ICONIQ AI Report: Margins, R&D, and Pricing Trends Revealed

ICONIQ Capital has released its January 2026 "State of AI: Bi-Annual Snapshot" report, offering crucial insights into the evolving landscape of artificial intelligence. The report, based on a survey of approximately 300 executives from software companies developing AI products, reveals a significant shift from experimental AI adoption to the challenge of scaling AI into sustainable, economically viable solutions. This marks a pivotal moment where building AI features is no longer a competitive advantage but a fundamental requirement.

The survey respondents include CEOs, Heads of Engineering, AI, and Product, CROs, and CFOs, representing companies with annual recurring revenues (ARR) ranging from under $5 million to over $1 billion. With 85% of companies headquartered in North America and 15% in Europe, the report provides a comprehensive overview of current AI trends. This second bi-annual survey from ICONIQ Capital allows for the observation of meaningful longitudinal trends, highlighting five critical takeaways for founders navigating the AI era.

1. Vertical AI Applications Drive Value Creation

The report indicates a clear trend towards specialized AI solutions, with nearly 70% of companies now developing vertical AI applications, a notable increase from 59% six months prior. Furthermore, 49% of teams identify application-layer innovation—such as unique user experiences, workflows, and integrations—as their primary source of differentiation.

The underlying model layer is rapidly commoditizing. Companies now utilize an average of 3.1 model providers, up from 2.8 previously. While OpenAI maintains dominance at 77%, Google/Gemini has seen a significant jump to 55% adoption (from 43%), and Anthropic/Claude stands at 51%. The message is clear: competitive advantage lies not in developing foundational models, but in building innovative applications on top of them. Success will favor companies that deeply understand specific industry workflows and create end-to-end AI solutions for those challenges. For more details, refer to the full report: ICONIQ Capital State of AI Report

2. AI Gross Margins Improve with Discipline

Projected average AI product gross margins are set to reach 52% in 2026, a significant improvement from 41% in 2024. However, this growth is contingent on strategic discipline. Companies demonstrating "balanced differentiation"—combining model and product innovation—report the highest margins at 53%, while pure application-layer companies sit at 45%.

The cost structure is also evolving as products scale. Talent costs are decreasing from 32% to 26% of total spend, while model inference costs are rising from 20% to 23%. This highlights that long-term margin leadership depends heavily on astute model selection, efficient routing strategies, and robust infrastructure. The most successful teams are directing the majority of tasks to smaller, more cost-effective models, reserving frontier models for only the most complex cases. Inefficient use of high-cost models, such as GPT-4 for tasks that GPT-3.5 can handle, directly impacts profitability. For more insights, view the report: ICONIQ Capital State of AI Report

3. AI Pricing Remains Unsettled, with Many Companies Planning Changes

The AI pricing landscape is dynamic and complex. While 58% of companies still rely on traditional subscription/platform pricing, usage-based (35%) and outcome-based (18%) models have seen substantial growth, with outcome-based pricing jumping from just 2% in Q2 2025.

A striking 37% of companies plan to modify their AI pricing model within the next year. The primary drivers for these changes include customer demand for consumption-based or outcome-based pricing (46%), a desire for more predictable pricing (40%), and competitive pressures (39%).

The report suggests an emerging best practice: initiate with a light subscription for platform access combined with usage-based pricing when outcomes are uncertain. As outcomes stabilize, transition towards a heavier subscription model for predictability. Hybrid models incorporating pricing safeguards, such as annual commitments (49%) and tiered overage rates (29%), are becoming a pragmatic solution. For further details on pricing strategies, consult the report: ICONIQ Capital State of AI Report

4. R&D Budgets Increasingly Dominated by AI

Companies are significantly increasing the proportion of their R&D budgets allocated to AI development. For companies under $100 million in revenue, AI's share of R&D surged from 25% to 45%. Mid-sized companies ($100-250 million) saw an increase from 15% to 36%, and larger companies ($250-500 million) from 15% to 33%.

Notably, high-growth companies (those with 100%+ year-over-year ARR growth) are investing even more aggressively, dedicating 57% of their R&D to AI, compared to the average of 38%. This trend underscores a clear correlation: the fastest-growing companies are those making the most substantial investments in AI development. Companies not allocating significant engineering resources to AI risk falling behind. Access the full report for more data: ICONIQ Capital State of AI Report

5. AI Acts as a Force Multiplier, Not a Headcount Killer (Yet)

Despite widespread speculation about AI replacing jobs, the data presents a nuanced picture. 42% of companies report no significant impact on headcount plans due to AI adoption. While 35% report a slight decrease, 15% are actually increasing headcount, specifically hiring for AI-related roles.

The productivity gains, however, are undeniable. High-growth companies report that 36% of their code is now written with AI assistance, up from 29% six months ago. Content generation and documentation show impressive productivity improvements of 35-42%, while coding assistance, testing, and code review see gains of 31%.

The reality is a shift in workforce composition. Companies are prioritizing AI-fluent talent and de-emphasizing administrative and repetitive roles. AI is empowering top performers to be significantly more productive and accelerating the upskilling of newer employees. For detailed findings, refer to the report: ICONIQ Capital State of AI Report

The Execution Era Has Begun

The report concludes that the experimentation phase of AI is over; we are now firmly in the execution era. Companies achieving success in AI are excelling across four key dimensions:

• Product: Developing vertical applications with deep workflow integration.
• Cost: Implementing efficient model routing and infrastructure management.
• Trust: Establishing robust evaluation frameworks and mitigating hallucination risks.
• Go-to-Market: Crafting pricing models that align with value delivery.

For founders building AI-powered B2B products, the focus should shift from model selection to ensuring application-layer differentiation, maintaining a sustainable cost structure, and effectively capturing the value created through strategic pricing. The time for execution is now.