The Builder's Economy: 10 Metrics from ICONIQ's Newest 2026 State of AI Report
ICONIQ's third State of AI report lands with a clear message: we've entered "The Builder's Economy," where investors reward founders who ship working products, not those who theorize about future capabilities. Surveying roughly 305 software company executives, the report reveals that AI products are approaching half of revenue, margins are expanding, and the technology stack is shifting decisively toward open-source models and vertical applications.
For founders planning their next move, these ten metrics chart the path from idea to fundable, sellable product—and they suggest that the window for theoretical AI pitches has closed.
AI Revenue Is Becoming the Majority of Business
AI products are no longer experimental line items. According to ICONIQ's survey, AI revenue for non-AI-native companies is projected to jump from 32 percent in 2025 to 53 percent by 2027. For companies born with AI at their core, the trajectory is even steeper. This means that in two years, most software revenue will come from AI-driven features or workflows, not legacy modules.
The implication for founders: your MVP must demonstrate that AI is not a nice-to-have feature but the engine of customer value. Investors want to see a product where AI drives the majority of usage and revenue potential, not a traditional SaaS tool with a chatbot bolted on.
Vertical AI Applications Lead at 43 Percent
The most common product type in the Builder's Economy is vertical AI applications, accounting for 43 percent of deployments. These are purpose-built tools for specific industries or workflows—medical coding, legal document review, logistics optimization—where domain expertise meets automation. Horizontal platforms and general-purpose models still exist, but the fastest growth and clearest path to revenue concentration are in verticals.
This shift reflects a maturation: founders who understand a specific industry can build AI that solves real problems, not generic productivity gains. If you're starting an MVP, pick a vertical where you can own the workflow end-to-end and where customers will pay for measurable outcomes.
Open-Source Models Now Represent 40 Percent of Usage
Perhaps the most striking finding is that open-source models have reached 40 percent of usage at the application layer. Builders are moving away from renting expensive frontier APIs and toward controlling their own stacks through fine-tuning and local deployment. This trend reflects cost discipline, data sovereignty concerns, and the improved performance of models like Llama, Mistral, and others.
For early-stage founders, this means two things: first, you can build a differentiated product without exclusive access to proprietary models; second, investors will ask how you plan to control costs and maintain margin as you scale. Fine-tuning an open-source model for your vertical gives you both.
Margins Are Expanding as Teams Optimize
The report also shows that margins are expanding across AI-native companies as teams learn to optimize inference costs, reduce prompt complexity, and fine-tune smaller models for specific tasks. The panic of 2023—when every AI call seemed to cost dollars—has given way to disciplined engineering.
This is where technical execution matters. Founders who ship fast but with architectural discipline—choosing the right model size, caching intelligently, batching requests—can demonstrate sustainable unit economics from day one.
Key Takeaways for Founders
- AI must be the majority of your business, not a feature. Investors want products where AI drives revenue and workflow ownership, not auxiliary tools.
- Vertical applications are the path to market leadership. Pick an industry or workflow where you can own the entire process and monetize usage, not seats.
- Open-source models offer cost control and differentiation. Fine-tune for your domain to build moats around data and performance.
- Show sustainable unit economics early. Demonstrate that your product can scale margins through intelligent model selection and optimization.
- The Builder's Economy rewards shipping, not theorizing. Investors back founders who have working products in market, gathering real usage data.
Ship a Working MVP in Days, Not Quarters
ICONIQ's report makes clear that the market has moved on from slide decks and prototypes. Founders who can show a live, vertical AI product with real users and usage data will capture investor attention and early revenue. The challenge is speed without sacrificing quality—building something that works, sells, and scales from the start.
Sources: https://www.saastr.com/the-builders-economy-10-metrics-from-iconiqs-newest-2026-state-of-ai-report/