OpenAI, Anthropic, and the War of the JVs

The frontier labs are raising billions to solve the least glamorous, most important problem in AI: making it work inside companies.

OpenAI and Anthropic are launching AI services JVs, which is a fancy way of saying the future of automated work still requires a lot of very expensive humans.

This is the part of the AI story that does not make the keynote. The product video invariably shows a serene agent gliding through the enterprise: drafting memos, updating systems, triaging customers, and making corporate workflows look like a clean, well-lit Apple Store.

Then the agent meets the actual enterprise. The actual enterprise has a 17-year-old ERP system, six disconnected databases, a security team with veto power, a compliance process that moves at geologic speed, and one mission-critical spreadsheet called “FINAL_v7_revised_REAL_final.”

This brings us to today. Both frontier model labs are setting up PE-backed “deployment” machines to buy or build services capacity, meaning engineers, consultants, implementation teams, and industry-specific operators who can actually get AI working inside companies.

Anthropic has announced an enterprise AI services company with Blackstone, Hellman & Friedman, and Goldman Sachs as founding partners. The venture is around $1.5 billion, with Anthropic, Blackstone, and H&F each expected to invest about $300 million, and Goldman contributing about $150 million.

The mission: help mid-sized companies bring Claude into core operations. Anthropic’s applied AI engineers will work alongside the new company’s engineering teams to find use cases, build custom systems, and support customers over time.

OpenAI is reportedly going even bigger. Its new venture, The Deployment Company, is said to be raising roughly $4 billion from a consortium of 19 PE investors, valuing the entity at about $10 billion. Reuters reports that the company is already in advanced discussions around three acquisition targets in AI services, engineering, and consulting.

So yes, the model labs are recreating services businesses around software that was supposed to reduce services work.

For two years, the narrative was that frontier AI would make enterprise software simpler, more automated, and less dependent on consultants. The reality is more humbling. Enterprise AI is often described as a high-margin software business. Deployment still looks suspiciously like labor-intensive, highly skilled services.

This is very Palantir-coded. Palantir’s real edge was never just software. It was the forward-deployed engineer model: put technical people close to the customer’s operations, learn the workflow, shape the product around the institution, and become hard to rip out. OpenAI and Anthropic are now replicating that playbook at scale.

The PE angle is the clever part. Private equity firms own or influence large portfolios of companies. Those portfolio companies are under pressure to improve margins, automate workflows, and show AI-driven productivity. The AI labs need distribution into enterprises. The PE firms need a credible AI transformation story for their portfolio companies.

So the bargain looks like this:

• OpenAI/Anthropic get distribution into hundreds of portfolio companies.

• PE firms get earlier access to AI tooling and implementation capacity.

• Portfolio companies get preferred access to engineers to make AI work.

• The JV captures services economics that might otherwise go to Accenture, Deloitte, boutique AI consultancies, or in-house teams.

There are four big implications.

1. This is a land grab for the deployment layer
The frontier model race is still important, but the next monetization frontier is operational. The question is no longer just “whose model is smartest?” It is “whose model can be embedded into real workflows with enough reliability, permissions, support, and ROI to become institutional infrastructure?”

2. Services are becoming a control point.

Everyone wanted software margins. The path to those margins may run through services first. The lab that owns the deployment layer learns the customer’s workflows, captures usage data, influences architecture decisions, and becomes hard to replace.

This also doubles up as market intelligence. Whoever owns the services layer gets a privileged feedback loop and can turn messy customer work into templates, agents, integrations, evals, and product features. The first deployment is services. The tenth might be product. That is the dream.

3. The AI labs are going around traditional SIs

The obvious losers, or at least threatened incumbents, are traditional systems integrators and consulting firms. Accenture, Deloitte, PwC, McKinsey, BCG, and boutique consultancies should pay attention.

Anthropic says its new company will still work alongside existing partners, including its Claude Partner Network, but the strategic tension is obvious.

The traditional SI pitch is: “We understand your enterprise and can help you choose among AI vendors.” The AI lab JV pitch is: “We built the model, we have privileged access to the product roadmap, and we can bring engineers who understand the frontier system itself.” That is a powerful wedge.

4. This is a pre-IPO story

Both companies are trying to justify enormous valuations. Enterprise revenue quality matters. Consumer AI can be massive, but enterprise AI is where investors look for recurring contracts, high ACVs, budget durability, multi-year commitments, lower churn, measurable ROI, and strategic lock-in.

These JVs help tell that story and create a pipeline from “model capability” to “enterprise deployment” to “revenue expansion.”

---

The irony is rich: AI was supposed to automate knowledge work. Its next act is hiring knowledge workers to help knowledge workers use tools that automate knowledge work. But that only makes the AI story more real. The demo phase was about intelligence. The deployment phase is about institutions.

---

Disclosure: I work at Bain Capital Ventures, which is affiliated with Bain Capital. I am not involved in this transaction or any related investment process, and this analysis is based solely on publicly available reporting. The views expressed here are my own and do not represent Bain Capital, Bain Capital Ventures, or any affiliated funds.

---