Gartner predicts that by 2028, 70 percent of enterprises will abandon agentic AI systems built with vendor assistance as costs climb and customers struggle to modify the technology without outside help.

What the consultancy calls forward-deployed engineering (FDE) embeds a vendor’s engineers with a customer to build and deploy software for its particular requirements. Gartner argues that the model can deliver rapid early progress but leave customers dependent on expensive external expertise.

According to Gartner, FDE engagements can fail when customers do not acquire the knowledge and control needed to maintain and develop the resulting systems after the vendor leaves.

FDE “success starts with getting the engagement structure right, from scope and incentives to governance, ownership, and exit,” said Gartner senior director analyst Mukul Saha.

“The best-scoped FDE engagements have clear guidelines on governance, business value delivery, IP ownership, project co-ownership, knowledge transfer, and an exit strategy from day one,” he added.

“Many providers now use ‘forward deployed’ as a label for implementation, professional services, solution engineering, or AI consulting; some thoughtfully, others because it sounds more strategic. Some charge premium fees without the delivery depth, program management, or change management maturity to justify them.”

Gartner advises software engineering leaders to use FDE only for problems requiring deep product expertise, rapid adaptation, or close integration between the vendor’s technology and the customer’s operating environment.

Otherwise, customers may see rapid early progress without building the internal capabilities needed to take control. They could then end up paying premium rates for work that a traditional services or partner model might deliver more cost-effectively and predictably.

Gartner also predicts that through 2028, fewer than 20 percent of FDE engagements will turn recurring customer requirements into features in the vendor’s core product. It warns of a growing risk of “FDE washing,” in which conventional consulting services are marketed as something more specialised.

This isn’t the first time that Gartner has warned of the potential costs and pitfalls of AI projects.

Earlier this year, it forecast that at least half of generative AI projects will exceed their budgets due to poor architectural choices and lack of operational know-how, while most organisations that try to build custom models will abandon them for similar reasons.

It also said it expects 40 percent of AI agent deployments would be scaled back or decommissioned as organisations run into governance problems with the technology.

More recently, the analyst predicted that nearly a third of employees laid off because of AI will need to be rehired again at a future date, often at significantly higher cost.

Worldwide AI spending will increase 49.5% in 2026, to $2.7 trillion, and grow another 36.2% in 2027 as AI usage continues to expand, Gartner predicts in the latest quarterly update to its IT spending forecast.

But there’s been no diversion of money from other areas of IT to feed AI, said Gartner Distinguished VP Analyst John-David Lovelock. CIOs got some net new money for AI back in 2024, and a little in 2025, he said, “so there wasn’t a diversion, and now more of their spending is about rebranding than diversion.”

By 2030, “every dollar is going to be an AI dollar in one way or another,” he said.

Instead of buying just a laptop, companies now buy a laptop with AI chips in it, while enterprise software now has AI embedded in it. And instead of doing a project about business strategy, they now do a project about how AI might change business strategy.

“In one way, it’s kind of rebranding,” he said. “Net new spending is going more towards AI, and existing spending is being transformed towards AI.”

It’s similar with hyperscalers: “The hyperscalers haven’t diverted one dime away from their cloud. So AWS, Google, Microsoft, Meta are all continuing to build out their cloud infrastructure at the same rate they were in 2022.”

The AI buildout “is the largest infrastructure project humanity has ever undertaken,” he observed.

Gartner updates its spending forecasts quarterly. In its latest report, it predicted AI infrastructure spend will grow 51.2% this year, and AI software 60.2%. Spend on AI agents and assistants, which it breaks out separately from other software, is forecast to increase by 77.3%, and AI security spend to almost double.

As AI development platforms prove themselves, Gartner sees spending on them accelerating: It has increased its 2026 growth forecast for them from 28% in its May report to 39%, attributing the change to the development of custom AI applications by enterprises and software and service providers.

AI model providers are seeing a “small but growing opportunity” for domain-specific language models aligned with enterprise use cases, upping the 2026 growth rate for generative AI models to 117%.

Companies are getting more selective about replacing legacy kit such as mainframe systems, and rising AI investment is one of the factors causing them to hold onto these existing assets for longer.

Managed services biz Ensono claims in its 2026 State of IT Modernisation report that 78 percent of IT decision makers regard legacy systems as more important today than they did two years ago, and this is because such systems are viewed as key to making AI work for their organisation.

The report is based on a survey of 500 IT decision-makers and line-of-business leaders conducted by the firm across the US and UK.

Its findings show that AI is reshaping modernisation priorities, but that corporates are still grappling with many of the challenges that have traditionally complicated modernisation efforts, such as budget overruns, delayed initiatives, and talent shortages.

Ensono found that 45 percent of firms are actively scaling AI deployments across their organisation, while 44 percent are investing in targeted, high-impact use cases.

Supporting AI, automation, and advanced data initiatives is now the top pressure driving modernisation, while more than half of companies say that AI is helping advance modernisation by delivering through greater automation and improved efficiency.

Among the barriers businesses face, the top challenges to achieving AI goals are said to be difficulty integrating AI into existing workflows and business processes (33 percent of respondents) and infrastructure limitations (28 percent).

For this reason, Ensono says that those legacy systems are taking on new strategic value as organisations attempt to scale their AI rollout. Nearly half of organisations view legacy systems such as mainframes as a critical foundation for AI and an important source of data for this purpose, while 47 percent use them selectively for specific AI applications.

More than half of organisations are optimising and extending legacy systems while modernising applications where they are, rather than replacing them outright. This is an approach that’s slightly more common in the UK (57 percent of respondents) than the US (48 percent).

“Enterprises are discovering that legacy systems, like the mainframe, are intensely powerful, reliable and efficient sources of computing that can now be augmented and made more agile thanks to AI,” says Ensono’s Chief Strategy Officer Brian Klingbeil.

“These systems contain decades of data and business logic that drive many companies. The advantage will go to organisations that understand their environments well enough to distinguish what should be replaced versus what should be augmented, and embrace new modernisation methods that simply were not available even 18 months ago.”

However, using the mainframe as a foundation for AI rollouts was noted by infrastructure services firm Kyndryl two years ago. It found that big iron was becoming a prime candidate to host and run AI workloads, while enterprises were increasingly integrating their mainframes with modern infrastructure.

This meant moving some workloads off the mainframe while updating others in place to continue to benefit from the security and reliability of the platform.

Analyst Gartner also said earlier this year that migrating workloads to a mainframe made more sense for VMware users than adopting Broadcom’s new licenses.

Elsewhere, HPE managing director for UK, Middle East and Africa said that enterprise customers are opting to sweat IT assets for longer, moving refresh cycles from five years to seven years, for example.

“I think we’re seeing earlier planning for critical projects. I think a lot of customers are looking much further ahead on their roadmap of spend at the things they have to do, because maybe some technology is going end of support or the lease is expiring at the datacenter,” he said.

“So I think you’re seeing a decision tree of, do we pause here, and do we sweat the asset for a little bit longer, or is this essential to move ahead? I think it’s fundamentally forcing a rethink for customers.”

Organisations are increasingly using agentic coding tools to build software in-house in lieu of purchasing it, according to a McKinsey report titled “The state of AI in 2026”.

Nearly a third of respondents (32 per cent) report that their organisations have decided against buying one or more software products or features because they could be built internally with agentic coding tools.

The trend spans several sectors, with technology companies leading the shift away from external software procurement. Across industries, 41 per cent of respondents in the technology sector report forgoing software purchases in favour of internal development using AI coding tools.

This is followed closely by healthcare payers and providers at 39 per cent, while both professional services and energy and materials stand at 38 per cent.

Financial institutions register 36 per cent, media and telecom at 34 per cent, and pharmaceuticals and medical products at 33 per cent.

Lower rates appear in sectors such as insurance at 19 per cent and the public and social sector at 17 per cent.

McKinsey Senior Partner Lieven Van der Veken said, “Not only are larger organisations scaling AI faster, but many are beginning to take greater ownership of their technology and change agendas. Until recently, many leaders assumed that AI was beyond the capabilities of their own technology teams, and that moving quickly meant securing the right external partnership.”

Van der Veken added, “Now, the tone is changing. Leaders are asking what their organisations need to build AI tools themselves. The rise of software coding agents and in-house development is one clear sign of this broader shift.”

The report noted that enterprises are progressing beyond the initial phase of AI experimentation. As per the report, nearly nine in ten respondents report regular use of AI in at least one business function. Overall enterprise-scale adoption expanded over the past year, with 44 per cent of respondents reporting that AI is scaling across their organisation, compared to 38 per cent previously.

Furthermore, the proportion of organisations deploying AI across three or more functions grew from 51 per cent to 56 per cent.

Addressing this strategic evolution, Van der Veken said, “This, of course, does not mean companies should build everything themselves or turn away from forging partnerships. It means that leading organisations are engaging partners from a position of agency.”

He stated, “The organisations moving fastest are becoming more deliberate about where to buy, where to build, and where to develop enough internal capability to integrate and scale what works. They are also treating operating costs as a design constraint, not an afterthought, and making the deeper changes in workflows required for value realisation.”

Adoption patterns show larger enterprises outpacing smaller firms in deploying AI capabilities. Fifty-four per cent of respondents from organisations generating at least USD 1 billion in annual revenue report scaling AI enterprise-wide, compared to one-third of smaller organisations.

Larger enterprises also registered faster momentum in adopting agentic AI, with the share scaling agents in one or more functions increasing from 27 per cent to 40 per cent, while smaller organisations remained flat at 22 per cent.

The cost of agentic AI workflows is forecast to increase more than fivefold by the end of 2028 as users adopt more complex applications of the technology.

As Nvidia and other tech giants push inference and agentic AI as the next stage of the AI wave, Gartner warns that the cost of implementing these systems will rise even as foundation models become cheaper.

The analyst firm has cast its eye over the nascent world of AI agents – systems designed to act independently in pursuit of a goal – and sees multiple challenges ahead.

Leaving aside the substantial security concerns, these software agents are considerably more complex than chatbots.

Gartner believes falling model prices are tempting users to build more complex workflows, whose greater token consumption can outweigh those savings and drive up overall inference costs.

In other words, tokens are becoming more cost-efficient, but those savings are not keeping pace with the rising cost of more advanced AI capabilities. The rate of innovation is outpacing the cost curve, Gartner claims.

“The harsh economics of the inference paradox are exemplified by the differences between a simple chatbot and an AI agent,” says Gartner senior director analyst Will Sommer.

“Where a simple chatbot must read and interpret a query and quickly respond with a probabilistic reasonable answer, an AI agent must constantly reason, negotiate, and question itself,” he explains.

Those processes add up: routing a task to an agentic reasoning model increases inference costs at least fivefold, and potentially by much more as the task becomes more complex.

Securing a return on investment from such advanced AI tools therefore demands either much greater returns than basic models provide or better optimisation of inference, routing, and orchestration, Gartner warns. This could mean assigning each task to the most cost-efficient model capable of handling it.

The move by some AI providers from flat-rate subscriptions to usage-based billing hasn’t helped. Token-heavy workflows can produce runaway costs under the new model.

Perhaps it is no wonder Gartner predicted earlier this year that 40 percent of organisations would demote or decommission AI agents because of problems with the heavily hyped technology.

The analyst biz also cheerily forecast last year that at least half of all generative AI projects would blow their budgets because of poor architectural choices and a lack of expertise, while most attempts to build custom models would be abandoned.

Microsoft is moving parts of its AI portfolio beyond pure per-seat pricing by pairing user licenses with charges tied to usage. The hybrid model keeps recurring seat fees in place while adding a variable billing unit as customers consume more AI capacity.

That model is beginning to show up in Microsoft’s financial results. Dynamics 365 Customer Service credit consumption rose fourfold quarter over quarter, while GitHub Copilot generated what Microsoft called “significant consumption revenue” after its June billing change.

Microsoft also reported more than 30 million paid Microsoft 365 Copilot seats, with net additions more than doubling from the previous quarter, and said GitHub continued adding Business and Enterprise seats.

The UK’s Competition and Markets Authority has opened a fresh investigation into Microsoft, this time over how the company is marketing its Microsoft 365 Personal and Family subscription plans.

A cursory glance at the probe highlights that the CMA is trying to prevent Microsoft from sneaking in Copilot and AI features as justification for recent subscription price hikes. A closer look at the investigation displays a pattern of friction Microsoft runs into with regulators regarding its subscription practices, automatic renewals, and software bundling almost every time it adjusts its services.

The CMA made the investigation public on July 29, 2026, though the case itself was opened two days earlier. According to the CMA’s case announcement, the price gap between plans is concrete. The Classic Microsoft 365 Personal plan costs £59.99 a year, compared to £84.99 a year for the Copilot-equipped version, a difference of £25 annually. The Family plan costs £79.99 a year for Classic versus £104.99 a year with Copilot bundled in.

The CMA’s announcement hinges on Microsoft’s change in January 2025, when the company automatically added new features such as Copilot to existing Microsoft 365 plans with no increase in fees for the remainder of the subscription period. Unfortunately, once the plans were up for renewal, customers who were not paying attention were re-enrolled into higher-priced plans. Customers who were paying attention could avoid the bait-and-switch by selecting a ‘Classic’ plan or cancelling during the re-enrolment period.

This price change was covered when it first happened in February 2025, and even then, Microsoft would not directly attribute the increase to Copilot in their own communications. Personal plans jumped from $6.99 to $9.99 a month in the US, and Family plans went from $9.99 to $12.99, with the Classic option only showing for existing subscribers who knew to look for it during cancellation.

CMA regulators are now examining whether Microsoft provided adequate messaging for customers to understand what changes occurred and how much more they were going to pay despite spending months using features that were initially presented as free.

As Senior Director for Consumer Protection, Hayley Fletcher puts it, “When a business changes its subscription plans, customers need clear and timely information about their options.”