The AI Market Intelligence Index
The ultimate collection of AI spend and momentum insights.
Last updated: Q4 2026
Top 25 AI-native and Applied AI tools (by popularity)
Findings
5 key points from the last 6 months
Over 33% of companies pay for a frontier AI model.
Open AI and Anthropic now dominate AI-native market and hold double the customers of the 3rd most popular tool (Cursor). Businesses are gravitating towards frontier models cecause of their operational flexibility, pace of innovation and multiple use-cases.
Large Applied AI vendors solidify market position further
SaaS juggernauts like Microsoft, Salesforce and Google continue to dominate the Applied AI market, consistently adding AI features to their product sets through innovation and acquisition. Salesforce's purchase of Fin, a former Top 25 Applied AI tool, exemplifies this.
Anthropic has doubled its customer base since July '26.
Anthropic is now level in size with Open AI, proving to be a genuine contender to its global supremacy, and even overtaking it in some markets (such as the UK).
Its blistering rate of innovation, releasing new Fable, Mythos, Opus and Sonnet models in a short 3 month window, has been central to this.
This has enabled it to close the gap to
Business-wide use over single department use
The vast majority of tools in both categories service the needs of multiple departments, rather than focus on a single one. This fits AI's well-defined ability to manage and connect processes across multiple siloes, departments and workflows, in areas like deep data research, documentation analysis and summary and retrieval operations. Engineering tools are the only outliers here, having far more specific use cases.
AI coding and ML development tools fuel agentic AI
"Proven" tools like Cursor and new entries such as CodeRabbit and Weights & Biases indicate that coding and development AI tools are being favored to create agentic AI setups. For example, the Cursor SDK can build agents, CodeRabbit can create the workflows and orchestration for agents to operate, and Weights & Biases trace and run evaluations and results.
Expert opinion
"AI's transformative power for businesses has driven demand to unprecedented levels, creating an AI arms race in every SaaS vertical.
As companies’ boards demand AI adoption, encourage experimentation and race to beat competitors, AI becomes more central to the standard tech stack. Which in turn is causing costs to skyrocket."
Watch
Are frontier models dominating already?
Calum Durrand, Procurement Director at Vertice, thinks that - while this is the current direction - we aren't there yet.
Teams still have all their data and prompts in Applied AI tools like Copilot, and will gravitiate to ease-of-use over better features at the moment.
Insight
As a total of all software spending in the average company, AI spending has increased by 254% YoY.
.json)
Going deeper
AI's increasing presence in budget allocations is down to multiple factors creating a perfect storm of rising costs.
Why is AI spend exploding?
79% of AI vendors are charging a premium to access AI features - sometimes up to 100% more than the original price.
The average value of an AI contract has increased 28% since June 2026. AI has gone beyond its experimentation phase, and vendors feel confident to charge more.
Healthcare, Media & Entertainment and Education industries have doubled their AI-native spending - indicating AI's expanding reach across all industries.
Confidence in embedding AI is growing. The average AI adopter now has 2.5 tools.
Watch
AI-native or Applied AI - who's getting the bigger chunk of investment?
Calum Durrand sees first-hand that AI-native platforms are growing their share of investment from customers faster than Applied AI tools.
And our data backs this up. Average AI-native spend per company has more than doubled YoY, whereas Applied AI spending is up just 33%.
However, an Applied AI contract is 6x larger than an AI-native one in pure dollar terms.
4x
as many companies commit over $250k or more to Anthropic than OpenAI
Insight
Anthropic has almost doubled its customer base in 5 months. From being a distant challenger it’s now a viable alternative to Open AI.
YoY growth for Anthropic and OpenAI
.json)
In June 2026, OpenAI had 39% more customers than Anthropic. In just 5 months, thanks to it's astronomic growth rate, Anthropic has drawn level.
Anthropic vs OpenAI - the key stats
37% of companies now use both OpenAI and Anthropic, and 33% are on paid contracts with both of them.
A typical Anthropic commitment is roughly 2x that of OpenAI.
For every customer that pays for an OpenAI contract, there are 3 that pay for Anthropic.
Anthropic commitments per company have increased more than 80% since April 2026. OpenAI commitments have stayed flat.
Despite OpenAI being purchased first in 4 out of 5 companies, Anthropic now holds the larger contract in 2 out of 3 of them.
Watch
Anthropic is breaking OpenAI's grip on global markets
Businesses seem to prefer Anthropic in the UK and EMEA regions, but OpenAI still reigns supreme in the US.
For Jess Poole, Procurement Director at Vertice, this is due to them having a much larger customer base and being first to the market in their space. But, Anthropic seem to understand where businesses want productivity gains and controls better, which helps explain their increasing popularity everywhere.
Expert opinion
“I'm really interested in seeing how large software vendors deal with OpenAI and Claude. Touted data walls and connection restrictions seem short-term and divisive.
It's likely they will form large mutual partnerships, exchanging hardware, cloud computing services, acess to markets and investment."
17%
The current overpayment for AI contracts vs Vertice pricing benchmarks
Insight
Despite still overspending on AI contracts, businesses have reduced this ratio by 5% since June 2026.
Are companies beginning to control their AI spending? In reality, it's a false dawn. The problem is, the floor has moved.
The rising cost of an AI contract has meant that, in actual dollar terms, a 17% overspend in November 2026 costs 23% more than in June 2026, even though the overpayment rate was higher.
The culprits? Unoptimized token spending coupled with a lack of visibility and control structures that aren't meant for consumption-based pricing models.
% of AI overspend compared to real-term contract cost increases
.json)
Unless AI costs are controlled, any overspending is likely to mean exponentially more in pure monetary terms as time goes on.
Watch
Where should AI spending control sit internally?
Jess Poole, Procurement Director at Vertice, discusses the importance of accountability for AI spendingbeing established, and how this unlocks visibility and meaningful spend optimization.
Expert opinion
"Token-based pricing models can cause costs to spiral very quickly without proper controls. Internal accountability, visibility and contractual guardrails such as overage limits and multi-year spending commitments are your best defences.”
Expert opinion
“AI-native tools, particularly frontier models, are proving their value, flexibility and potential more effectively than Applied AI platforms.
Consider how companies use OpenAI and Claude to cover multiple departments, democratize data access for all users, and customize them to the unique needs of different teams - all in one tool".
Insight
AI-native tools are offering companies far more flexibility and power than an AI tool that has to come as a bolt-on to a larger SaaS platform.
AI-native tools per company
.json)
This is a clear indication that companies are going multi-modal - realizing the power of each AI-native tool and frontier model and deploying them where they work best.
But, will this approach last? Or will AI stacks homogenize?
Watch
Diversifying your AI vendors protects against risk
Relying on one AI tool - particularly a frontier model - is risky at the moment, according to Calum Durrand.
Until risks from rising token costs, cybersecurity, and even how AI vendors change terms of their service, are understood and guarded against, it's beneficial for customers to diversify their AI tools.
