In a 10-day window in May 2026, the two leading frontier AI labs each made the same move: backed by over $5 billion in private equity capital, they bought services firms.
I. Two Deals, One Pattern
Two firms with a combined ~$5M raised across them. Acquired in deals backed by trillions in PE AUM. Something is happening.
II. Humans are the Bottleneck
Jon Gray, President of Blackstone, said it plainly: "We want to overcome the shortage of engineers capable of implementing advanced AI systems at the required pace."
But the more interesting question is: why are Anthropic and OpenAI putting billions of their own capital behind services firms? Deployment isn't just a moat — it's the catalyst. By owning the service layer, the labs ensure customer satisfaction and successful outcomes with the most important enterprise deployments, and tighten the feedback loop between the technology and such customers. That creates a flywheel: better deployments → happier customers → deeper integration → more references → more customers. This is an all-out sprint to get their model — Claude vs. GPT — entrenched into the Fortune 500 first. The lab that wins enterprise share wins the next decade.
Is this the SAP and AWS playbook running again? For these transformative technologies, they weren't constrained by what they could do — they were constrained by who could install, customize, and operationalize them. Implementation services were initially owned in-house, then outsourced to the likes of Accenture, Deloitte, and IBM, who built decade-long franchises off that gap. The parallel has merit. But AI introduces a critical wrinkle: outputs are probabilistic, not deterministic — and that changes everything about what "deployment" means.
III. What's an FDE?
The role at the center of this trade is the forward-deployed engineer (FDE) — a term Palantir coined in 2011 for what had previously been called "solutions engineer" or "integration engineer." The work didn't change. The title did. And the new title attracted a different caliber of person — high-agency, high-EQ engineers who would otherwise have viewed client-facing work as beneath them.
What an FDE is: a production engineer who writes real code inside the customer's environment. They sit with the client for weeks or months, frame the actual business problem, ship the solution into production, and own the outcome end-to-end.
That combination — an engineer who can also bridge to a Fortune 500 buyer — is exceedingly rare. The math is harsh:
This is what the Tomoro and Fractional deals are really buying. Not technology. Not IP. Best-in-class, pre-assembled FDE teams.
IV. Will It Self-Resolve?
Reasonable people disagree on whether the services demand is structural or transitional. Three forces could compress it over time:
- Tooling matures. As agentic patterns standardize, more deployment work gets automated. Cloud followed this arc — DevOps tooling compressed the infrastructure consulting market.
- Supply expands dramatically. Universities, bootcamps, and self-taught engineers are racing into the gap. Five years out, the qualified candidate pool will be meaningfully larger.
- AI eats its own deployment work. The most provocative scenario: AI becomes capable enough to map enterprise systems, design implementation plans, and execute them — with human supervision. If frontier labs succeed at this internally, the FDE shortage starts to self-correct.
The bull case for sustained demand: enterprise AI complexity grows faster than tooling, talent, or AI itself can automate. Every prior generation of platform technology was supposed to eliminate consulting. Each one expanded the market for it.
What's clear: capital is voting that the next several years are services-heavy regardless of which way this resolves long-term.
V. The M&A and Venture Read
Three forces shape the next few years of dealmaking in this category:
- Supply-demand imbalance persists → rapid growth → robust venture & M&A activity. As long as enterprises can't hire FDEs fast enough, services firms with proven teams will grow quickly, raise capital, and trade at premium multiples.
- Frontier labs are racing to own the top talent to own the Fortune 50. Anthropic and OpenAI are buying services firms to address the bottleneck ASAP and block competitors from doing the same. This is a land grab. Expect more transactions, more bidders, and more competitive dynamics in every process.
- Enterprise AI adoption remains in its infancy. Most Fortune 500 companies have AI pilots, not AI in production. The gap between aspiration and operational reality is enormous. Expect a multi-year wave of financings and acquisitions as that gap closes.
Cyrus Maghami is the Founder & Managing Director of Harbor Ridge Capital, a SaaS- and tech-services-focused M&A advisory firm that has completed 69 transactions representing $2.3B in value.