Prioritizing Workflow Integration and Physical Implementation Over Model Power

The AI cycle has moved past speculative infrastructure bets and into a phase of fast, tangible scaling. While public market investors focus on justifying heavy capital spending, private markets show that AI-driven returns are real. The shift highlighted at the Morgan Stanley Spark Conference is the move from building to implementing. Competitive advantage is shifting away from the models themselves and toward deep integration with mission-critical workflows. For investors and operators, the advantage lies in finding companies that can navigate the process of AI integration without compromising the reliability of their existing, protected businesses. Success in this phase will not be defined by who has the most powerful model, but by who can build the most effective business solutions at the intersection of digital intelligence and physical execution.

The Complexity Gap: Why Growth Looks Different

The difference in growth rates between AI-native startups and established software companies is not a sign of failure for the latter; it is a result of the software development lifecycle. Coding and inference tools, the building blocks, are growing quickly because they are being adopted during a bespoke phase where companies build custom solutions.

In contrast, application software, which manages finance, customer service, and complex data, requires more than just raw intelligence. It requires integration, workflow understanding, and strict controls. As Adam Wood noted, these applications demand a higher level of reliability that naturally slows adoption. The takeaway for the market: do not mistake the rapid growth of infrastructure tools for the coming obsolescence of incumbents. Incumbents have deep, mission-critical moats, but their transition to AI is harder because they cannot afford to break existing, high-value workflows.

> "An incumbent embedded in complex mission-critical workflows is hard to replace because of the data integrations and customer trust it has by definition. But those same requirements can make introducing AI into its platform more demanding. You cannot compromise the reliability of that work."
>
> -- Adam Wood

The Commoditization of Intelligence

The biggest shift in the AI landscape is the transition of models from moats to table stakes. In a multi-model world, value is shifting toward orchestration, or the ability to benchmark dozens of models against hundreds of criteria to find the right tool for a specific task.

This creates a secondary, often overlooked, financial opportunity: the potential for margin expansion. While conventional wisdom suggests that the cost of running inference tokens will permanently compress software gross margins, the reality may be more nuanced. If software companies can lower inference costs through efficient orchestration while maintaining their pricing power, they may preserve margins better than the market currently expects.

The Physical Moat: Where AI Meets Reality

A non-obvious insight from the conference is that as software becomes easier to build and deploy, the most defensible competitive advantages may shift toward the physical world. AI is not just a digital phenomenon; it is becoming an industrial one.

> "I think it is important for people to remember that AI is not just software. AI is going to get physical through new types of robots that impact the entire globe."
>
> -- Brian Nowak

By integrating AI into logistics, manufacturing, defense, and agriculture, companies are creating physical moats that are much harder to replicate than purely digital ones. This physical implementation is where the next wave of durable value will likely accrue, as these systems become deeply intertwined with the fundamental operations of global industry.

Key Action Items

• Evaluate Incumbents by Journey Durability: Over the next 12 to 18 months, assess software incumbents not just by their AI growth, but by how well they protect their existing workflows during the transition. The journey is a marathon, not a sprint.
• Prioritize Orchestration Capabilities: Look for companies, both startups and incumbents, that are building layers to manage model selection and cost-efficiency. This orchestration layer is where the value is shifting.
• Monitor Gross Margin Trends: Watch for software vendors that successfully integrate AI without sacrificing their gross margins. A bull case exists for those who can lower inference costs while holding onto existing pricing structures.
• Shift Focus to Physical-Digital Intersections: Over the next 3 to 5 years, pivot attention toward companies applying AI to robotics, logistics, and industrial operations. The competitive advantage here is structural and likely more defensible than pure software.
• Identify Forward-Deployed Talent: Look for companies mirroring the strategy of hiring domain experts, such as forward-deployed lawyers, to bridge the gap between AI capabilities and actual business workflows. This investment in human expertise is a leading indicator of long-term product-market fit.