On the same day in May 2026, Anthropic and OpenAI each launched a joint venture worth billions of dollars — not to build a better model, but to deploy the models they already have inside enterprises. That timing wasn’t a coincidence. It marks the moment frontier AI labs stopped competing only with each other and started competing with Accenture, TCS, and Infosys.
Two rival labs, one identical bet
Anthropic’s venture, “Ode with Anthropic,” launched May 4 as a $1.5 billion joint venture with Blackstone and Hellman & Friedman each committing $300 million, alongside Goldman Sachs, Apollo Global Management, General Atlantic, GIC, Leonard Green & Partners and Sequoia Capital. Built on the acquisition of Fractional AI, it runs on roughly 100 forward-deployed engineers operating under a “Claude-first” principle for client work.
Hours earlier that same day, OpenAI unveiled “The OpenAI Deployment Company,” valued at roughly $10 billion, with over $4 billion from 19 outside investors led by TPG, Brookfield Asset Management, Bain Capital and Advent International, plus about $1.5 billion from OpenAI itself. The venture acquired UK consultancy Tomoro for its roughly 150 forward-deployed engineers, and lists Bain & Company, Capgemini and McKinsey as delivery partners.
Two labs, two structures, the same underlying thesis: the model is no longer the differentiator. Getting it running inside a real enterprise workflow is.
The gap these ventures are chasing
The commercial logic behind both bets traces back to a single number Cisco’s president and chief product officer, Jeetu Patel, cited at RSA Conference 2026: 85% of enterprises are running AI agent pilots, but only 5% have moved them into production. That 80-point gap between experimentation and deployment is, in effect, the addressable market both labs are now chasing with services revenue instead of API revenue.
Closing that gap requires a scarce and expensive role: the forward-deployed engineer. Indeed data shows postings for the role grew roughly 7x year-over-year between April 2025 and April 2026. A survey of 1,500 FDEs by Perspective AI put median total compensation for a senior FDE at a frontier lab at $485,000, staff-level FDEs above $725,000, and principal-level applied AI engineers at Anthropic reportedly above $1.2 million — among the most expensive engineering roles in tech.
Why this puts AI labs on a collision course with IT services
Forbes reported that these forward-deployed-engineer investments from Anthropic, OpenAI and Meta are putting frontier labs “on a collision course” with Accenture, TCS and Infosys — the first time AI labs have been framed as direct competitors, not just customers or partners, of the global IT services industry. That framing matters for any technology leader choosing an implementation partner over the next few years, because it changes who is actually bidding for that work.
It also opens a window for a different kind of provider. Lab-run implementation arms carry structural incentives an independent consultancy doesn’t: Ode’s “Claude-first” principle for client work is a case in point. Enterprises that want an AI implementation partner without a specific vendor bias — or without paying for $485K-plus FDE salaries baked into the bill — have reason to look at specialized, nearshore consultancies that can deliver the same modernization and AI-integration work model-agnostically.
What to weigh before picking a deployment partner
Before defaulting to a lab-run venture or a traditional systems integrator, it’s worth weighing:
- Vendor neutrality — does the partner default to one model provider, or choose the right model for the workload?
- Cost structure — FDE-driven engagements carry some of the highest per-engineer costs in the industry; a leaner delivery model can close the same pilot-to-production gap at a fraction of the price.
- Delivery speed and specialization — a smaller, focused team embedded in your existing engineering org often moves faster than a large SI or a newly formed joint venture still building its own delivery playbook.
The takeaway for 2026
The market-size numbers for “AI implementation services” are still all over the place — estimates for the broader AI consulting category range from roughly $12 billion to $38 billion depending on the research firm, which is itself a sign of how early this category is. What’s not in question is the direction: two frontier labs just backed that direction with roughly $11.5 billion in combined venture capital on the same day. For enterprises still stuck in pilot mode, the real decision isn’t whether to get help closing the gap — it’s who to trust to do it without locking you into one model provider’s roadmap.