Hidden Dangers in the AI Boom: The Asset-Light Model of the 'Big Seven' is Being Disrupted, and the Return on Massive Investments Faces Uncertainty

Financial Services Association, August 4 (Editor: Xiaoxiang) In the past two weeks, US Technology giants after another announced strong second-quarter results, thanks to the industry's widespread embrace of artificial intelligence.

But on closer inspection, some investors may also find the unsettling side of the AI boom: all spending on chips, data centers, and other AI infrastructure is draining American companies' Cash/Money Market reserves.

This scene may be highlighting the hidden risks behind the AI boom — no one doubts the potential of artificial intelligence to enhance growth and productivity over the long term, but the huge financing that supports this boom may put pressure on businesses and capital markets.

The AI boom is changing the business models of giants

According to the data, since the first quarter of 2023, the scale of investment in information processing equipment in the US has increased by 23% after adjusting for inflation, while the total US gross domestic product (GDP) increased by only 6% during the same period. In the first half of this year, investments in information processing contributed more than half of America's overall GDP growth rate of 1.2%.

These figures unquestionably reflect a phenomenon: at a time when US consumer spending is stagnating, AI spending is gradually becoming the “backbone” driving US economic growth.

A large portion of this investment consists of graphics processing units (GPUs), memory chips, Server, and networking equipment to train and running the Big Language Model (LLM) at the core of this AI craze — all of this computing power requires significant building, land, and electricity support.

And such huge capital investment is actually changing the business model of American Technology giants for a long time...

One of the main reasons investors have loved these companies over the years is because of their “asset-light” — Technology giants make profits through intangible assets such as intellectual property, software, and digital platforms with a “network effect.” Users flock to Facebook, Google Search, iPhone, and Windows because other users are also generally using it. This means that Technology giants will increase their revenue and hardly need to invest more in construction and equipment costs — making them Cash/Money Market machines.

This trend can be fully reflected through the “free cash flow” indicator — this indicator is roughly defined as Cash/Money Market flow from Operation activities minus capital expenses, and excludes items such as non-cash impairment charges that may distort Net income.

This is probably the purest measure of a company's internal ability to generate Cash/Money Market. For example,$Amazon (AMZN.US)$“Our financial focus is on the long-term sustainable growth of free cash flow,” he said to investors.

From 2016 to 2023, the US Technology giants clearly did this -- $Alphabet-C (GOOG.US)$ , Amazon, $Meta Platforms (META.US)$ with$Microsoft (MSFT.US)$Free cash flow and Net income grew roughly at the same time during this period.

Since 2023, however, the two have begun to clearly diverge. During this period, the four companies' consolidated Net income increased 73% from two years ago to $91 billion, while free cash flow fell 30% to $40 billion, according to FactSet. Although Apple is relatively conservative in terms of capital expenditure, its free cash flow is also beginning to lag behind Net income growth.

Prospects for return on huge investments face question marks

While the economic potential of artificial intelligence is clear, the financial return is clearly still a question mark. Currently, although the two leading independent developers of large language models — OpenAI and Anthropic — are growing rapidly in business, they are still experiencing losses to date.

In fact, the latest profits of US Technology giants still largely reflect their mature core Business — such as Meta and Alphabet advertising revenue, Apple's iPhone revenue, etc. As to when the huge investment in artificial intelligence hardware will pay off, even the executives of these companies are advising investors to be patient.

Facebook's parent company Meta reported a 36% increase in second-quarter earnings last week, but free cash flow fell 22%. The company said capital expenditure in 2025 will be about double that of last year and is expected to maintain a “similar amount” of growth in 2026.

Susan Li, chief financial officer of Meta, told investors that “we are still in the early stages of the investment life cycle” with regard to investments in the big language model, and “we do not expect to generate significant revenue from any of these investments in the short term.”

Amazon, on the other hand, began reducing the construction scale of its fulfillment center (automated logistics facility) in 2022, so that its free cash flow can be corrected. But over the past year, it has also increased its investment in AWS, its cloud computing service unit — hosting data and running AI models for customers, which reduced its free cash flow by two-thirds compared to the same period last year.

Currently, investors are still betting on the pricing of large technology companies on their asset-heavy business model, which will make the same profit as the previous asset-light model. But Jason Thomas, head of research at The Carlyle Group Group, said, “So far, we have no evidence to prove this.”

He added, “The variable people overlook is the time frame. All of this capital expenditure may eventually become extremely efficient, but the time it will pay off may exceed the return on investment period that Shareholder are concerned about.”

The “echo” of the internet bubble

This situation is also easily reminiscent of the internet bubble at the turn of the century. In the late 1990s and early 2000s, the nascent Internet boom prompted investors to invest heavily in network startups and broadband telecom operators.

At the end of the day, their predictions that the internet would drive productivity growth were right, but they got it wrong in terms of financial return — many companies went out of business because the money they made wasn't enough to cover expenses. The resulting decline in capital expenditure was also one of the reasons for the mild recession in the US economy in 2001.

Today, a collapse similar to an internet bubble seems hard to imagine. Huge investments in the field of artificial intelligence come from mature companies that can achieve stable profits, and their demand for computing power exceeds supply.

However, if their revenue and profit expectations are too optimistic, it will be difficult to maintain the current pace of capital expenditure.

The challenge of high interest rates may persist for a long time

Finally, another problem that cannot be ignored comes from interest rates — after the 2007-09 global financial crisis, large technology companies were both beneficiaries of low interest rates and drivers of low interest rates.

Thomas of The Carlyle Group Group estimates that between the financial crisis and the COVID-19 pandemic, these companies generated five to eight times the amount of Cash/Money Market they invested, and that excess Cash/Money Market would flow back into the financial system — a high US federal deficit at the time, combined with lower inflation than the Federal Reserve's 2% target and operations such as the Federal Reserve's QE, these factors combined to depress long-term interest rates. Low interest rates, in turn, make investors value these companies' future profits more.

But now, the US government deficit has grown larger, the US inflation rate continues to be above 2%, and the Federal Reserve has been reducing its Bonds holdings. At the same time, giant companies are facing huge investment needs in order to chase the AI boom and return production to avoid taxes.

According to Thomas's estimates, since 2020, the cumulative free cash flow of these companies in relation to gross domestic product (GDP) has been 78% lower than in the same period after 2009.

All of this suggests that interest rates in the US are likely to be much higher than they were before the outbreak of the pandemic in the next few years — another risk both the US economy and these companies will face, and investors may not be fully aware of this.

Still want to know more about the market?Futubull AI is online!Accurate answers, comprehensive insight, and seize key opportunities!

Edit/rice