B2B SaaS marketers are paying for their employers’ AI transformation out of their own pockets, according to new research that puts numbers on a gap the industry has felt for two years but never measured. The AI Funding Gap, a SaaStrix study of 739 B2B marketers released this week, found that 71% cover at least some of the cost of the AI tools they use for work, and that among those self-funders, 98% spend more than £20 a month of their own money.

The finding lands at a moment when the instruction to produce AI-assisted work has become close to universal while the money behind it has not moved. Gartner’s 2026 CMO Spend Survey found 56% of CMOs say their organisation lacks the budget to deliver this year’s strategy, and 2026 has produced a run of high-profile corporate reversals in which staff were pushed onto AI tools and then watched access disappear when the bills arrived. The study measures what that squeeze looks like from the marketer’s desk.

The full study can be found on the SaaStrix website: The AI Funding Gap.

How many B2B marketers pay for their own AI tools?

SaaStrix’s data comes from 739 B2B marketers — 693 through a structured survey, plus focus groups and one-to-one interviews — most of them at manager or senior-manager level in B2B SaaS companies of fewer than fifty people, split between the UK and North America. Within that group, 84% use AI in their work every day. The support has not followed the mandate: 88% say they are allowed to use AI but given no real guidance, 74% have no dedicated budget for AI learning, and 68% say their employer funded none of the four things that build AI capability in the past year — not a course, not an event, not a community membership, not a training session. Meanwhile 81% paid for at least one of those themselves.

“I’m paying for both Claude and ChatGPT every month. It’s too much of a headache to get everything signed off all the time,” one respondent told the study’s interviewers.

Independent research points the same way. Recruitment firm 3Search’s 2026 survey found 53% of marketers funding their own AI training, and the paper cites broader workforce research putting personal payment for AI tools at around 29% — a rate B2B SaaS marketers are running at more than double. One caveat the paper makes itself: respondents belong to SaaStrix’s AI-focused marketing community, a self-selected group of engaged, self-investing practitioners, so the self-funding rates likely sit above the profession-wide average. The study treats that as the point — these are the marketers setting the pace, and the distance between their effort and their employers’ investment is what it set out to measure.

Why are employers pulling AI access back?

Nearly half of the marketers surveyed (47%) have had an AI tool they used for work blocked, restricted or withdrawn with no funded alternative, and 43% saw an AI or learning budget cut or frozen in the past year. Of those, nearly a quarter (23%) were given no reason at all.

The same pattern has played out at enterprise scale through 2026. Microsoft cancelled most internal Claude Code licences in its Experiences and Devices division with a 30 June cut-off, moving thousands of engineers to GitHub Copilot CLI after token costs consumed the annual AI budget, as The Verge’s Tom Warren first reported on 14 May. Uber ran internal leaderboards ranking teams by AI usage, then its chief technology officer told The Information in April that the company had burned through its entire 2026 AI coding-tools budget in four months; by June, Bloomberg reported a cap of $1,500 a month per engineer, per tool. Accenture told associate directors and senior managers in February that regular AI adoption would be a condition of leadership promotion, in Financial Times reporting the firm confirmed to CNBC — then by June was restricting generative AI for routine tasks, such as converting PDFs into slide decks, as token costs became material, according to 404 Media’s reporting of leaked internal audio.

The sequence repeats: mandate adoption, reward usage, withdraw funding when the invoice lands. For the individual marketer, the paper argues, the worst position is the gap in between — the free tool is blocked, no enterprise licence replaces it, no training is offered, and the expectation to produce AI-assisted work remains. “One email and they told us we can’t use it anymore. Yet the volume of work was still expected. So now I’m using my personal account on my own laptop, which I know I’m not really meant to be doing,” another respondent said in interview.

What is the funding gap costing marketers?

Confidence is low across the whole sample — just 19% of respondents feel confident in their AI skills — but the anxiety concentrates among those paying their own way. Among marketers whose employer funds everything, 11% are seriously concerned about keeping pace with AI. Among the self-funders, the figure is 80% — roughly seven times higher. The people carrying the personal cost are, by a wide margin, the people most afraid of falling behind.

The fear has a backdrop. More than half of respondents (57%) say they personally know someone who lost a marketing job in the past year where AI was cited or suspected as a factor — a figure the paper explicitly reports as perception rather than verified cause. The reward side looks thin against it: 60% say their AI skills have brought them no career return at all — no pay rise, no promotion, not even recognition.

“Companies keep calling themselves AI-first while a £20 subscription takes three approvals to sign off. Marketers stopped waiting. They are funding the transformation themselves, and right now their employers are banking the gains,” said SaaStrix’s founder, a practising CMO.

Why are employers stuck — and who breaks the loop?

Gartner’s data suggests the squeeze is structural rather than simple meanness. In its survey of 401 CMOs and marketing leaders, largely at companies with revenue above $1bn, 56% said their organisation lacks the budget required to deliver its 2026 strategy, with overall marketing budgets effectively flat at 7.8% of company revenue. The paper describes the resulting trap plainly: leadership cannot prove the return on AI without trained people and proper tools, and cannot win budget for trained people and proper tools without first proving the return. Most AI spend is reallocation, funded by cutting elsewhere.

The cost of staying stuck is now measurable. PwC’s 2026 Global AI Jobs Barometer, published on 15 June, puts the average wage premium for AI skills at 62%, up from 57% a year earlier, with jobs requiring AI skills growing roughly eight times faster than the overall market. Every quarter a team spends inside the trap, the skills become more expensive to buy in later.

SaaStrix says the study is part of a commitment to original research built with its community on the reality of marketing work in the AI era, and that the aim is to give B2B marketers a voice — and evidence they can put in front of leadership teams. The findings are free to read in SaaStrix’s public write-up of the study, with the complete paper available through its community.

The wage premium for AI skills has climbed from 25% in 2024 to 57% in 2025 and 62% this year, PwC’s data shows. The capability B2B marketers are buying with their own money travels with them — not with the employers who declined to fund it.