Hear more from me on this topic and others by connecting with me on LinkedIn; if you’re building in this space, I’d love to chat! Reach out at jake@wing.vc.
For AI-powered app companies, the fast-moving waters of artificial intelligence have created a challenge: Where can they safely create value without risking direct competition from foundation model companies?
Foundation model companies are not merely infrastructure providers; these companies have already developed direct-to-user chat products, recognizing that end-user experiences offer superior monetization potential compared to API access. And, they are strategically positioning themselves to own portions of the app layer, enhancing their data collection capabilities and overall defensibility.
I see three key areas where AI-powered app companies might find relative safety and opportunities for growth:
1. Vertical/Domain-specific software is safe
Vertical or domain-specific software presents a promising avenue. Industries with incredibly specialized needs, such as legal, financial, and healthcare sectors, require tools that are intricately tailored to their unique domains. These verticals demand a level of specificity and domain expertise that may be challenging for generalist foundation model companies to replicate effectively.
2. Apps with strong network effects will be resilient
Applications with strong network effects are more likely to demonstrate resilience against competition from foundation model companies. The value derived from user interactions and connections in social networks and marketplaces creates a moat that AI alone cannot easily replicate. These network-driven platforms benefit from a compounding effect where each additional user enhances the overall value of the service.
3. Horizontal software markets are most likely to face competition
The landscape becomes more competitive in horizontal software markets. Foundation model companies are already making inroads here, including in AI-powered developer tools for coding and testing, assistants and productivity tools that are replacing standalone applications for note-taking, writing, and summarization, and even search engines that can quickly and efficiently summarize information rather than relying on traditional link-based approaches.
Looking ahead, we can anticipate foundation model companies expanding their efforts into search and knowledge management, potentially disrupting traditional document search, internal wikis, and knowledge bases. Customer support and chatbots may evolve into distinct offerings, and media generation tools could revolutionize the creation of images, videos, and music.
What horizontal areas will foundation model companies NOT compete directly?
Despite this encroachment, some specific horizontal areas have the chance to remain relatively insulated from direct competition by foundation model companies:
- Enterprise Resource Planning (ERP) systems, with their complex and industry-specific implementations, present a significant barrier to entry.
- HR, payroll, accounting, and finance applications, burdened by high compliance and regulatory requirements, may prove less attractive targets.
- Enterprise security and compliance solutions, facing similar constraints, might also remain the domain of specialized providers.
App-layer AI companies must carefully navigate these waters, identifying niches where they can provide unique value. The challenge for these companies is not just in selecting the right domains, but also in constantly innovating to stay ahead in categories where foundation model companies may eventually enter.
The competitive pressure from foundation model companies is undeniable, but there’s nothing like pressure to drive innovation, and opportunities remain for AI-powered app companies that can leverage domain expertise, harness network effects, or address highly specialized needs. The key to success will likely involve a combination of strategic positioning, continuous innovation, and a deep understanding of specific industry requirements that generalist AI models may struggle to fully address.
Hear more from me on this topic and others by connecting with me on LinkedIn; if you’re building in this space, I’d love to chat! Reach out at jake@wing.vc.
