Goldman Sachs questions the real impact of AI on economic growth

While tech giants like Meta, Amazon, Google, and OpenAI have made colossal investments in artificial intelligence, reaching tens of billions of dollars in 2025 and planning nearly 700 billion more in 2026 for new data centers, hopes for a massive economic transformation through AI appear tempered. Goldman Sachs, a major player in financial analysis, published a cautiously toned report, citing a limited economic impact despite the investment frenzy. This questioning raises several questions: would artificial intelligence technology really stimulate economic growth? Or are we witnessing an illusion in the face of promises ineffective on a productive level? Between expectations of spectacular profits on the financial market and a moderate reality of economic fallout, Goldman Sachs’ analysis invites a reconsideration of certainties about the AI revolution. This observation takes place in a context where massive investment in hardware infrastructure seems to benefit foreign semiconductor manufacturers investigated in Taiwan or Korea more than the American economy itself, thus blurring the direct link between spending and domestic growth. At the same time, many user companies still struggle to observe concrete productivity gains, even though macroeconomic indicators remain unclear on the real impact. Finally, this questioning also has political and strategic repercussions, between calls for measured regulation and the need for a coherent industrial strategy. This complex and nuanced overview of artificial intelligence leads to profound reflection on its real role in the current and future economic dynamic.

Goldman Sachs and the critical analysis of the economic impact of artificial intelligence

Faced with the rise of artificial intelligence, Goldman Sachs has taken an analytical position concerned with precisely examining the consequences on economic growth. Although AI investments have reached historic records – notably in computer hardware and electronic chips – the institute observes that these expenses do not mechanically translate into a substantial effect on the U.S. gross domestic product (GDP). One major reason for this gap is the very nature of the economic flows involved. Indeed, a significant proportion of these investments benefits manufacturers from Asia, particularly Taiwan and South Korea, specializing in the production of semiconductors and advanced equipment. This import of hardware thus reduces the direct circulation of added value on American soil.

Joseph Briggs, analyst at Goldman Sachs, notably points out that this intuitively attractive interpretation – that AI would be an immediate growth driver – could mask more complex dynamics. Economic data show that, in fact, the effects of technology on economic activity remain weak, even nonexistent in some cases. Jan Hatzius, chief economist of Goldman Sachs, confirms this idea by stating that investment in AI has had “practically no” influence on GDP growth.

This analysis thus invites a rigorous re-examination of statistics and a questioning of overly optimistic forecasts often reported by the financial markets. The excitement around AI, symbolized by the record market capitalization of the S&P 500 estimated at more than 670 billion dollars for tech companies linked to artificial intelligence, must be tempered by a cold and measured look. For technology, as promising as it may be, does not automatically generate spontaneous economic growth.

Massive investments in AI: a promise difficult to convert into real growth

Since 2025, the world’s leading technology companies have multiplied projects and expenditures to integrate artificial intelligence into their business model. Dedicated infrastructures, such as data centers, are expanding at an unprecedented pace, supported by a colossal investment of nearly 700 billion dollars in 2026. These heavy facilities are necessary to train and operate advanced AI models that underpin current innovations in voice recognition, natural language processing, robotics, and more.

However, despite this volume of investment, economic returns are still awaited. The phenomenon can be explained by several interdependent factors:

• Nature of algorithms: although sophisticated, existing AI models often require adjustments and continuous learning, which limits their immediate effectiveness on productivity-generating processes.
• Adoption costs: companies must reorganize their operations, train their employees, and rethink their value chains to fully leverage AI tools, a long and costly process.
• Technological dependency: a large part of the hardware is imported, which disperses economic gains and prevents full integration into the national industrial fabric.
• Measurement of effects: the absence of reliable and standardized indicators makes the precise evaluation of AI’s impact on productivity and growth complex.

These blockages partly explain why the private sector and institutions do not yet observe tangible results commensurate with financial incentives. Therefore, American economic growth has not experienced sustained dynamism due to this single technology, contrary to what financial market speculation might suggest.

A concrete example comes from the Federal Reserve Bank of St. Louis which, despite initial optimistic studies attributing 39% of growth in the third quarter of 2025 to AI, has significantly tempered this figure in view of analysis difficulties and sector disparities. This caution aligns with that of Goldman Sachs and recalls the complexity of transforming technological innovation into a massive growth lever.

Table: comparison of AI investments and impact on economic growth

Year	Global investments (in billion $)	Estimated contribution to economic growth (%)	Main economic beneficiaries
2024	650	1.5	Asian semiconductor manufacturers
2025	900	2.2	American tech giants (hardware investment)
2026 (forecast)	1,200	2.5	Predominantly Asian manufacturers

Limits in measuring the economic impact of artificial intelligence according to Goldman Sachs

A major obstacle pointed out by Goldman Sachs in its analysis is the difficulty in effectively quantifying the economic fallout of AI. Data volatility, combined with the absence of standard methodologies, produces sometimes contradictory results. Encouraged by enthusiasm, companies tend to overestimate the immediate effects of their AI investments, while economists prefer to temper their interpretation of the figures.

This uncertainty was also recently highlighted by Mary Daly, president of the Federal Reserve Bank of San Francisco, who insists on the need for cautious observation. Even if technology raises significant expectations, it remains difficult to measure on a productive level. Prudence therefore invites decision-makers to avoid hasty conclusions and to analyze data in depth before modifying economic policies.

This finding is confirmed by a survey from the National Bureau of Economic Research (NBER) involving nearly 6,000 business leaders in North America, Europe, and Australia. This survey reveals that, despite the active adoption of artificial intelligence by 70% of companies surveyed, about 80% of them observe no significant change either in employment or productivity.

This paradox clearly illustrates the concrete difficulties in capitalizing on AI investments over a short or medium time horizon, thus corroborating the skepticism expressed by Goldman Sachs. To establish a reliable link between technological innovation and economic growth, measurement tools and analysis criteria will likely need to be reviewed.

Strategic and industrial stakes behind investments in artificial intelligence

Beyond the simple economic question, investments in artificial intelligence are part of a critical geopolitical and industrial dynamic. For the United States, heavy dependence on imported equipment weakens national technological sovereignty. This raises important questions about the ability to lead a truly local “technological renaissance.”

This situation encourages public policies to promote reshoring and support domestic manufacturing of essential components, notably semiconductors. Global competition in the advanced technology sector, notably between the United States and China, imposes a strategic imperative that goes beyond mere economic analysis.

In this context, Goldman Sachs points out that without greater mastery of the technological value chain, high AI spending risks reinforcing industrial dependency, to the detriment of real growth dynamics on American soil. The fragmentation of economic benefits by geographic regions thus complicates the implementation of a coherent industrial strategy.

An illustrative example is the rise of Taiwanese chip manufacturers, such as TSMC, who dominate the supply of key components for AI. Their economic success benefits American growth only minimally, even though American tech giants remain the main drivers of innovation.

Artificial intelligence and the real transformation of companies: the implementation challenge

Having the best artificial intelligence technology does not guarantee immediate economic success. Successful AI adoption requires a profound change in organizations, a redesign of business processes, and appropriate training of employees. This work is often complex and lengthy, slowing down the announced growth potential.

A large number of companies thus face a gap between massive injection of AI technologies and real, measurable benefits. Installing artificial intelligence software in a department does not automatically transform work methods. It is often necessary to rethink corporate strategy to align tools with a clear vision of efficiency and profitability.

Observed results in certain sectors show a diversity of effects: some organizations indeed experience a positive return on investment, especially in automating repetitive tasks or improving customer relations. Others struggle to generate a clear advantage, due to a lack of coherent integration or full appropriation of the technology.

• Appropriate training and skills development of teams to fully use AI
• Clear identification of key processes that can be optimized by AI
• Rigorous monitoring of performance indicators after technological deployment
• Organizational adaptation to accompany cultural change
• Balance between technological innovation and budget control

In sum, to transform AI investments into real growth engines, it is essential to build an agile, evolving strategy centered on concrete economic value.

Implications for the financial market: realigning expectations in the face of AI technology

The frenzy around artificial intelligence has had a considerable echo on financial markets. Many investors have embraced the idea that AI would quickly generate profits and disrupt sector dynamics. Thus, the market capitalization of AI-related companies has reached dizzying heights, fueling a speculative wave.

Goldman Sachs nonetheless draws attention to the need to realign these expectations with tangible economic reality. Productivity gains, on which stock market valuations rely, are slow to appear. This gap may trigger a market correction or even a refocusing of technological projects toward more concrete and profitable applications.

This evolution requires investors to adopt a more cautious approach, favoring rigorous evaluations of performance and actual return on investment, rather than ambitious projections that remain uncertain. Companies themselves, in turn, will need to demonstrate with greater transparency their results and their ability to turn innovation into sustainable growth.

Regulatory framework and political perspectives on artificial intelligence

The argument advanced by certain political actors, notably former President Donald Trump, who claims that investment in AI energizes the American economy and requires relaxed regulation to stimulate innovation, is challenged by recent analyses such as Goldman Sachs’. If AI does not generate massive economic growth in the short term, the discourse justifying less regulatory intervention loses strength.

This situation opens the way to a more balanced reflection on legislative frameworks that could better regulate the development of artificial intelligence without slowing technological advances. Indeed, regulators can now afford to adopt a firmer stance, since AI does not constitute an uncontested economic locomotive likely to evade all control.

Moreover, this dynamic could favor regulation aimed at protecting public interests, securing data, ensuring fairness, and preventing abuses without fearing to slow economic growth. This evolution contributes to a more mature and balanced political dialogue, essential for the proper integration of AI into society.

In this context, debates focus on:

1. Establishing coherent federal standards to avoid excessive regulatory fragmentation
2. Protecting users’ rights and data privacy
3. Supporting initiatives promoting reshoring and technological sovereignty
4. Preserving economic competitiveness by ensuring an environment conducive to responsible innovation

This more nuanced approach corresponds to the principles of economic and political prudence recommended by Goldman Sachs and other major financial institutions.