Something significant just happened in enterprise AI — and it has nothing to do with a new model release. Anthropic partnered with Blackstone, Goldman Sachs, and Hellman & Friedman to launch Ode, a $1.5 billion AI implementation company. OpenAI responded with The deployment Company, a $4 billion subsidiary backed by a 19-firm consortium. The Big Four are spinning up forward-deployed engineering practices. Every major AI lab is making the same bet at the same moment: the next trillion-dollar opportunity in AI isn't building better models. It's putting those models to work inside real companies.
That shift should reframe how every CIO, CTO, CFO, and CEO thinks about their AI strategy.
Why AI Labs Are Suddenly in the Services Business
For the past three years, the AI narrative was about model capability. GPT-4, Claude 3, Gemini Ultra. Each new release prompted boardroom conversations about switching providers, upgrading tiers, running benchmarks. The implicit assumption was that if you just had access to the best model, the rest would follow.
The problem: it didn't follow.
Enterprise AI spending hit roughly $2.5 trillion globally in 2026. Yet only 12% of CEOs report seeing both higher revenues and lower expenses as a result of AI. An estimated 95% of AI pilots deliver zero measurable profit-and-loss impact. Companies widely deploying AI are six times more likely to report clear ROI than those still running pilots — but most organizations are still stuck in pilot mode.
The frontier AI labs saw this gap and recognized it as a business opportunity. If their best customers couldn't successfully implement the technology, it wasn't just a customer problem. It was a threat to the entire AI value proposition.
Ode CEO Chris Taylor, who co-founded Fractional AI (the boutique AI engineering firm that became Ode's foundation), put it directly: "Non-AI companies are going to be among the big winners of this whole AI moment — if they adopt the technology the right way." The qualification matters. The opportunity is real, but access to it requires something most companies don't yet have.
The Model Selection Myth
Here's the quote that should reshape how your leadership team talks about AI procurement.
Eddie Siegel, Ode's chief technologist, described model selection this way: "I think model selection matters, but it's not where the majority of calories are spent. It's one ingredient in a system that has to be engineered. It's like the choice of programming language when you build a piece of software."
Think about what that means. When you built your core business applications, you probably didn't agonize for months over whether to use Python or Java. You picked something reasonable, hired engineers who knew it well, and focused energy on what the software actually needed to do.
Enterprise AI works the same way — but most procurement conversations are still stuck in the language-selection debate. Too many AI strategy meetings are consumed by model comparisons when the harder conversation is: who is going to do the actual implementation work, and how do you build the institutional capability to keep it running?
This isn't a knock on model quality. It matters. Claude, GPT-5, Gemini — these are meaningfully different for different use cases, and choosing well has cost implications. But in conversations with peers across industries, the companies seeing real ROI aren't the ones who picked the "best" model. They're the ones who invested in building the right systems around whichever model they chose.
The New Competitive Landscape
What's emerging is essentially a three-tier market for enterprise AI implementation talent:
Tier 1: The Lab-Backed Boutiques
Ode (Anthropic/Blackstone) and The Deployment Company (OpenAI/TPG consortium) sit at the top. Ode has 100 elite engineers — over half are former founders — who operate with what Siegel describes as the ability to "juggle a really challenging technical problem, but also own something end-to-end." The Deployment Company acquired Edinburgh AI firm Tomoro and its ~150 forward-deployed engineers as its foundation.
The pitch is quality and proximity to the model makers. Ode operates on a "Claude-first" principle but isn't exclusive. Both ventures will take on only the highest-priority enterprise transformations — the kind that represent, as Taylor said, "the most important product feature a company is going to build over the next two years."
Tier 2: Big Consulting with FDE Practices
Accenture launched a Microsoft Forward Deployed Engineering practice. Deloitte has created its own FDE teams. The Big Four bring scale, existing enterprise relationships, and brand risk management that many Fortune 500 boards find reassuring.
The tradeoff: implementation costs range from $250,000 to over $2 million for initial engagements, with senior consultant rates at $400-$800 per hour. A significant portion of work often gets executed by more junior staff. Hidden costs — internal team time, platform expenses, change management — add another 20-40% on top of quoted fees.
Tier 3: Internal Build
Some organizations are choosing to develop their own forward-deployed engineering capability. Forward-deployed AI engineers command median base salaries around $195,000, with total compensation ranging from $215,000 at established firms to over $1.2 million for principal-level talent at frontier AI labs competing for the same people.
The scarcity is real. Siegel acknowledged the challenge: these are elite generalist engineers who combine systems-first thinking with AI expertise and enterprise product judgment. They're not easy to find or train at scale. The good news is that building even a small internal FDE team creates compounding value — institutional knowledge, proprietary tooling, and the ability to move faster than any external consultant.
What This Means for Your Executive Team
The emergence of this implementation ecosystem gives enterprise leaders a clearer framework for the decision in front of them. But it also surfaces a question that deserves honest internal debate: what does your company actually need right now?
If you have a CEO-level AI priority with a clear business outcome: Ode, The Deployment Company, or a comparable boutique may be the right first call. Taylor's description of Ode's ideal customer — a company where AI transformation is "the top one or two priority for the CEO" — is a useful filter. If you're not at that level of urgency and executive commitment, the boutique model may not deliver what you expect.
If you need broad transformation with change management at scale: The big consulting firms bring process maturity and implementation frameworks built for complex organizations. They're more expensive, slower, and sometimes less technically elite — but they know how to work through procurement, legal, HR, and the organizational politics that derail most AI implementations.
If you're building for the long term: The companies I've seen build durable AI advantage are the ones investing in internal capability alongside any external engagement. Hiring even 3-5 strong implementation engineers — people who can bridge business problem and AI solution — creates a foundation that external consultants alone can't provide. The institutional knowledge stays.
The ROI Reality Check Every CFO Needs
There's a data point worth putting in front of your CFO. Companies that are widely deploying AI across their operations are six times more likely to report clear ROI (39%) compared to organizations still running early pilots (6.5%). That gap isn't about model quality. It's about deployment depth.
The implication is uncomfortable: the cost of staying in permanent pilot mode is high. Not the direct cost — the opportunity cost. Every quarter a company runs AI experiments rather than committing to production-grade deployment is a quarter competitors are capturing the compounding returns of a fully deployed system.
The math changes when you shift from "how much does implementation cost?" to "what is continued non-deployment costing us?" In customer service alone, an AI-handled interaction runs roughly $0.50 versus $2.50-$4.00 for a human-handled ticket. That's a 5-8x cost difference that only materializes at production scale.
Call center AI has demonstrated cost reductions of up to 70% at scale. Compliance monitoring AI cuts audit preparation time by 50% in financial services. Software development teams using AI assistance are completing 126% more weekly projects. None of these outcomes exist in pilots. They require the kind of full-stack implementation that Ode and The Deployment Company are explicitly designed to deliver.
The Strategic Framing Question
Here's the question worth bringing to your next AI leadership discussion: Are you thinking about AI as a technology decision, or a capability-building problem?
Technology decisions are procurement exercises. You evaluate vendors, run benchmarks, select a model, sign a contract. The work ends when the contract is signed.
Capability building is different. You're not just buying access to AI — you're building the organizational muscle to use it, iterate on it, and expand it over time. That requires a different investment thesis, a different hiring plan, and a different definition of success.
The $5.5 billion bet that Anthropic, OpenAI, and their backers have placed — collectively — on the implementation side of this equation isn't a coincidence. It reflects something the model makers learned from watching thousands of enterprise AI projects: the model is not the bottleneck. Getting it deployed, integrated, and scaled inside real business processes is.
Blackstone didn't need a political or philosophical reason to back Ode. They saw the gap directly: when they brought large consulting firms and AI boutiques into their portfolio companies, the results varied widely. The firms that stood out weren't the ones with the best AI models. They were the ones who could figure out where AI would have impact, build custom systems for each organization, and measure the business results in terms that mattered to the board.
That observation — documented in a $1.5 billion investment — is one of the clearest signals yet that the enterprise AI market is entering a new phase. Phase one was about access to capability. Phase two is about the ability to deploy it.
What Leaders Should Do Now
Short term: Map your current AI portfolio against business outcomes. Distinguish between pilots with defined endpoints and production deployments generating measurable returns. If most of your AI budget is still in the pilot column, that's a strategic gap — not just a technical one.
Medium term: Clarify your implementation model. Are you building internal capability, engaging external partners, or planning a hybrid? There's no universally right answer, but there is a wrong answer: spending more time and money on AI tools without a parallel investment in the implementation capability to use them.
Long term: Think about AI implementation as infrastructure. The companies building durable competitive advantage aren't just buying better models every six months. They're building systems, processes, and teams that make them measurably better at using AI than their competitors. That capability compounds. Model selection doesn't.
The AI labs have made their bet explicit. Anthropic and Blackstone are putting $1.5 billion on implementation. OpenAI and TPG are putting $4 billion on deployment. Every major consulting firm is building forward-deployed engineering practices.
They're not betting on models. They're betting on execution. That's probably the most important signal enterprise leaders can act on today.
Follow Rajesh on LinkedIn and X/Twitter for daily enterprise AI insights. Subscribe to THE D*AI*LY BRIEF at beri.net.
