Enterprise spending on artificial intelligence is showing a widening gap between the companies investing most aggressively in the technology and the broader group of adopters.
According to data highlighted by Andreessen Horowitz (a16z) in its State of Markets II report, which tracked enterprise AI-vendor spending through July 2026, the median AI-vendor spending of companies in the top 1% has risen to roughly eight times that of companies in the top 10%.
By July, median spending among the top 1% had climbed to approximately $800,000, while spending among the broader top 10% remained below $100,000.
Register for the next Tekedia Mini-MBA.
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
The divergence became particularly visible from late 2025, suggesting that AI adoption is not progressing evenly across enterprises.
Instead, a relatively small group of companies appears to be moving from experimentation into much larger-scale deployment, committing significantly more capital to AI vendors, cloud infrastructure, models, and related tools.
The spending gap is significant because it points to a two-speed AI adoption cycle. On one side are companies that remain in the early stages of evaluating and deploying AI tools.
On the other are a small number of large adopters that are spending at a substantially higher level as they integrate AI into core business processes.
For the leading adopters, AI spending increasingly extends beyond individual productivity tools. Companies are using AI across software development, customer service, data analysis, automation, and other enterprise workflows, creating demand for more computing power and model usage.
That dynamic also has implications for the companies supplying the AI infrastructure. As enterprise deployments become larger, demand can flow through the broader AI stack, including cloud platforms, computing infrastructure, model providers and specialized software vendors.
The concentration also helps explain why the AI market continues to generate substantial infrastructure demand even while many businesses remain cautious about large-scale deployment.
A relatively small number of high-spending companies can account for a disproportionate share of overall AI consumption, creating substantial demand for compute and model capacity before adoption becomes widespread.
At the same time, the data does not necessarily mean that most companies are rejecting AI. Rather, it indicates that enterprise adoption remains uneven.
The widening spending gap could represent a transition period in which leading companies are moving first into production-scale AI while other businesses are still testing use cases, determining returns and establishing the infrastructure needed for broader deployment.
For the AI industry, the trend therefore highlights an important characteristic of the current market: AI adoption is expanding, but the intensity of spending is concentrated among a relatively small group of early enterprise leaders.
Perhaps one of the most important findings in the report is that nearly 30% of S&P 500 companies report some quantifiable impact from AI, but only around 2% report having a tracked metric for that impact.
The distinction suggests that many companies are already experiencing or identifying benefits from AI, but relatively few have established mature systems for measuring those benefits.
The same pattern appears with AI agents. Although agentic AI has become a major focus of enterprise technology development, only a small proportion of users are deploying agents at meaningful scale
If adoption broadens over time, the current concentration could eventually give way to a much wider distribution of AI spending across businesses.
Outlook
Looking ahead, the concentration of enterprise AI spending is likely to remain a defining feature of the market as companies move at different speeds from experimentation to production-scale deployment.
The leading adopters are expected to continue increasing spending as AI becomes embedded in software development, customer service, analytics, automation, and other core business functions.