AI Economics — March 27, 2026
AI Labs Spend $25B. Harvey Raises at $11B.
Here Is Who Actually Captures Value.
AI labs spend $25 billion per year running frontier models. Harvey raised at $11 billion building legal agents on top of them. Here is where the money actually goes, who captures value in the AI stack, and the gap between what agents cost and what they can do.
Sources: OpenAI financials; Harvey funding announcement; Epoch AI agent capability data; a16z AI market report 2026.
Global enterprise spending on AI agents is projected to reach $47 billion by the end of 2026, up from $18 billion in 2024 (Gartner). 79% of organizations have adopted AI agents to some extent (PwC 2025). 40% of enterprise applications will embed AI agent capabilities by year-end 2026 (Gartner). 86% of respondents in NVIDIA‘s 2026 State of AI report said their AI budgets will increase this year. The money is real. The question everyone avoids asking is simpler: who is actually making money from AI agents, and who is just spending money on them?
The answer, as of March 2026, is that the infrastructure layer is profitable, the platform layer is growing revenue, and the application layer is mostly still proving ROI. The economics of AI agents follow a pattern similar to several previous enterprise technology waves: the companies selling picks and shovels tend to profit first. The companies using the tools profit later, if their implementation is disciplined. The companies buying tools without a clear unit economics framework are at the highest risk of not profiting at all. This section distinguishes disclosed financial figures (revenue, spending, deal terms as reported) from this publication’s inference about unit economics and value capture, since the latter involves judgment calls the underlying disclosures do not make for us.
The Three-Layer Economics
The AI agent stack has three economic layers, and the profit distribution is not equal across them, based on the disclosed figures below.
The infrastructure layer (GPU compute, cloud capacity) is dominated by NVIDIA, which sells the hardware, and the three hyperscalers (Microsoft Azure, Amazon AWS, Google Cloud) which sell the compute. This layer is unambiguously profitable per each company’s disclosed financials. NVIDIA’s data center revenue exceeded $115 billion in fiscal 2026. AWS, Azure, and Google Cloud all reported double-digit growth driven by AI workloads. The infrastructure providers profit regardless of whether any individual enterprise’s AI agent deployment succeeds or fails, because they charge for compute consumed, not value created; that structural observation follows directly from the pricing model, not from an inference this publication is making.
The platform layer (model providers and agent frameworks) includes OpenAI, Anthropic, Google, Microsoft (Copilot Studio), Salesforce (Agentforce), and ServiceNow. These companies charge per API call, per seat, or bundle agent capabilities into existing enterprise licenses. Revenue is growing rapidly per company disclosures. OpenAI’s annualized revenue reportedly exceeded $11 billion in early 2026. Salesforce and Microsoft are embedding agent features into existing enterprise agreements, which increases lock-in but makes it difficult, including for this publication, to isolate the revenue contribution of agents specifically from the disclosed figures.
The application layer (enterprises deploying agents for their own operations) is where the economics get murky, and where the numbers below mix disclosed figures with reasonable but inferred unit-economics framing. Enterprise AI agent deployments cost $150K to $800K for initial setup with $50K to $200K in annual operating costs (Sustainability Atlas analysis). Organizations report 40 to 60% reductions in manual processing time and 30 to 60% cycle time reductions in targeted workflows, per the cited surveys. But integration costs regularly exceed initial estimates by 30 to 50%, per the same sources. And the critical metric, cost per successful task versus the cost of the human equivalent, appears positive for narrow, high-volume tasks and negative for complex, low-volume tasks, based on the case studies available; this publication has not independently audited any individual company’s internal unit economics.
The Unit Economics Problem
The central tension in AI agent economics in 2026 is what AnalyticsWeek calls the “inference paradox”: while the unit cost of AI is down (token prices dropped 95% since 2023, per that report), total enterprise spending is up because volume has exploded. An autonomous agent that reasons in loops hits the LLM 10 or 20 times to solve one task. RAG systems send thousands of pages of context with every query. Always-on monitoring agents consume compute 24/7. Inference now accounts for 85% of the enterprise AI budget, per AnalyticsWeek’s reported figure.
The unit economics test is straightforward in principle: if an AI agent saves a customer service representative 15 minutes of work but costs $4.00 in inference tokens to run, the ROI is negative on that specific comparison. The winning deployments in 2026, per the case studies this piece cites, are the ones where the task is high-volume, the agent’s token consumption is optimized, and the human-equivalent cost is high: insurance claim processing (10,000 claims/month, $370K monthly savings, 2.3-month payback), IT ticket triage (60 to 80% deflection rate), purchase order automation (80% of transactional decisions automated, $15M annual savings at Danfoss). The losing deployments, by the same reasoning, are the ones where the task is complex, the agent loops extensively, and the human being replaced was not expensive enough to justify the compute cost; this framework is this publication’s synthesis of the pattern across the cited cases, not a universal formula independently verified for every deployment.
Who Actually Profits (Disclosed Figures vs. This Publication’s Inference)
The FinOps for AI Discipline
A new discipline is emerging in 2026: FinOps for AI. The concept mirrors the original FinOps movement that brought cost accountability to cloud computing. The goal is not to cut AI costs. It is to optimize unit economics so that every dollar of inference spending generates measurable business value. The key metrics are shifting from technical (latency, accuracy) to financial: cost per resolved ticket, human-equivalent hourly rate (comparing agent compute cost to the human labor it replaces), and revenue velocity (how much faster a deal moves from lead to closed when AI handles qualification).
The tiered compute strategy is the primary cost optimization lever practitioners in this space describe. Route simple queries to small, cheap models. Route complex queries to larger, expensive models. Cache frequent responses. Compress context windows. Kill idle agents. Whether a given company is “getting this right” or “getting it wrong” in the framing below is this publication’s characterization of a pattern observed across public case studies, not a claim about any specific named company’s internal engineering practices: some organizations appear to be treating inference optimization as a first-class engineering problem, while others appear to be running larger, more expensive models for tasks a smaller fine-tuned model could handle at a fraction of the cost.
The enterprise AI agent market in 2026 is real, growing, and, per the case studies available, economically viable for disciplined deployers. It is also a market where 60% of projects fail per the cited survey, where the infrastructure providers capture profits documented in their own financial disclosures while application deployers take on implementation risk, and where the difference between a positive and negative ROI often appears, based on the available case studies, to come down to whether someone measured cost per successful task before signing the compute contract. The $47 billion in enterprise agent spending will likely generate substantial value for some companies and substantial waste for others. Disciplined unit-economics practice looks, from the available evidence, like a major differentiator between those outcomes, though this publication cannot rule out that other factors this dataset does not capture also play a role.
Sources: Gartner Market Guide for AI Agent Platforms (enterprise spending projections); NVIDIA 2026 State of AI Report; PwC 2025 (adoption data); AnalyticsWeek (inference economics analysis); Sustainability Atlas (deployment cost benchmarks); NovaEdge Digital Labs (implementation guide); Forrester TEI study on Microsoft Foundry (327% ROI, February 2026); G2 Enterprise AI Agents Report; Danfoss case study; Apify (production deployment analysis); Harvey funding disclosures. Updated 2026-08-18: labeled valuation and value-capture conclusions (Harvey’s multiple being “justified,” the profit-distribution categorization, which companies are “getting inference optimization right”) explicitly as this publication’s interpretation, and separated those from the disclosed financial figures (revenue, spending, deal terms) they are based on.
One pattern worth watching, in this publication’s reading of the available disclosures: the bundling strategy. Microsoft, Salesforce, and ServiceNow are embedding agent capabilities into existing enterprise agreements rather than pricing them separately, per their public pricing pages and packaging announcements. This removes the procurement barrier (no new budget line item) but also, in this publication’s assessment, obscures the cost. When an enterprise pays $150 per seat per month for Salesforce and agent features are “included,” the cost of agents becomes difficult for the buyer to isolate. It can appear free. Seat prices have risen at some of these vendors over the prior year in ranges reported around 15 to 20%; whether that increase specifically funded agent feature development, as opposed to other product investment, is this publication’s inference rather than something the vendors have stated directly. The broader pattern, that vendors profit from bundling regardless of whether the agent features deliver value to any specific customer, mirrors dynamics observed in the earlier SaaS pricing expansion, in this publication’s reading of that history.