Does GenAI Come With Costs That Make Scale-Up a Challenge?

 Scale up is always the aim for insurance brands, rather than niche usage. Here are some thoughts from SAS on the topic;

Dr Iain Brown, Global Head of AI & Data Science at SAS, warns UK and Ireland businesses that it is no longer a case of ‘does GenAI work’ but ‘can organisations control costs’ as they attempt to scale up its use.

In a new piece of research on GenAI usage, which SAS commissioned Coleman Parkes to undertake across 100 senior UKI enterprise technology decision-makers, it found that for some enterprises, GenAI is not just a technology challenge but an economic one. The deeper the level of adoption, the more significant the financial pressures become.

The research was launched today at SAS Innovate on Tour at Anfield, Liverpool. It found the number of UKI enterprises to fully integrate GenAI into their regular processes is at 22%, compared with only 9% in 2024. Among this more advanced group, 41% already cite prohibitive LLM costs as an active implementation barrier, compared to 32% across the overall market. Within this same segment, 45% say GenAI has delivered below-expected ROI, the highest dissatisfaction rate in the research.

But the picture isn’t completely negative. Nearly half of UKI enterprises actively using GenAI report significant improvements in operational costs and time savings, and over two in five say customer satisfaction has significantly improved.

Dr Brown said: “GenAI is genuinely working for the organisations that have deployed it thoughtfully – we see that in the data and hear it from customers. But what we’re also seeing is the first signs of a cost challenge that most UK and Irish businesses aren’t prepared for, because they haven’t moved out of pilots. Nearly half of the organisations that have moved beyond pilots are telling us clearly – the costs at their current scale are becoming prohibitive.

“I’m hearing the same thing directly from enterprise customers we work with. The cost of AI tokens has fallen in recent times, but consumption at such a scale is more than wiping out any savings. They moved first, they scaled first, and they’re feeling the cost challenges first because they didn’t have the luxury of market proof. That’s why governance and cost controls need to become part of GenAI strategies much earlier than many organisations anticipate.”

Why are falling prices not cutting costs?

While the cost of individual AI interactions has fallen dramatically, overall enterprise spending continues to rise because usage grows much faster than unit costs decline. Unlike traditional enterprise software where costs are relatively predictable, GenAI costs scale directly with how much your people use it and most enterprise budgets aren’t built to absorb this.

SAS has spent five decades helping enterprises govern and control complex data and AI systems at scale. That experience increasingly puts it at the centre of conversations about how organisations structure and manage their GenAI usage before costs become unmanageable.

Dr Brown continues: “The organisations we work with that have moved GenAI into production aren’t asking us whether it works anymore, they’re asking how to stop the costs running away from them. That’s a fundamental shift in the conversation, and it’s happening faster than most anticipated.

“What organisations need to do is focus on good governance of their GenAI and AI systems first, then look at how they can control costs and treat GenAI like an enterprise capability first before scaling.”

How are global brands reacting to the challenge? 

Large organisations are already experiencing the realities of AI economics at scale.

Uber burned through its entire 2026 AI coding tools budget in just four months after rolling out tools including Claude Code across its 5,000-strong engineering team, with monthly API costs per engineer running between $500 and $2,000.

Amazon ran an internal AI adoption leaderboard, called Kirorank, designed to incentivise engineers to use AI tools. Employees began generating unnecessary AI workloads to improve their leaderboard position, inflating costs without delivering productivity gains. Amazon subsequently shut the programme down. This same dynamic dubbed ‘tokenmaxxing’ has also been reported at Microsoft and Meta too.

Agentic AI will accelerate the problem, if not governed

The SAS research finds that 30% of all UKI enterprises are already investigating or piloting agentic AI, systems that independently handle multi-step tasks like case handling, onboarding and investigations without human prompting at each stage.

Gartner’s analysis confirmed that agentic AI models require 5–30x more tokens per task than standard generative AI. Where a standard chatbot query generates one round of token consumption, an agentic workflow makes 10–20 separate AI calls to complete a single task. Enterprises that have modelled their agentic business cases on standard GenAI pilots will find real-world costs are much higher than their projections if not governed effectively.

What do businesses need to do to manage GenAI costs?

The SAS data shows difficulty proving ROI is already a barrier for 38% of UKI enterprises, and that’s before the majority have reached production. Dr Brown says businesses can do a few things to get ahead of the challenge;

“Stop measuring GenAI success by how much your people are using it. Governance, budgets and usage controls need to be in place before you scale. And critically, before any organisation moves from pilot to production, especially into agentic AI, they need to stress-test what the costs might look like at full deployment. Good foundations and planning are key to generating strong returns from GenAI.”