Gartner puts worldwide IT spending at $6.31 trillion for 2026, growing 13.5%. After thirty years of IT value research, the question of what it all produces is still not settled.
Gartner's April 2026 forecast puts worldwide IT spending at $6.31 trillion for 2026. That is 13.5% growth from 2025. Data center systems alone are projected to pass $788 billion. When I read those numbers, my first thought is not "wow, technology is taking over." My first thought is: what fraction of that is producing measurable value, and what fraction is organizational inertia dressed up as investment?
The question is not rhetorical. It is the foundational empirical problem of IT value research, and after thirty years of serious IS scholarship on it, the honest answer is that we still do not have a clean predictive model. We have frameworks. We have mechanisms. We have firm-level and sector-level evidence. What we do not have is a reliable way to look at a specific IT investment, in a specific organizational context, and predict whether it will produce performance gains commensurate with its cost. The $6.31 trillion is real capital allocated in real organizations by real managers, and a substantial portion of it will not produce what the business cases promised.
Brynjolfsson's 1993 productivity paradox named the problem at the macroeconomic level, and the early 2000s resolution at the firm level was genuine but partial. Brynjolfsson and Hitt (1996) showed that IT capital has a higher marginal product than non-IT capital when organizational structures support it. The key phrase is "when organizational structures support it." The productive firms are not simply the ones that spent more. They are the ones that spent and simultaneously restructured decision-making, empowered workers, and realigned processes. Those organizational changes are not visible in the IT spending figure. They are the complementary conditions that convert spending into value.
Melville, Kraemer, and Gurbaxani (2004) formalized this through their integrative model. IT resources interact with complementary organizational resources to improve business process performance, which then drives organizational performance, moderated by the competitive environment and trading partner characteristics. The model is important because it refuses the direct link. There is no path from IT spending to organizational performance that bypasses business process change and complementary organizational investment. That means every dollar in the $6.31 trillion that flows to an organization without the organizational capacity to restructure around it is spending that the Melville model predicts will not produce the expected performance gains.
Bharadwaj (2000) made the complementary point from the Resource-Based View. IT capability, defined as the organizational ability to mobilize IT infrastructure, human IT resources, and IT-enabled intangibles, is what produces sustained performance differences between firms. The infrastructure is necessary but not sufficient. Two organizations can spend the same amount on cloud infrastructure and produce very different performance outcomes, because one has the organizational capability to deploy it well and the other does not. Barney (1991) gave us the VRIN framework to understand why: the infrastructure is not rare or hard to imitate, but the capability to use it well can be, because it is embedded in organizational routines, skills, and culture that are socially complex and path-dependent. The $6.31 trillion is spending on a mixture of infrastructure, which is commodity, and capability, which is scarce.
What troubles me about reading the IT spending forecast in 2026 is that the AI component is distorting the signal in a particular way. A large portion of the 13.5% growth is AI-related: compute infrastructure, AI software licensing, consulting fees for AI implementation programs. McKinsey's finding that only 7% of organizations have fully scaled AI sits right next to Gartner's $6.31 trillion. The spending is growing at the fastest rate in years. The scaling rate is not. That combination is precisely the structure of the productivity paradox: spending outpaces demonstrated value, and the gap between them accumulates in the firms that bought without first building the organizational conditions for exploitation.
I think the IT value research agenda has two urgent questions right now. The first is how to measure value from AI-embedded systems where the productivity gains are diffuse and the quality improvements are not captured in standard output metrics. This is Brynjolfsson's mismeasurement problem reasserting itself in a new domain. The second is how to explain and predict the gap between IT spending and organizational performance across different organizational capability profiles. This is the Melville et al. mechanism question: what are the specific complementary resources and process changes that convert AI spending into performance, and which organizations are structurally positioned to make those changes?
The $6.31 trillion does not answer either of those questions. It just makes them more urgent. When IT spending is modest, the cost of poor value extraction is modest. When it is the largest single investment category in the modern economy, exceeding the GDP of most countries on earth, the cost of poor value extraction is enormous and it mostly falls on organizations and their employees rather than on vendors, who get paid regardless.
Thirty years after Brynjolfsson named the paradox, the research question it raised has not been answered in a general way. We know the answer depends on complementary conditions, organizational capabilities, and the competitive environment. We do not know precisely which conditions matter most, in what sequence, and for which types of technology. The $6.31 trillion is the urgency case for finding out.
About the author
Share
More notes
Related notes