The Future of Generative AI: Exploring Ownership, Opportunities, and Infrastructure

The Future of Generative AI: Exploring Ownership, Opportunities, and Infrastructure

Introduction:

Generative AI has rapidly gained momentum, revolutionizing various industries with its innovative applications. From image generation to copywriting and code writing, generative AI has already generated over $100 million in annualized revenue in multiple product categories. However, amidst this growth, questions arise regarding ownership, sustainability, and the role of infrastructure providers in this evolving landscape. In this article, we will delve into these topics and explore the potential future of generative AI.

The Landscape of Generative AI:

While generative AI applications have experienced staggering growth, challenges remain for companies operating in this space. Application companies, although witnessing rapid revenue growth, often struggle with retention, product differentiation, and gross margins. On the other hand, model providers, responsible for the existence of the market, are yet to achieve large-scale commercial success. This raises the question of whether selling end-user apps is the only path to building a sustainable generative AI business.

The Role of Hosting and APIs:

One key takeaway for model providers is the importance of hosting. Demand for proprietary APIs, such as those offered by OpenAI, is rapidly increasing. Additionally, hosting services for open-source models, like Hugging Face and Replicate, are emerging as useful hubs for sharing and integrating models. These services not only facilitate collaboration but also create indirect network effects between model producers and consumers.

Considering the Public Good:

The potential of generative AI is both immense and potentially harmful. To address concerns, many model providers have organized as public benefit corporations (B corps) or incorporated the public good explicitly into their mission. Surprisingly, this has not hindered their fundraising efforts. However, the debate remains as to whether most model providers should focus on capturing value or prioritize the welfare of all stakeholders.

The Dominance of Infrastructure Providers:

Behind the scenes, infrastructure companies play a crucial role in the generative AI market. On average, app companies spend a significant portion of their revenue on inference and fine-tuning, either directly to cloud providers or third-party model providers. Approximately 10-20% of total revenue in generative AI is estimated to flow to cloud providers. Nvidia, with its data center GPU revenue, stands out as a major winner in this space, highlighting the profitability and defensibility of the infrastructure layer.

Moats and the Future of Ownership:

While infrastructure providers enjoy certain moats, such as scale, supply-chain, ecosystem, algorithmic, distribution, and data pipeline advantages, none of these moats tend to be durable over the long term. Furthermore, it remains uncertain whether strong network effects will emerge in any layer of the generative AI stack. The future of ownership in generative AI remains open, with the potential for both horizontal and vertical companies to succeed, depending on end-markets and end-users.

Conclusion:

As generative AI continues to thrive, it is essential to navigate the evolving landscape strategically. Here are three actionable pieces of advice to consider:

1. Embrace vertical integration: If the AI itself is the primary differentiating factor in the end-product, consider tightly coupling the user-facing app with a home-grown model to drive differentiation.

2. Prioritize hosting and collaboration: Whether through proprietary APIs or open-source model hosting services, leverage the power of collaboration and integration to foster innovation and network effects.

3. Evaluate the balance between value capture and public good: Model providers should carefully consider their mission and the impact of their actions on all stakeholders. Balancing value capture with societal benefit can lead to sustainable growth and support in fundraising efforts.

In conclusion, the future of generative AI ownership is still unfolding. While infrastructure providers currently dominate the market, the landscape remains open to horizontal and vertical companies. By embracing strategic approaches, prioritizing collaboration, and considering the greater good, stakeholders can navigate this exciting and transformative field.