While investors fixate on the AI hardware boom, a Goldman Sachs report reveals a larger, trillion-dollar opportunity in the non-hardware investments required to make AI work.
A new Goldman Sachs report argues that the market’s focus on AI hardware overlooks a larger, more complex spending wave, projecting global non-hardware investment could exceed $1 trillion in the coming years. The analysis by economists including Joseph Briggs concludes that for every dollar spent on AI chips and servers, companies are spending two dollars on the intangible capital required to make them useful, such as data infrastructure, software development, and organizational restructuring.
"Every dollar of hardware investment historically brings two dollars of intangible investment," the Goldman Sachs Global Economics Analyst report states, highlighting the scale of often-overlooked spending. This complementary spending is not just a secondary cost but a primary driver of value, determining which companies will successfully translate AI potential into productivity gains.
The report quantifies the scale of this hidden investment, estimating that U.S. companies are already incurring $153 billion in annual labor costs for internal IT teams to develop and manage AI tools. Furthermore, it calculates that executive time dedicated to AI-related organizational change represents another $40 billion in yearly investment, while the total cost for workforce reorganization over the entire AI adoption cycle could reach between $800 billion and $900 billion. The accelerating revenues of data-management firms like Snowflake, Databricks, and Palantir—whose combined enterprise value has surged from under $100 billion in 2022 to a projected $650 billion by 2025—serve as a market-based validation of this trend.
This massive wave of intangible spending suggests the companies best positioned to win the AI race may not be the ones selling the hardware, but those that most effectively invest in the data, software, and organizational overhauls to deploy it. According to the report, this dynamic is creating a "J-curve" effect, where initial productivity gains are masked by heavy upfront intangible investment, potentially understating current GDP growth by as much as 2 percent.
The 'Intangible' Multiplier Effect
The core of Goldman's thesis rests on a historical 1-to-2 ratio between hardware and intangible capital. The bank's analysis of EU KLEMS data found that for every $1 of information and communications technology (ICT) hardware investment, a corresponding $2 is spent on intangible assets. This is broken down into approximately $1.30 for data and software and $0.50 for organizational capital.
Applying this multiplier to the current AI hardware boom, which has already seen $360 billion in U.S. capital expenditure, points to a future wave of non-hardware spending that could top $1 trillion globally. This spending is already visible in the financial results of cloud providers like Amazon AWS, Microsoft, and Google, whose collective cloud service revenues have grown from about $200 billion in 2022 to over $500 billion today.
From J-Curve to Superstar Firms
The report argues that the nature of intangible capital—characterized by high fixed costs and low marginal costs—naturally favors early and effective adopters, creating a winner-take-all market dynamic. Economists refer to this as the rise of "superstar firms." Companies that invest more effectively in data architecture, employee retraining, and process redesign can achieve higher productivity, capture larger revenue shares, and generate better returns on investment.
This investment pattern also creates a "J-curve" for productivity. In the short term, as companies divert resources to internal restructuring and training, measured output can stagnate or even fall, as these activities are often expensed as costs rather than capitalized as investments. However, as these investments mature, they are expected to unlock significant productivity gains. Goldman maintains its forecast that AI will ultimately boost annual labor productivity growth by 1.5 percentage points and lift global GDP by 15% after the technology is fully adopted. The companies making the smartest non-hardware investments today are the most likely candidates to become the next generation of superstar firms with超额估值.
This article is for informational purposes only and does not constitute investment advice.
