Many economists anticipate a significant shift in the global economy due to the rise of generative artificial intelligence (AI). Ege Erdil and Tamay Besiroglu from Epoch, an analysis firm, argued last year that there could be “explosive growth” leading to increased income if AI can effectively replace human labor on a wide scale. Stanford University’s Erik Brynjolfsson also predicts that AI will drive productivity growth in the near future.
To facilitate this economic transformation and integrate AI into production processes, companies will need to invest heavily in new technologies, communication systems, factories, and equipment. Similar investment booms were necessary in the past to support the diffusion of earlier technological innovations like automobiles and personal computers throughout the economy. For instance, between 1992 and 1999, nonresidential investment in the United States increased by 3% of GDP, largely driven by greater spending on computer technology. However, the current level of AI investment, often referred to as capital expenditure (capex), remains relatively low globally.
Following the easing of lockdowns post the COVID-19 crisis, capex saw a gradual increase, rising at an annual rate of approximately 8% earlier in 2022. Some companies were optimistic about technological advancements, while others focused on fortifying their supply chains. However, capex growth slowed down afterwards due to political uncertainties and rising interest rates. By the time OpenAI launched GPT-4 in March 2023, international capex spending was growing at an annualized rate of around 3%.
Currently, some businesses are ramping up their capex once again to seize the perceived opportunities presented by AI. Forecasts suggest that Microsoft’s spending, including research and development, may increase by nearly 20% this year, while Nvidia’s expenditure is expected to rise by over 30%. Mark Zuckerberg, Meta’s CEO, highlighted that AI will be a major focus in terms of architecture and resource allocation.
However, the scenario is more conservative internationally. Excluding leading AI-driven companies like Microsoft and Nvidia, as well as those in the S&P 500, overall budget increases are expected to be modest, around 2.5% in 2024, adjusting for inflation. The broader economic landscape is even bleaker, with Goldman Sachs’ American expenditure tracker indicating a 4% annual decrease in corporate spending.
Despite the hype surrounding the potential of AI, investments in information technology are not experiencing a significant surge. American firms’ investment in “information-handling equipment and software” actually declined by 0.4% year over year in the fourth quarter of 2023.
On a global scale, purchase spending, including institutional investments, is growing at a slower pace compared to pre-pandemic years, as indicated by federal data up to the third quarter of 2023. JPMorgan Chase’s high-frequency measure of global capex also suggests minimal growth. These trends point to a lack of substantial productivity improvements amidst weak investment levels.
While a study in Japan hints at a promising future with strong capex growth, the outlook is less favorable outside the United States. European businesses, for example, are less optimistic, with service sector investment plans in the European Union dropping by more than half compared to early 2022. British businesses also show a modest intent to increase capex by only 3% in the coming year, down from 10% in early 2022.
These developments suggest two possible scenarios. The first is a skeptical view where AI fails to live up to expectations, leading to oversupply of advanced technologies by tech giants with limited demand. The second, more plausible scenario, involves a gradual adoption curve for general-purpose solutions like AI. Just as it took time for businesses to increase spending on software after Microsoft introduced a groundbreaking operating system in 1995, the full impact of AI may unfold gradually over the next few years. Research indicates that while most CEOs anticipate AI’s impact on their companies within the next three to five years, only a small percentage expect significant effects in the short term. The market is poised for transformation, albeit at a measured pace, influenced by the gradual integration of AI technologies.
