Shipping Company, underscoring the growing recognition among frontier AI labs that winning business customers requires much more than shipping the best models.

Ode was started by Blackstone, which recognized a gap when it roped in large consulting firms and smaller AI services shops to implement AI across its portfolio companies. One of these outlets, AI engineering services startup Fractional AI, apparently stood out, and the joint venture got a start soon after it was announced. (Fractional ended an 11-month relationship with OpenAI when it was acquired.)

Fractional has become the foundation of what is now Ode – a sort of “scaled boutique” AI services firm. And its leaders have big goals.

“It’s very easy to imagine this as a billion-dollar company one day if we do it right,” Chris Taylor, CEO of Ode and founder of Fractional, told TechCrunch in an exclusive interview. “The key business challenge is how do you get through that hyper growth phase without losing the emphasis on quality?”

Ode currently employs 100 engineers, and works closely with Anthropic’s applied AI team to identify where the technology can impact different businesses, and build systems tailored to each organization’s operations.

Anthropic’s internal team will continue to focus on strategic, automated applications, a spokesperson told TechCrunch. The private companies that support Ode will refer their portfolio companies to the joint venture as potential customers, although Ode will not limit the sale of its services to those companies.

For Ode, the ideal customer is one whose CEO buys the offer, according to Taylor.

“A lot of the work we do is very important to the CEO of the company,” said Taylor. “It’s the most important product feature the company is going to build over the next two years, or it’s redesigning the most important business process it has.”

Ode will operate under a “Claude-first” principle, meaning it will use Anthropic technology, including features like Claude Tag in Slack, whenever possible. The company is not limited to Anthropic technology, however, and will use rival AI products if needed.

Eddie Siegel, chief engineer at Ode and founder of Fractional, says the secret sauce of the business is its level of initiative, and the ability to create custom solutions to business problems.

“I think the choice of models is important, but it’s not where most of the calories are wasted,” Siegel said. “It’s one ingredient in a system that needs to be developed, it’s like choosing a programming language when you’re building a piece of software. […] I’m not going to define a business transformation in terms of whether they choose Python or Java.

Taylor added Ode’s founding belief is that “non-AI companies will be among the most successful in this AI moment if they use the technology correctly.” But taking AI, “this magic, tricky ingredient,” and core business processes or customer experience with it requires a lot of help, he said.

“That requires high-level AI talent, which most companies don’t have,” Taylor said.

Ode’s executives describe their team as specialized software engineers, more than half of whom are former founders — the kind of people who “can tackle a challenging technical problem, but also have something to do with it,” according to Siegel. Or as one Blackstone executive put it: a team of “mature,” “special forces” rather than an army of forward-deployed engineers (FDEs).

As several people involved in the business told TechCrunch, demand for such FDE teams far outstrips supply. Ode’s goal is to continue to scale, also internationally, while maintaining its strong boutique position – in other words, conducting continuous experiments to measure the business impact of AI implementation.

But in a world where top engineering talent is already scarce, retaining and growing such a team presents a real challenge. If becoming a top AI engineer requires experience as an entrepreneur, systems thinking first, AI chops, and business product judgment, can Ode train enough people to meet the demand?

Combine that difficulty with the fact that Ode will not only be competing with OpenAI’s The Deployment Company, but also consulting giants like Deloitte and Accenture, which have established their own FDE teams.

Siegel isn’t too worried about the shrinking pool of senior civil engineers.

“It has never been an easy time to be an entrepreneur,” he said. “You learn a lot by trying to end up with problems, try to find a product market, move the needle in business. You learn a lot when you don’t by solving a small problem. That’s a skill set that fits well with Ode.”

Whether enough developers will show up is still an open question. But if Ode and his supporters are right, the next big AI race won’t just be about the best models, but about who can successfully put those models to work inside the world’s biggest companies.