Economic overview
Figure 1: Key indicators and trends
Source: Nasdaq Economic Research, FactSet
AI stocks dominate market returns after strong performance
Artificial intelligence (AI) has become a big topic in the market these days. The benefits are so great that some openly wonder if there is a bubble in AI. However, the data shows that over the past few years, stock returns (Figure 2, bars) have been driven primarily by revenue (Figure 2, circles) and not by multiple business expansions.
The Nasdaq-100® is home to the Magnificent 7 stocks as well as Broadcom and AMD, making it the major stock index with the most exposure to AI. It's up more than 50% so far in 2024, with 35 percentage points of that increase coming from revenue as AI drives demand for chips, cloud computing, and AI chatbots for these companies. Other indexes with less exposure to AI, such as the S&P 500, Europe's Stoxx 600, and the S&P 600 Small Cap, have seen revenue growth slow over time, and thus returns.
Therefore, despite concerns about the AI bubble, stock returns have been primarily driven by real profits.
Figure 2: Index price and 12-month forward returns (% change from 2024)
Data through November 7, 2027. Asia Pacific region is compiled by FactSet. Source: FactSet, Nasdaq Economic Survey
AI hyperscaler capex has nearly tripled in the past two years
To meet the demand for AI services, AI hyperscalers are increasing their capital expenditures, especially in data centers and the chips and power needed to run them. The total capital investment of the top five AI hyperscalers is tripled According to JPMorgan, the increase over the past two years has been driven by investments in AI (Figure 3), with the majority of the increase going to AI investments. Future earnings data suggests that this capital investment will continue to increase for at least the next few years.
Importantly, companies investing in AI are already seeing real returns. The company currently spends only about 60% of its operating cash flow on these AI investments, according to the data.
Figure 3: Capital expenditures by leading AI hyperscalers
Source: MUFG
P/E ratios for the largest AI companies are falling, while the rest of the S&P 500 are rising.
This is a big difference from the dot-com bubble, where many companies failed to realize profits. Price-to-earnings (PE) ratios for the top five stocks in the S&P 500 have declined recently as earnings have grown faster than prices (Figure 4). However, the P/E ratios of the remaining 495 stocks in the S&P 500 index have risen to near all-time highs, pushing up the valuation of the S&P 500 index as a whole.
Figure 4: Largest stocks trade at 28x, below 2021 and tech bubble levels
Five major stocks today: NVDA, MSFT, AAPL, GOOGL, AMZN. Source: Goldman Sachs Global Investment Research
AI capital investment is increasingly contributing to U.S. economic growth and is likely to continue to do so
All that additional capital investment is also reflected in the real economy. By some estimates, investments in AI accounted for 92% of US GDP growth in the first half of 2025. Furthermore, spending on data center construction has increased by 400% since 2019. As a result, by mid-2025, data center construction will rival office construction in the United States (Figure 5).
Figure 5: Data center construction spending grows faster than office spending
Expenditures are at the Seasonally Adjusted Annual Rate (SAAR). Source: FactSet, Nasdaq Economic Research
Of course, investing in AI involves more than just building data centers. According to Citi research, investment in AI equipment has increased by 0.9% as a share of GDP since 2023. By comparison, this kind of investment grew by more than 1.25% over six years during the internet revolution, showing that there is potentially even more upside in AI investing.
Whether AI investment is “too much” depends on its impact on productivity
With spending on AI reaching such significant levels, some companies are wondering whether AI will create stranded assets and wasted investments. That's possible as leaders in the AI space emerge, but it's worth considering that AI has the potential to become a transformative infrastructure. AI has the potential to match the scale of other historical investment cycles, such as rail, automobiles, power, and internet, which at their peak accounted for 6% of GDP (Figure 6). These estimates vary, but they show that AI is still a relatively small share of GDP.
Figure 6: Historical investment cycles
Source: Paul Kedrosky, Nasdaq Economic Research
Both facts seem to indicate that AI development may continue for some time yet. The hope for AI investments is that they will deliver productivity gains. For example, in the late 1990s and early 2000s, PCs and the Internet contributed to an average annual productivity increase of about 3%. double Pace from 2007 to 2019 (Figure 7).
Figure 7: Five-year average (Q/Q) productivity of US non-financial companies
Source: MUFG
Enterprises still underutilize AI
As it stands, AI appears to have little impact on the job market. That may be because data shows that corporate adoption of AI remains low. According to Census data, as of late September 2025, only 10% of U.S. companies were using AI to produce goods and services. Among large companies with more than 250 employees, only 13% are using AI. However, Goldman Sachs indicates that 37% of its clients are using AI in regular production, and it expects that number to reach 50% next year.
For now, the weak labor market appears to be due to the Fed's 2022-2024 rate hike cycle and the budget burden that is finally hitting low-income consumers.
5 questions leaders should ask right now
- If our customer base includes low-income consumers facing headwinds from rising inflation and slowing wage growth, how do we serve them?
- How are debt refinancing and interest costs planned as interest rates fall?
- What are the dependencies for chips, power, and critical inputs? And what are the contingencies for rising costs, delays, export and administrative disruptions, etc.?
- With only 10 to 13 percent of companies currently using AI, what is the roadmap for incorporating AI into core workflows, and what measurable productivity goals will executives have to achieve?
- Even if SCOTUS rules against current implementation, tariffs are expected to remain in place, so how will we be affected and how will this change our supply chains?
