OpenAI, Anthropic embed engineers

The AI labs are moving down the stack. Not just building models, not just selling API access, but helping companies actually rewire work around those models. That is the real news here. On May 4, Anthropic announced a new enterprise AI services firm with Blackstone, Hellman & Friedman, and Goldman Sachs, and OpenAI finalized a separate vehicle called the Deployment Company with backing from 19 investors. ### What did they actually launch? Anthropic’s new company is a standalone AI services firm built to bring Claude into core business operations at mid-sized companies. Anthropic said its applied AI engineers will work alongside the firm’s own engineering team to find use cases, build custom tools, and support customers over time. OpenAI’s structure is similar in spirit but bigger in financial scale — a separate venture focused on helping businesses deploy OpenAI software, majority-owned by OpenAI and backed by private-equity and consulting partners. (anthropic.com) ### Why are Goldman and Blackstone involved? Because private equity controls a huge distribution network. These firms own or influence hundreds of portfolio companies that all have the same problem — they want AI gains, but they do not have enough people who know how to integrate models into messy real workflows. So the PE firms are not just financing the rollout. They are supplying a captive customer base. Anthropic’s backers include Blackstone, Hellman & Friedman, Goldman Sachs, and a wider consortium including Apollo, General Atlantic, GIC, Leonard Green, and Sequoia. (anthropic.com) ### Why embed engineers? Because enterprise AI keeps stalling at the same point. Buying a model is easy. Connecting that model to internal data, compliance rules, legacy software, and actual employee workflows is the hard part. Anthropic spelled this out pretty directly — its engineers will work with the services firm to build and maintain custom deployments. Bain said something similar about OpenAI’s Deployment Company on May 11, noting that Bain clients and portfolio companies will get priority access for joint work. (anthropic.com) ### How big are these bets? Anthropic’s venture launched with $1.5 billion in backing. OpenAI’s Deployment Company is valued at $10 billion and raised more than $4 billion from investors including TPG, Brookfield, Advent, and Bain Capital, with other partners such as Dragoneer and SoftBank also tied to the vehicle. That means the combined headline scale is at least $11.5 billion, and arguably more if you count follow-on capital and partner resources. (anthropic.com) ### Is this basically consulting? Yes — but with a twist. Traditional consultants sell advice and implementation. These ventures sell implementation that is tied to one model provider’s stack. That matters because the lab is no longer waiting for Accenture, Deloitte, or an internal IT team to make the product useful. The lab is building its own forward-deployment layer, which looks a lot like Palantir’s old playbook — put technical people next to the customer and make adoption somebody’s full-time job. (cnbc.com) ### What changes for customers? The pitch gets simpler. A portfolio company does not need to assemble vendors, integrators, and model providers on its own. It gets a bundled path — model access, engineers, workflow redesign, and capital-market sponsorship in one package. The catch is lock-in. If the people redesigning your operations are tied to Claude or OpenAI from day one, switching later gets harder. That is not a bug in the model. It is part of the business design. (msn.com) ### Why now? Because the easy part of the AI boom is over. Labs proved demand for chatbots and copilots. But large companies still struggle to turn that interest into durable productivity gains. These ventures are a response to that gap. Turns out the bottleneck was not just model quality. It was deployment labor. Marc Nachmann at Goldman put it bluntly — there is a big shortage of people who know how to integrate AI with existing business processes. (anthropic.com) ### Bottom line? This is the enterprise AI market growing up. The winners may not be the labs with the best benchmark scores. They may be the ones willing to show up on site, plug into ugly systems, and do the integration work themselves. (anthropic.com) (siliconangle.com)