Google chose licenses over consulting: a new strategy in the battle for enterprise

Google writes licensing agreements with private investors, while OpenAI and Anthropic are building their own consulting divisions. The difference in approach may determine who will take the leading position in what is arguably the most important new distribution channel for AI in the enterprise — the first so significant since the emergence of cloud computing.

Three strategies for fighting for enterprise

OpenAI and Anthropic are investing in classical consulting. OpenAI created a multi-billion-dollar division, valued at $10 billion. Anthropic is taking a similar path with an investment of $1.5 billion. Both companies hire consultants, open offices, prepare diagnostics and customization for each client. This is a proven approach: McKinsey, BCG, Deloitte show that consulting works. But it is a resource-intensive path. Each new client requires people, time, and deep knowledge of their business processes.

Google chose something completely different. Instead of creating its own consulting firm, the company concludes licensing agreements directly with private investors like Blackstone and KKR. These funds manage enormous portfolios: hundreds of companies of different sizes, industries, geographies. When Google licenses its models to such a fund, it gains access not to one company, but to the entire portfolio at once.

Why licensing scales better

Consulting requires linear scaling of people. Each new client task requires an expert who understands both Google Gemini architecture and the client's operations. If a consulting firm needs to work with a thousand companies, it needs a thousand consultants. Plus training, plus quality management, plus staff turnover. Hiring qualified consultants, their continuous training and quality management grow non-linearly in cost. This is why even large consulting firms grow slower than SaaS companies. They are limited by the number of talented people on the market and the time to train them.

A license scales completely differently. One contract, one integration, and the entire infrastructure of the partner gets access to Google's models. There are virtually no restrictions on the number of users. An investor can distribute the license among hundreds or thousands of portfolio companies without additional negotiations.

Portfolio companies as a new distribution channel

The main advantage of Google's approach is not the cost of the contract, but the volume of coverage and speed of access. Portfolio companies of private investors are not only Fortune 500, but also:

• Startups in AI, SaaS and fintech
• Manufacturing enterprises and logistics
• Companies in healthcare and retail
• Energy and infrastructure projects

Each needs AI, but doesn't want to independently negotiate with Google, conclude a separate contract, train specialists, integrate the API. When Blackstone or KKR license Google's models, they become an internal distributor. The fund offers AI to its portfolio companies as a tool for improving operational efficiency. A company from the portfolio doesn't pay Google directly — it pays the investor or receives the tool as part of a support program. For investors, this adds value to each investment. For Google — this is access to thousands of companies without launching its own consulting.

Essentially, Google uses private capital as the infrastructure of its distribution.

What this means

Cloud computing went through a path: from technology to service to a distribution channel for software. AI is following a similar route. The difference is that cloud was distributed through IT departments of companies and consultants, while AI can be distributed through financial partners and portfolios.

The company that first builds strong relationships with the largest investors and gets licensing contracts will capture access to tens of thousands of companies simultaneously. The result: not a consulting firm, but an infrastructure company that manages AI access through capital.