Artificial intelligence, once regarded primarily as a promising area of research, is now exerting a profound and rapidly expanding influence on corporate America. Large and small businesses alike are leveraging machine learning systems, advanced data models, and generative tools to drive growth and efficiency. Yet, paradoxically, despite this explosion of activity visible in the corporate sector, the magnitude of the boom is only faintly discernible within the official records of national economic growth. According to a detailed analysis published by Goldman Sachs, the official government statistics appear to capture only a fraction of AI’s true contribution to the United States economy, leaving a significant portion of its effects essentially invisible in GDP figures.
Goldman’s analysts emphasized the immense scale of the transformation in a recent report. They pointed out that revenues collected by U.S.-based companies building and supplying the infrastructure essential for the deployment of AI — such as cloud computing platforms, specialized servers, and the increasingly indispensable semiconductor chips — have skyrocketed by approximately $400 billion since the beginning of 2022. At first glance, figures of this magnitude would naturally suggest that artificial intelligence is already functioning as a substantial accelerator of national economic expansion. However, when one shifts to the lens of government statistical reporting, a considerably more subdued picture emerges.
The economists at Goldman estimated that AI-related innovation and spending have, in fact, lifted real economic activity in the United States by about $160 billion over the same period, a figure that corresponds to roughly 0.7 percent of national GDP. Yet strikingly, official statistics have managed to register only about $45 billion of this growth — or approximately 0.2 percent of GDP — within the formal accounts of measured output. This leaves an estimated $115 billion of AI-driven activity effectively absent from the government’s tally, a discrepancy that speaks volumes about the challenges posed by measuring the contributions of new and intangible technologies.
The disparity can largely be traced to the methodology employed by the U.S. Department of Commerce’s Bureau of Economic Analysis (BEA), which is responsible for compiling national accounts. Because of the BEA’s framework for calculating GDP, certain critical inputs to the AI economy are treated in a way that minimizes their visible impact. For instance, high-performance semiconductors — the powerful chips that are indispensable for training complex AI models — are classified as what economists call “intermediate goods.” These inputs are deducted from GDP totals when they are imported, and their deployment in building extensive AI capability does not appear in government statistics as an independent form of investment. This accounting rule means that billions of dollars in expenditures are acknowledged in corporate statements yet remain invisible when translated into national-level metrics.
As Goldman’s analysts elaborated, these chips, though currently captured on the books as mere materials used in production, are actually being employed to create systems and models that function as long-term intangible assets. These AI models, once trained, underpin a vast array of future applications ranging from automated services to predictive analytics in healthcare and finance. However, the ultimate economic value generated by such models has not been fully integrated into GDP measurements because intangible, software-driven assets are much harder to quantify and assign in national aggregates than physical products like industrial machinery or consumer electronics. Thus, the current official data significantly underrepresents the full scope of investment occurring beneath the surface.
In fact, Goldman’s estimates suggest that approximately $75 billion spent since 2022 on developing enterprise-level AI models and deploying applications within cloud infrastructures has gone unrecognized within the investment portion of GDP statistics. This omission contributes directly to the $115 billion blind spot and illustrates how contemporary accounting conventions may fail to capture new forms of economic capital that derive primarily from knowledge and software rather than from tangible goods.
At the same time, the situation has been further complicated by shifting trade and import policies. During the first half of 2025, for instance, business investment data exhibited a sudden spike in spending on information-processing equipment such as high-capacity servers and advanced networking gear. This surge, however, was not purely a reflection of organic demand for AI-related equipment. Rather, many companies moved to accelerate purchases in order to secure supplies ahead of tariffs scheduled to be imposed under policies introduced by President Donald Trump. While such frontloaded imports boosted short-term investment figures, they also exaggerated the appearance of underlying trends, creating the risk that the true pace of AI investment might be overstated in certain quarters. Moreover, because imported goods subtract from GDP calculations, much of the apparent boom is effectively negated when incorporated into the official economic accounts.
The difficulty of capturing AI’s impact is not restricted to GDP figures alone; it also extends to other indicators of business performance. Even though a record proportion of S&P 500 companies publicly referenced artificial intelligence during their earnings calls in the second quarter of the year, relatively few of them were able to provide precise quantifications of how these technologies were affecting their profitability. Goldman Sachs, in another recent report, noted that although corporate enthusiasm for AI appears widespread and genuine, the majority of firms remain cautious about assigning numerical values to its contributions at this early stage of adoption. This underscores the broader problem: while AI’s influence on economic structures is undeniable and perhaps transformative, the methodologies available to measure it with clarity and precision remain incomplete and underdeveloped.
Taken together, the evidence illustrates a clear paradox. On the one hand, AI is driving significant investments, fostering the creation of new intangible assets, and delivering tangible improvements in productivity. On the other, the existing frameworks for measuring economic activity seem ill-equipped to capture these effects in real time. This mismatch has produced what can be seen as a vast blind spot in our official statistics — a multibillion-dollar gap between economic reality and what our metrics are able to reveal. Goldman Sachs’ analysis ultimately poses a critical question: as AI continues to reshape industries and redefine the contours of economic growth, can our methods of measurement evolve quickly enough to reflect this transformation with accuracy, or will official statistics continue to lag behind the technological frontier?
Sourse: https://www.businessinsider.com/ai-tech-economy-us-gdp-boost-chips-blindspot-goldman-sachs-2025-9
