Anthropic's Ode Acquires Casper Studios in First Deal Since Launch

Ode with Anthropic, the enterprise AI services venture Anthropic founded with Blackstone and a group of Wall Street firms, has made its first acquisition since launching in July.

In exclusive reporting from The Information published Thursday morning, Julia Hornstein broke the news that Ode is buying Casper Studios, a four-year-old consultancy that helps companies build applications on top of AI models. Casper employs about a dozen technical consultants and has worked with Netflix, Pepsi, private equity firms and hedge funds, according to the report. The purchase price could not be learned.

Ode is announcing the deal at a point when corporate buyers are trying to work out what they are actually getting for their AI spend, and growing uneasy about the bills.

A familiar kind of target

Casper does not come with proprietary technology attached. What it has is a group of engineers and designers who have put AI systems into production for large brands and financial clients, plus its standing as a Select Services Partner in Anthropic's Claude Partner Network. Jay Singh runs the firm as CEO, and its practice areas cover enterprise, financial services, healthcare and media.

Ode was assembled through a nearly identical purchase. The joint venture acquired Fractional AI in May, and that team became the operational core of the company when it launched publicly in July under its permanent name, as we covered at the time. Fractional's co-founders, Chris Taylor and Eddie Siegel, run Ode as CEO and CTO. The company now employs more than 100 people out of San Francisco, many of whom work alongside Anthropic engineers on deployments in industries such as accounting and healthcare.

Buying two small teams of senior implementation engineers inside four months reads less like opportunism than like a considered way to grow. Ode has no shortage of capital, and capital on its own does not produce engineers who have shipped production systems at a bank or a retailer. It can, however, buy the ones who already exist.

Taylor's argument about tooling

Taylor told The Information that a good deal of what enterprises need can be handled with "the tooling that Anthropic is building on top of their models."

That is a claim about labour economics as much as about technology. Traditional consultancies sell hours, staffed in pyramids, with junior headcount absorbing the volume work. If Anthropic's product surface takes on enough of that volume, a small team of senior engineers can plausibly outrun a much larger team of average ones, and the margin structure of AI services stops resembling systems integration at all.

Whether that holds up is the central open question about Ode, and it is not one the company can answer yet. In the meantime, it is hedging by buying more senior engineers.

The rest of the field is doing the same thing

OpenAI's counterpart venture, The Deployment Co., has raised $4 billion at a $10 billion valuation from backers including TPG, Brookfield Asset Management, Bain Capital and Advent International. It has already spent some of that money acquiring firms that employ hundreds of the consultants who help companies integrate customised AI systems, The Information reported.

Thrive Holdings, the one-year-old holding company started by OpenAI investor Thrive Capital, raised $2 billion recently to take stakes in businesses that own small accounting and services firms, with the aim of automating their routine work. OpenAI supplies researchers, product and engineering to that venture, though the companies under it are not restricted to OpenAI models.

The three approaches differ in shape. DeployCo is buying consultants at scale, Thrive is buying the underlying businesses, and Ode is buying a dozen engineers at a time. They rest on the same conclusion, which is that better models do not convert into enterprise revenue without someone on site doing the integration work. Both labs have now said as much publicly.

The portfolio question

Anthropic created Ode alongside Blackstone, Hellman & Friedman, Goldman Sachs, Sequoia Capital and others, which together put $1.5 billion into the company. Many of the businesses Ode serves are backed by those same private equity firms.

That arrangement has produced early momentum. Ode announced work this summer with the IT consulting and advisory firms Baker Tilly and Citrin Cooperman, both mid-market professional services firms of exactly the kind the venture was designed to reach. Rodney Zemmel, who leads Blackstone's operating team, said at launch that Ode was already creating "poetry in our portfolio."

It is an efficient way to build a book of business, and it is also a closed loop. The investors supply the customers, the customers supply the revenue, and the revenue validates the venture the investors funded. Nothing about that is improper, but it does make the early growth numbers harder to read. A services firm winning competitive bids against Accenture is telling you something. A services firm winning work inside its own shareholders' portfolios is telling you less.

Why Anthropic needs this to work

Anthropic is growing quickly, mostly on the back of coding models that developers keep choosing even as the cost per user climbs. Companies including Restoration Hardware are buying Claude and related tools to automate finance and project management work. At the same time, OpenAI is repositioning itself around the enterprise, and open-source models are giving cost-conscious buyers somewhere cheaper to go.

The pressure that creates is straightforward. Every deployment that fails to produce a result a finance department can point to is a renewal at risk, and rising per-seat costs make those conversations harder rather than easier. Consultants are the answer both labs have landed on, which is why both are buying them.

Casper Studios is twelve engineers, which will not move Ode's numbers much on its own. What it signals is that Ode intends to grow the same way it was created, and that the market for people who can turn a Claude licence into working software is about to get considerably more expensive.