Distribution Is King

In our first post, Data As The Moat, we discussed how proprietary data gives companies a competitive edge in the age of intelligent technologies. But having a data moat is only part of the equation. The key to turning that moat into sustainable profits lies in the second piece of our investment framework: distribution.

As intelligent technologies continue to evolve at an unprecedented pace, having cutting-edge technology is no longer enough. The real challenge now is how companies bring their innovations to market profitably and efficiently. The ability to distribute intelligent technologies at scale—getting them into the hands of users quickly and seamlessly—has become the defining factor for sustainable success.

There are only a handful of companies worldwide that are achieving positive unit economics from their AI products and services, and most of them are companies with existing distribution networks where AI is being inserted as a feature or an add-on to an existing experience. These companies have the infrastructure to scale AI rapidly and integrate it into user experiences that customers are already comfortable with, eliminating friction and speeding up adoption.

Why Distribution Matters in AI

Companies like Amazon, Microsoft, and Meta are not necessarily leading in AI innovation but are leveraging their vast networks to roll out AI features on top of established products. This allows them to scale AI quickly, achieve market penetration, and turn a profit faster than startups building standalone AI products.

Established distribution networks bring several critical advantages to the table:

• Instant Scale: Companies with broad reach can introduce AI features to millions, if not billions, of users almost overnight. This allows them to capitalize on AI’s potential without needing to build a user base from scratch. In contrast, reaching scale takes time for startups, during which they may burn through capital without generating sufficient revenue, making it difficult to achieve positive unit economics in the early stages.
• Customer Trust: Users are more likely to embrace AI enhancements when they come from a brand they already trust. Trust minimizes the barriers to adoption, making it easier for established players to scale their AI offerings.
• Cost Efficiency: By leveraging existing infrastructure, these companies significantly reduce the cost of deploying new AI features. They don’t have to create new platforms or acquire users—AI simply becomes a new layer on top of what already works. Without existing infrastructure, standalone AI startups have to spend heavily on marketing, customer support, and scaling operations—all of which erode margins.
• Data Collection: A larger user base feeds more data back into AI models, enabling continuous refinement and improvement. This feedback loop creates a compounding advantage that startups often lack.

The Cost Decline Dilemma

The traditional "fast-follower" strategy, where incumbents let startups de-risk new technologies before scaling, is becoming less viable as the speed of AI cost reductions increases.

For example, Google delayed the public release of its large language model for over three years after OpenAI launched GPT-3. By the time Google entered the market, its model was 40% more expensive per unit of performance compared to OpenAI. Apple has been even slower, with plans to launch its first AI-driven products four years after GPT-3, acknowledging that its models will be less performant than OpenAI, Anthropic, and Meta’s open-source Llama 3.6.

This hesitation isn’t just about technology—it’s also about protecting reputations. AI is unpredictable, and deploying it at scale can lead to unintended consequences, as seen with Microsoft’s ChatGPT generating controversial responses.

Startups can leverage partnerships with established companies to embed their AI technologies into larger ecosystems, gaining access to the partner’s existing distribution networks. These collaborations benefit both sides: the startup gets instant distribution and credibility, while the larger company gains access to innovative AI features without incurring development costs. This strategy helps startups bypass high customer acquisition costs and compete against incumbents by focusing on innovation while their partner handles distribution.

The Power of Existing Distribution Networks

Charging for AI as a feature or add-on to existing products is not only more feasible but also far more obvious than monetizing a standalone product. By embedding AI into existing platforms—such as adding AI-powered grammar checks in Word or recommendation engines in Amazon—these companies are able to charge for AI’s added value without requiring customers to adopt entirely new tools or workflows.

This approach creates a more straightforward path to profitability because customers are paying for a feature that enhances an already valuable service, not an entirely new offering that must prove its worth from scratch.

With more than 8.5 billion daily queries, Google has an enormous data advantage. While it faces competition from emerging AI-first companies like OpenAI, Google’s entrenched user base gives it the ability to incorporate AI directly into search, using distribution as a competitive moat.

Similarly, Meta’s social platforms—Facebook, Instagram, and WhatsApp—collect massive amounts of user-generated data. This gives Meta a unique edge in training AI models at scale and deploying AI features in a way that feels seamless to users. The size and engagement of Meta’s platforms create a powerful feedback loop that continuously improves its AI offerings.

Conclusion: Why Distribution Is King

In the world of AI, distribution is king. Companies with the ability to integrate AI into existing products and deliver them through established channels are far more likely to succeed than those starting from scratch. They benefit from customer trust, cost efficiencies, and the ability to scale rapidly, all while refining their models through a steady influx of user data.

As we refine our investment framework, the importance of distribution becomes more apparent. It’s not just about building the best AI technology; it’s about delivering it to the masses in a way that’s scalable, efficient, and profitable. Companies that master this balance—those that combine a data moat with powerful distribution—will emerge as the true leaders in the intelligent age.

In a world where AI technology is becoming increasingly commoditized, distribution is what sets the winners apart.