Telecoms beat tech on AI payback as KPMG survey finds returns gap
Only 56% of tech CEOs report significant AI value versus 70% in telecoms, KPMG finds, as 28% of tech leaders demand AI payback within six months.

Why it matters
- 56% of technology CEOs say AI has delivered significant or transformational value, versus 70% of telecoms CEOs (KPMG, 19 Nov 2026).
- 28% of tech CEOs expect AI investment returns within six months, against a 3% cross-sector average.
- 88% of tech CEOs and 82% of telecoms CEOs rank AI among their top investment priorities.
- 94% of tech CEOs and 96% of telecoms CEOs cite concern about AI's environmental impact.
- 92% of CEOs in both sectors expect headcount to grow over the next three years.
Only 56 per cent of technology CEOs say AI has delivered significant or transformational measurable value, compared with 70 per cent of telecoms CEOs, according to new KPMG research that points to a widening gap between expected payback and realised results across the two sectors.
The finding lands at an awkward moment for tech. Technology chief executives are the most impatient investors in AI of any industry: 28 per cent expect a return on their AI investment within six months, against a 3 per cent average across all sectors surveyed. The faster they want the money back, the harder the results question becomes.
The survey, dated 19 November 2026, captures a sector under pressure to justify rising AI spending. It also suggests the conversation has moved on from whether to adopt AI to whether the economics of that adoption actually work.
What does the survey actually show?
KPMG's data splits the picture into investment priorities, adoption confidence and realised value. On the first two, tech CEOs look bullish.
- 88 per cent of technology CEOs rank AI among their top investment priorities; 82 per cent of telecoms CEOs do the same.
- 74 per cent of technology leaders say they are moving at the right pace or ahead of the curve on adoption, versus 61 per cent in telecoms.
- 80 per cent of technology CEOs and 76 per cent of telecoms CEOs expect agentic AI to have a significant or transformational impact, well above the 57 per cent figure across all sectors.
The gap sits in the results column. Telecoms leaders report higher levels of measurable value from AI than their technology counterparts, even though they are less confident about their pace of adoption. The implication, on KPMG's reading, is that telecoms operators may be further ahead in turning AI spending into practical gains, while technology companies face greater pressure to defend their budgets.
The takeaway: enthusiasm for AI is nearly universal in both sectors, but the returns are arriving unevenly, and tech is currently on the wrong side of that line.
"We're entering the AI economics phase"
Chad Seiler, Americas Head of TMT at KPMG US, framed the shift in blunt terms. "We're entering the AI economics phase now," he said. "Over the next few years, I think we're going to see a real separation between companies that turn AI investment into revenue, margin and enterprise value and those that struggle to make the economics work."
Seiler also pointed at the capital being poured into the physical layer of AI. "At the same time, the amount of capital going into compute, data centers, networks and power is extraordinary. How efficiently we power that infrastructure, who bears the cost and whether it ultimately creates enough economic value to justify the investment are becoming much bigger questions."
His conclusion raises the stakes beyond individual company results: "I think those questions are going to shape the technology and telecoms industries as much as the technology itself."
The takeaway: the industry's binding constraint is shifting from model capability to whether the capital stack behind it pays for itself.
How confident are the CEOs?
Company-level confidence remains relatively strong despite broader economic uncertainty. KPMG found 84 per cent of technology CEOs and 82 per cent of telecoms CEOs were confident in their own companies' growth prospects.
Confidence in the wider economy is thinner. It sits at 72 per cent among technology CEOs and drops to 57 per cent among telecoms CEOs.
Earnings expectations follow the same pattern. Among technology executives, 36 per cent expect growth above five per cent and 13 per cent forecast double-digit growth. In telecoms, 22 per cent expect earnings growth above five per cent.
Anna Scally, Global Head of TMT at KPMG International, said belief in the company is not enough on its own. "Technology and telecoms CEOs remain confident that their organisations can outperform the uncertainty around them," she said. "But delivering on that confidence will require more than continued investment. It will depend on their ability to turn technology investment into sustainable value, while strengthening the critical infrastructure and skills needed to support future growth."
The takeaway: executives trust their own firms more than the economy around them, and they know investment alone will not close that gap.
Why is infrastructure the pressure point?
The survey also documents the strain AI places on infrastructure, energy use and risk management. Concern about AI's environmental impact is close to universal: 94 per cent of technology CEOs and 96 per cent of telecoms CEOs cited it.
Sustainability appears to be moving from slide deck to balance sheet in the technology sector. Some 70 per cent of technology CEOs said sustainability was fully embedded in their business, up from 59 per cent the previous year. A further 63 per cent said sustainability was being integrated more deeply into core strategy, up from 49 per cent.
As AI systems demand more computing capacity, spending on defence is rising too. The survey found 51 per cent of technology CEOs and 59 per cent of telecoms CEOs increasing investment in cyber and digital risk mitigation.
Regulatory preparedness is the weak spot. Some 42 per cent of technology CEOs and 34 per cent of telecoms CEOs said they were only moderately prepared for evolving rules covering AI, cyber, privacy and platform safety. Rising investment in risk mitigation has not translated into confidence about the rulebook itself.
The takeaway: the compute build-out is forcing both sectors to spend on power, sustainability and security faster than they can prepare for the regulation coming with it.
What happens to the workforce?
Both sectors are redesigning how humans and machines work together. Eighty-four per cent of technology CEOs and 88 per cent of telecoms CEOs said they were redesigning roles, career paths and team structures to reflect that shift.
That reorganisation does not mean shrinking payrolls. In both sectors, 92 per cent of CEOs said they still expected total headcount to grow over the next three years.
Scally said the impact will differ by industry, with agentic AI posing a specific challenge to software business models. "The next phase will look different across technology and telecoms," she said. "In technology, agentic AI could challenge traditional software economics and reshape where value is captured. For telecoms, rising demand for connectivity, compute and trusted data creates an opportunity to move beyond traditional connectivity and capture more of the value created by the AI economy."
Her overall verdict sets the bar for both: "Across both sectors, the winners will likely be those that can convert AI and infrastructure investment into sustainable growth, stronger economics and higher enterprise value."
The takeaway: the workforce question is being reframed around role design rather than headcount cuts — for now, executives expect to hire more people, not fewer.
Why does this matter?
The KPMG numbers capture an industry at an inflection point. Tech firms are spending the most, expecting the fastest returns and reporting the weakest measured value; telecoms operators, long cast as the slower half of the TMT pair, are currently converting AI spend into results more reliably. If that pattern holds into the "AI economics phase" Seiler describes, the winners of the next few years will be decided less by who builds the best models and more by who can pay the compute, power and data centre bill while still producing revenue, margin and enterprise value.
Source: itbrief.com.au
More from Sophie Lindqvist
Show full bio
Staff writer covering regulation, policy and their impact on artificial intelligence and machine learning at AI In Context.
777 articles
Related articles
- Only 10% of Enterprises Let AI Agents Decide Without Human Approval
- Productivity leads enterprise AI, but autonomy remains scarce
- Enterprises Kill 46% of AI Pilots Before Production, S&P Global Finds
- 51% of AI projects miss targets, yet 85% of firms raise budgets
- Forrester: AI bubble fears premature, but math is getting harder