New Survey: Technology Outgrew Its Measurement Systems. Even GDP Can't Keep Up.

The instruments used to measure technology’s impact—at both the enterprise and national economy levels—haven't kept pace with what technology has become, according to a new survey from The Technology Economists. Founder Howard Rubin explains why the old instruments no longer work, and what a new one might look like.

Ask U.S. technology and business leaders whether GDP and productivity statistics capture what technology is actually doing to the economy, and the answer is overwhelmingly no. In a September 2026 global survey of senior business and technology leaders from The Technology Economists, 90.7% of respondents say conventional macroeconomic measures capture technology's true economic impact only partially, or not at all. Just 9.3% express confidence in them.

That's a striking number for a statistic most people treat as settled economic fact.

"GDP misses a view that's inclusive of technology economics," says Howard Rubin, founder of The Technology Economists and the survey's lead author. "It was born late in the industrial revolution, in 1934. It'll soon be 100 years old, and it doesn't capture anything about technology capital or production."

Rubin has spent his career building the field of technology economics, and that level of doubt doesn't surprise him. It confirms something he's argued for years. "GDP, the Consumer Price Index, and even the labor indices are archaic," he says. "Classical economics, going back to Adam Smith, focuses on input-output and the allocation of resources. Today we need to focus on technology investment, the cost of technology capital, and its value to nations, businesses, and people."

AI Made the Problem Impossible to Ignore

That doubt deepens when the questions turn to AI. Just 3.7% of respondents call their AI value measurement fully effective—the lowest confidence score anywhere in the survey. Another 51.9% call it limited or insufficient, and 24.1% say they have no meaningful AI measures at all. Altogether, 75.9% describe AI value measurement as limited, insufficient, or simply absent.

AI is compounding a measurement problem that already existed, according to Rubin. "AI and its massive cost are all over the headlines daily. Yet, “enterprise consumers of AI technology can barely understand the cost, and few can really show the value that they are getting in business terms,” says Rubin. “That’s the gap: connecting costs to outcomes, or what I call the ‘value reality’.”

AI programs can generate plenty of activity data—models deployed, users enabled, tokens consumed—without any of it connecting to realized economic value, the survey notes. Rubin distinguishes between AI's role and more familiar infrastructure spending. "AI is aimed at productivity and society as a primary impact," he says. "Cloud and other infrastructure are a different issue. They’re critical to run the business. Their value is in cost optimization, performance, and risk reduction, and measures for those are crystal clear.”

Counting Costs, Missing Value

The problem is simple to state and hard to fix: it’s far easier for organizations to measure what technology costs and does than what it changes and creates.

While 61.1% of leaders surveyed see technology primarily as a driver of growth or transformation, only 7.4% say they can consistently connect specific technology investments to business outcomes. Another 61.1% say they can make that connection only in some cases, and 31.5% say it happens rarely or not at all. When respondents were asked to rank the biggest limitations in how they currently measure technology's value, "failure to capture business impact" topped the list, followed by "overemphasis on cost versus value” and "fragmentation and inconsistency across the organization”.

"What's breaking down is exactly the point," Rubin says. "IT financial management is a big deal—knowing what things cost. That's IT financial management and FinOps and more. What’s missing is IT value management.”

The findings come from a September 2026 survey distributed to technology and business leaders through Rubin's Technology Economists network, the TBM Council, and Heller Report readership. Respondents skewed senior: 37% identified as CIO-, CTO-, or CDO-level technology leaders, 31.5% as strategy or transformation leaders, and 16.7% as CEO, CFO, or COO executives. This isn't a scientifically sampled, statistically representative poll. But the results are still valuable as a directional signal: a real-time read on how senior business and technology leaders are thinking about this problem, even if it doesn't represent the full population of IT decision-makers, says Rubin, who has spent three decades building benchmarking databases on technology economics.

Why This Can’t Wait

None of this is theoretical, Rubin says. Earlier this year, AI agents built by OpenAI compromised infrastructure at Hugging Face, a platform for AI developers, and similar incidents followed at other AI labs—a reminder that AI risk is more than financial. "It's urgent because the expenses and safety implications are now at the forefront," he says. "It's all about making choices that make the value greater than the cost and the risk."

That's the logic behind an idea Rubin raised in a July 2026 article he wrote for Heller on IT inflation: that technology spending needs to be managed like an economy, with a technology economist translating costs and returns into language a board can act on. Asked what a CIO or CFO should do first after reading these survey results, Rubin doesn't hesitate. "Become the company's chief technology economist," he advises, "and deal with the cost-risk-value equation in every decision."

Technology Value Index: Separating Price from Consumption

The Technology Economists report recommends that organizations build what it calls a Technology Value Index: a standardized way to connect technology spending to business outcomes, rather than relying on cost ratios alone. Rubin has already begun sketching what that could look like in practice.

The starting point, he says, is separating true price inflation from simple changes in consumption. "If cloud spend rises because a company doubles its consumption, that's not inflation," he says. "If the same standardized compute unit becomes 12% more expensive, that is." Most organizations conflate the two, he argues, reporting rising technology costs without distinguishing how much reflects buying more versus paying more for the same thing.

That distinction, Rubin says, needs to be made at multiple levels. Enterprises need a technology value index to understand their technology spend, so they can tell whether rising costs reflect genuine price increases or simply heavier use—and negotiate, budget, and report to their boards accordingly. And nations need an equivalent to track technology’s impact on the broader economy. Even long-term technology contracts could use a similar mechanism—one that adjusts pricing based on verified cost changes rather than a generic inflation index, Rubin says.

No More Flying Blind

The technology value gap starts inside the enterprise, where measurement stays anchored in cost and activity rather than outcomes. AI only widens it, creating forms of value—speed, learning, augmentation—that don't fit within conventional IT metrics, such as cost per license, system uptime, or tickets resolved.

The fix, in Rubin's view, is a different discipline altogether with clear measures of what's being invested, what it's producing, and whether the value justifies the cost. "If you're not tracking that," Rubin has argued before, "you're budgeting blind."