Something significant happened this week. Two big announcements by the leading AI labs on the same day.
OpenAI finalized a $10 billion joint venture backed by 19 private equity firms including TPG, Bain Capital, and Brookfield. They simultaneously acquired Tomoro, an AI consulting firm, bringing 150 engineers and deployment specialists into the new entity. The venture is called the OpenAI Deployment Company. Its job: embed engineers directly inside enterprises to redesign how they operate.
Anthropic closed a $1.5 billion joint venture anchored by Blackstone, Hellman & Friedman, and Goldman Sachs, with Apollo, Sequoia, and Singapore's GIC also joining. Same model: Applied AI engineers sitting inside portfolio companies, building Claude into actual workflows.
The shared conclusion behind $11.5 billion in new ventures:
Models are not sufficient on their own, and the people who can embed AI into how a business actually operates are now the most valuable resource in enterprise AI.
📢Upcoming Event Announcement: Exclusive Workshop | AI Deployment in PE Portcos: The 2026 Playbook. Joining forces with enable2grow on 2nd June 2026 to share a practical AI implementation roadmap for PE investors & operators. Join the waitlist here: [LINK]
What these ventures are actually doing
It is worth being precise, because the headlines do not quite capture it.
This is not traditional consulting. McKinsey writes strategy. Accenture integrates systems. What is new here is the pairing: OpenAI and Anthropic are now fielding people who understand both the technology and how the business actually works, embedded inside the client organization, owning the outcome.
OpenAI's own description of the role is telling. Forward Deployed Engineers (FDEs) work with business leaders, operators, and frontline teams to identify where AI makes the biggest impact, redesign the organizational infrastructure around it, and turn those gains into durable systems.
Anthropic's structure is similar. Applied AI engineers embedded inside the new services company, working alongside portfolio firms to identify use cases, build custom solutions, and support long-term adoption. Blackstone put it plainly in their announcement: the new company exists to break the biggest bottleneck in enterprise AI adoption by expanding access to highly skilled implementation partners.
Both labs have now staked $11.5 billion on that being true.
Why private equity, and why now
The choice of PE as the distribution vehicle is deliberate.
PE firms own portfolios of operating companies. They sit on the boards. They set the mandate. When a PE firm decides AI is a priority, it becomes a priority across every company they own, without an 18-month procurement cycle. Both JVs use the investors' own portfolio companies as the first proving ground. Blackstone's portfolio. TPG's portfolio. Bain's portfolio. Hundreds of captive clients, board-level mandate already in place.
Google is running the same play through a different mechanism. Its $750 million innovation fund announced in April is earmarked for McKinsey, PwC, and TCS to accelerate agentic AI rollouts in enterprise clients. The hyperscalers have reached the same conclusion: the model is not the constraint. The humans who can deploy it are.
What this also means is that the capacity being built right now will be directed at the largest portfolios first. The Blackstones and TPGs of the world hold large-cap companies. That is where the embedded engineers will go first. The mid-market, smaller PE-backed businesses, owner-operated companies: this segment will feel the capacity shortage for longer. The gap between organizations that get embedded AI expertise and those that do not is widening, not closing.
What this signals for the market
AI vendors are becoming integrators. The line between "we sell you the model" and "we redesign your operations around the model" is disappearing. That has real consequences for how you evaluate vendors, who you trust to give you objective advice, and what conflicts of interest exist when your AI provider is also your deployment partner.
The scarce resource has shifted. It is no longer compute or model capability. It is people who can marry technical knowledge with deep operational understanding of a business. OpenAI is acquiring that talent. Anthropic is building partner networks around it. Google is funding it through systems integrators. Everyone has reached the same conclusion at the same time.
Having the model alone doesn't change your workflows or how you operate. You need people who can combine the technology with what's actually happening in the business and implement those changes.
Enterprise AI spend is heading toward $665 billion in 2026. The majority is not delivering meaningful P&L impact. These deployment ventures are the labs' answer to that. But the answer is expensive, and it is being pointed at the top of the market first.
We have been doing this for the last eight months
The reason I started PowerUp AI was that I kept watching the same pattern play out. A company buys access to a model. Maybe they run some upskilling sessions. The tools sit underused, the workflows do not change, and six months later someone asks why the investment has not moved the numbers.
Buying access to a tool and teaching people how to use it is not enough. What actually creates change is someone getting inside the business, understanding how it works, identifying where AI connects to real outcomes, and then staying embedded long enough to make the new way of working stick.
That is the work we have been doing.
We start with a diagnostic workshop: map the workflows, understand where the real bottlenecks are, identify and prioritize the AI use cases that will move the business. Then we build and implement, use case by use case, with adoption coaching built in at every step. We do not move to the next use case until the team is actually using the previous one. We stay until the work is embedded, not just installed.
The results are visible. A fintech company we worked with improved their marketing team's productivity by 53%. More importantly, that translated directly into business outcomes: email open rates from prospects and consumers are now running at 50%, which is a number that moves pipeline.
That is what the labs just spent $11.5 billion trying to build at scale. It is the work that was always missing.
If you are a PE operating partner or a portfolio company leader navigating this, join the free workshop where I will share the well-proven 2026 AI playbook for portcos.
Talk soon,
Pooja
PS: If this was useful, forward it to a portfolio company leader or operating partner who is still trying to figure out where to start. That is exactly who I wrote it for.