The AI era is forcing startups to rethink what they should own

The era of artificial intelligence is prompting entrepreneurs to reconsider what aspects of their businesses they should own. Traditionally, startup success hinged on possessing specialized technology, specialized teams, proprietary tools, and infrastructure that competitors found challenging to replicate. However, AI is altering this landscape by enabling startups to acquire sophisticated capabilities without building them in-house.

This development raises a strategic question: which assets should startups retain control over to maintain value, independence, and bargaining power?

The answer varies depending on the nature of the business. In certain companies, proprietary technology will continue to be the core asset. In others, the technology might evolve into strategic infrastructure, while the genuine competitive edge lies in customer data, distribution, brand, workflow integration, or direct market access. The key distinction now lies not in the dichotomy of building versus outsourcing, but in understanding where dependence is acceptable and where it inadvertently undermines the company's value.

Technology's centrality to startup defensibility has long stemmed from its expense and difficulty in replication. Yet, with advanced capabilities readily available via APIs, cloud services, or specialized providers, internal ownership may become less crucial than effectively leveraging these technologies. Developing a unique system may still make sense when it yields a tangible technical advantage; nevertheless, constructing it internally merely because technology is deemed strategic can drain capital, engineering resources, and managerial focus without yielding much differentiation.

Harvey, a legal AI company, exemplifies this distinction. Its product relies heavily on foundation models from external sources, yet its competitive advantage extends beyond mere access to those models. Harvey has developed tailored workflows, evaluation systems, and integrations around legal work and is diversifying its technology base by supporting models from multiple providers—including OpenAI, Anthropic, and Google.

Similarly, Cursor, a software development tool, competes through its interface, understanding of codebases, and integration within the development environment. In both instances, the startup does not need to own every layer of the technological stack to create added value.

However, this does not imply that dependency on technology is inconsequential. Suppliers can raise prices, alter terms, restrict access, or introduce competing services of their own. The strategic objective is thus not merely outsourcing as much as possible but identifying which technological components are essentially commoditizing and which remain central to the company's differentiation.

A startup may reasonably lease computing infrastructure, foundation models, or specific automation tools while preserving proprietary systems in areas where technical superiority directly influences customer choice.

Ownership is becoming selective rather than ideological. Moreover, data plays a crucial role when it generates knowledge that competitors find challenging to reproduce. While startups have been told that data is an asset, not all data possesses strategic value, as competitors can access similar material. The critical question is whether the data enables the company to understand or serve customers in ways others cannot easily replicate.

Data generated through direct customer relationships, such as usage patterns, purchasing behavior, operational histories, and accumulated feedback, can be significantly more valuable. This information might reveal customer interactions, friction points, and future needs, allowing the company to learn and improve over time.

However, physical possession of servers or databases does not equate to strategic control. Instead, founders should safeguard the ability to access, interpret, and utilize the information necessary to enhance the business without relying on another company's permission. In this context, customer relationships are becoming increasingly valuable as technology becomes more accessible.

The shift created by AI has profound implications, emphasizing the importance of understanding where dependence is acceptable and where it inadvertently diminishes a startup's value.