How Will Most Firms Monetize Generative AI?

Some firms will make money from generative AI by selling the enabling infrastructure, including graphics processing units and licensing of models. But what are the revenue models for the typical business that does not sell GPUs or license GenAI models?

In some cases, there are direct models.

Revenue Model	Description	Example
Subscription	Users pay a recurring fee to access the generative AI service.	Adobe Creative Cloud, Netflix
Pay-per-use	Users pay a fee each time they use the generative AI service.	Google Cloud Text-to-Speech, Amazon Rekognition
Advertising	The generative AI service displays ads to users.	Google Search, Facebook
License fees	Businesses pay a fee to use the generative AI technology.	OpenAI GPT-3
Selling data	The generative AI service collects data from users and sells it to third parties.	Google Analytics, Facebook Ads
Improved customer experience	The generative AI service can be used to improve the customer experience, such as by generating personalized recommendations or creating customer support chatbots.	Amazon Personalize, Drift
New product development	The generative AI service can be used to develop new products or services, such as by generating new designs or creating marketing materials.	DALL-E 2, Noun Project
Brand reputation	The generative AI service can be used to improve the brand reputation of a business, such as by generating positive reviews or creating viral content.	The New York Times, Coca-Cola

In other cases, the payback will be indirect. 

Direct Revenue Models	Indirect Revenue Models
Subscription	Reduced costs
Pay-per-use	Increased sales
Advertising	Improved customer experience
License fees	New product development
Selling data	Brand reputation

In the early stages of a technology change cycle, the firms that benefit the most are those that provide the enabling platforms. These are the firms that develop the basic technologies that other firms will use to build new products and services. 

In the generative AI business, that means Nvidia selling graphics process units. That fits an older pattern of new technology opportunities, where infra has to be built first, before use cases, apps and industries can develop. 

By 2030, $79 billion will be spent annually on specialized applications designed to improve automation and increase productivity, especially in the security, health, and content marketing industries, according to Forrester Research analysts Michael O'Grady, Mike Gualtieri and Michael Kearney.

The goals will typically be productivity improvements, as almost always is the case when new technologies get adopted. That includes using generative AI to develop code as well as content useful for training, customer service and marketing. 

By 2030, $42 billion will be spent annually on generalized use cases, such as research, writing, and summarizing tools, Forrester also predicts. More than 50 percent of this spend will be on chatbot and communications platforms, which will drive substantial improvements in customer and employee experience, Forrester says. 

Logical questions already are being asked about generative AI revenue models. Perhaps the better question is “revenue models for whom?” Nvidia and others already have a revenue model: sell more graphics accelerator units. 

In a similar way, some firms that can generate GenAI models can earn revenues by licensing those models to end user firms. 

But that is not the question many are asking. Instead, they want to know how firms that do not sell GPUs or develop GenAI models are going to earn a return on the costs of creating such capability. And the generic answers are somewhat obvious: 

As RBC analysts note, there are several generic approaches for GenAI monetization. 

Firms can charging for GenAI solutions as a separate SKU, perhaps pricing on a 

per-seat or consumption basis. 

Also firms can add GenAI capabilities as a feature of an enhanced or premium tier, much as Netflix offers ad-free and ad-supported versions of its streaming service, or as linear video service providers offer multiple packages of channels. 

In many--perhaps most cases--GenAI will simply be incorporated into the base product at no extra charge, the financial returns being driven indirectly, such as higher market share or lower churn. 

For example, in the early days of the internet, the firms that made the most money were the ones that built the infrastructure, such as server suppliers, then internet service providers and the web hosting companies. These firms provided the platform upon which other firms could build their businesses.

As the technology matures, use cases will expand from “automation”  and “content generation” to “new products,” at least in some cases. 

But it is reasonable to expect that the revenue models in most cases will involve indirect monetization. Amazon might use GenAI to expedite or improve product descriptions, which will lead to boosts in sales volume.  Similarly, GenAI might enhance the product search process, or provide better previous-buyer feedback, which will also tend to boost sales. 

Generative AI can be used to create personalized experiences for customers, which can lead to increased customer loyalty and repeat business.

When Generative AI automates tasks, lower labor costs can result, reducing operating costs overall. 

Generative AI also can be used to create engaging content, which can help to increase brand awareness and attract new customers. That, in turn, might aid advertising sales. 

Personalized recommendations are an obvious way to produce more transactions, as well. Such capabilities should also lead to higher user satisfaction, which might reduce churn and stimulate incremental transactions.