Key Takeaways

  • For most application-layer AI startups, the existential "lab risk" from large foundation models is significantly overblown. Founders worry too much about frontier labs steamrolling their niche.
  • Instead of competing on AI models, focus on defensibility through strong network effects, especially in marketplaces. Ev Randall calls these “a drink of water in the desert” for investors.
  • Frontier labs like Anthropic naturally prioritize huge total addressable markets (TAMs), meaning smaller, specialized niches are unlikely to be their immediate focus, even if they're technically capable.
  • The growing trend of app companies hiring dedicated "labs teams" is often a misallocation of resources and unlikely to produce truly groundbreaking research compared to specialized frontier labs.

Your AI App Layer Isn't as Vulnerable as You Think

Founders often live in fear of the next large language model update, convinced that one release from a frontier lab will obliterate their entire business. But Ev Randall, a keen observer of the venture capital landscape, argues this "lab risk" is largely overblown for most application-layer AI companies. He says it’s not rational to assume giants will automatically crush every niche. While foundation models might technically be able to replicate parts of an app's functionality, their strategic focus is usually elsewhere.

Randall paints a clear picture of this prioritization, using a specific example: “My response has always been like look like Anthropic in like 3 months of 2026 probably added the amount of like near like even medium or long-term TAM that exists in AI legal which is still like a ton of revenue... Like what? Like it just doesn't like it just doesn't make any sense at all in terms of like the highest and best use of like the labs time.” The core idea here is that dedicated labs will naturally chase the largest possible markets. If your niche is smaller, even if it's substantial, it's simply not their highest and best use of time or resources. This creates a powerful window of opportunity for focused app-layer startups.

The Oasis of Network Effects

If you're not competing on the foundational model, where do you find defensibility? Randall points to an old truth that AI hasn't changed: network effects. He describes finding “a nice atscale marketplace with clear network effects. It's like a it's like a drink of water in the desert.” This isn't groundbreaking new advice, but it's a vital reminder when founders are caught up in the AI hype cycle.

True network effects, where the value of a product or service increases as more people use it, create a formidable barrier to entry. This applies whether your product is powered by cutting-edge AI or not. For app-layer companies, integrating AI to enhance these network effects (e.g., better matching in a marketplace, smarter content recommendations, more efficient collaboration) is far more valuable than trying to build a new foundational model from scratch. Your defensibility comes from the entrenched user base, the proprietary data generated by their interactions, and the friction of switching – not from the specific model architecture you're using.

Stop Building a 'Labs Team'

Perhaps the most counter-intuitive piece of advice from Randall concerns a growing trend he sees: app companies trying to build their own internal "labs teams." He notes that “a lot of these app companies now feel like they need to have like a labs team where it's like every single app company um you know they'll hire like you know a few researchers for meta or something and then all of a sudden it's like well we're you know we can defend ourselves from the labs.” Randall is deeply skeptical of this strategy.

He argues that these internal teams are often unlikely to produce truly groundbreaking research that could compete with the dedicated, massive resources of frontier labs. “I just think that like there's not there's not that many opportunities for like an in-house research team to be doing that much groundbreaking work,” Randall explains. For most app companies, their R&D budget is better spent on product development, user experience, and deepening those network effects, rather than chasing foundational AI breakthroughs. Trying to keep up with the cutting edge of AI research when your core business is an application is a distraction and a drain on resources that could be better deployed elsewhere.

What to Do With This

Pull your last three R&D budget reviews and identify any resources (time, money, talent) dedicated to foundational AI research, model pre-training, or attempting to replicate lab-level work. If it's more than 5% of your total R&D, reallocate those funds this week towards enhancing your product's core network effects or deepening your moat in a specific, under-served niche. Stop trying to out-innovate Google or OpenAI; instead, focus on creating an indispensable application layer where the AI is an enabler, not the entire strategy.