Not the Federal Reserve and not tariffs, AI will become the most important macro variable in 2026

Not the Federal Reserve, nor tariffs, Barclays has recently made it clear that artificial intelligence will become the most important macro factor in 2026, with its influence surpassing traditional monetary policy and trade policy drivers.

On November 21, according to news from the Wind Trading Desk, Barclays stated in its latest macro outlook report that as traditional macro drivers gradually exit the stage, AI has "no turning back" and will continue to drive U.S. economic growth and financial market performance.

According to Barclays analysts Ajay Rajadhyaksha and Amrut Nashikkar, AI not only directly contributed nearly 1% to U.S. GDP growth but also supported consumption through rising valuations of AI-related stocks and wealth effects, becoming a dual engine for the U.S. economy, offsetting the negative impacts of weak job growth and tariff shocks.

At the same time, the importance of traditional macro drivers will decline significantly. Most major central banks are expected to maintain a wait-and-see attitude for most of 2026, the European Central Bank has ended its rate hike cycle, the Bank of England is nearing the end, and the Bank of Japan has only one or two rate hikes left. The Federal Reserve is expected to cut rates three more times, but the appointment of a new chair is unlikely to raise investor concerns about the independence of the Federal Reserve.

The report emphasizes that although the AI narrative may face challenges such as valuation, financing needs, and power constraints, these concerns are overly exaggerated in the coming quarters. The bank expects the U.S. to grow by 2.1% in 2026, with global economic expansion continuing for the fourth year. However, if the AI narrative encounters significant setbacks in 2026, it will be the biggest risk facing the U.S. and even the global economy.

AI has become the dual engine of U.S. economic growth

The Barclays report shows that the largest tech companies in the U.S. will invest about $400 billion this year in data centers, model training, and cloud infrastructure. Even more astonishing, McKinsey previously estimated that by 2030, AI-related capital expenditures could reach an astonishing $5-7 trillion.

The bank emphasizes that these figures are not castles in the air. Meta CEO Mark Zuckerberg stated in September 2025 that "spending hundreds of billions of dollars incorrectly" would be "very unfortunate," but he insisted that the greater risk is that Meta does not invest enough or fast enough in AI. Google's Larry Page was quoted as saying he "would rather go bankrupt than lose this race."

At the same time, the report points out that OpenAI has signed agreements worth hundreds of billions of dollars with U.S. tech giants, including NVIDIA, AMD, Microsoft, and Oracle. The company announced plans to invest $500 billion in its Stargate project over the next few years to build AI data centers across the U.S. Despite OpenAI currently generating less than $20 billion in revenue and expected to lose hundreds of billions annually in the coming years, it possesses ChatGPT, the globally dominant AI platform.

Barclays estimates that AI spending by hyperscale cloud service providers will grow by 70% in 2025, with plans for another 30% increase next year. **In 2025, it is estimated that nearly 1% of U.S. GDP growth will come from AI spending, providing important cushioning for the U.S. economy amid the trade war **

In addition to the highly watched capital expenditures, the supporting role of AI in American consumption is little known. Barclays pointed out that as of April, the U.S. household savings rate was extremely low. In comparison, during 2019, when the unemployment rate was lower, inflation was below 2%, and the Federal Reserve was easing policies, the average savings rate was 6.5-7%.

The report stated that after the tariff shock, the savings rate should have risen significantly, which would have had a major impact on consumption. However, consumption did not decline significantly, averaging a growth of 1.5% in the first half of the year, and may strengthen further in the third quarter. The wealth effect played an important role, as the U.S. stock market quickly recovered from the shock in April and continued to rise, with the Nasdaq index expected to close the year with a 20% increase.

According to the rule of thumb, 2.5-3% of the wealth created in the past 12-18 months will be converted into consumption in the next 12 months. Since the recovery from the COVID-19 pandemic, U.S. household wealth has increased by more than $55 trillion, with AI-focused stocks contributing the vast majority of the stock market gains.

Traditional macro drivers will take a back seat

Barclays stated that by 2026, the importance of tariffs and monetary policy will decline significantly.

From a monetary policy perspective, the bank believes that most major central banks will remain on hold for most of next year: the European Central Bank has ended its actions, the Bank of England is nearing the end, and the Bank of Japan has only one or two rate hikes left. The Federal Reserve may also cut rates three times.

Barclays does not believe that the appointment of a new chair will raise investor concerns about the independence of the Federal Reserve, as the government’s confirmed candidates (including Hassett, Waller, Rieder, etc.) are likely to be accepted by the financial markets.

Most importantly, Barclays believes that the monetary policy in 2026 will not be primarily politically motivated, regardless of the economic backdrop.

Regarding tariffs, although some expect a favorable ruling from the Supreme Court on tariffs related to the International Emergency Economic Powers Act (IEEPA), Barclays believes this hope is overestimated— even if the Supreme Court rules IEEPA unconstitutional, the U.S. will achieve the same level through industry tariffs.

More importantly, the market has already turned the page on tariffs, with diminishing reactions to each news headline, and the current economic entities are adjusting well.

However, the bank also cautioned in the report that whether the labor market can hold up is key. The report noted that employment growth has been very weak in recent quarters, with a three-month moving average of only 29,000.

Barclays believes the U.S. labor market is facing serious supply-side shocks. Between 2022 and 2024, immigrant labor will add 2 to 2.5 million workers to the job market each year. With the sudden closure of new immigration, new supply has disappeared.

Layoff announcements from large tech companies have raised concerns about AI causing unemployment. In 2025, Amazon announced layoffs of 30,000, Microsoft 15,000, and Oracle 10,000. However, Barclays is skeptical that AI will cause massive unemployment in this business cycle, as the labor force participation rate for the core age group remains high If labor demand collapses, the labor participation rate is expected to decline. The U.S. economy has long been in a "no hiring, no layoffs" state, and Barclays believes this state will continue.

AI Faces Challenges but Remains Optimistic

The report states that while valuation, financing needs, and power constraints are concerning issues, these worries are overly exaggerated in the coming quarters.

Regarding power demand, Barclays estimates that $3 trillion to $4 trillion in AI spending corresponds to a power demand of 70-95 gigawatts. The U.S. Department of Energy believes that data centers currently account for 4.4% of total power demand, but this is expected to grow exponentially in the coming years.

However, Barclays believes that power will not be a serious constraint on capital expenditures at least in 2026 and 2027.

In terms of financing, publicly listed mega-cap companies have strong balance sheets and generate substantial cash flow. However, some estimates suggest that free cash flow may only cover about half of AI capital expenditures in the coming years, with the remainder needing to come from capital markets.

Nonetheless, these concerns must be balanced against actual computing power and reasoning demands.

• Google recently pointed out that the supply-demand environment is expected to remain tight in 2026, with strong demand from cloud customers.

• Microsoft stated that computing power demand exceeds the supply of all workloads, and capacity is expected to remain constrained at least until the end of 2026, with hopes to double data center capacity within two years.

• Meta noted that its computing power demand continues to expand "significantly," and Amazon expressed a similar viewpoint.

Barclays believes that the likelihood of a sudden demand-supply mismatch (similar to the slowdown in demand during the internet bubble burst while a large amount of new supply comes online) is extremely low at least in the coming quarters.

AI has already brought efficiency improvements and revenue growth in the tech sector, and the narrative around AI will continue to influence capital expenditures and valuations. This is a truly game-changing technology that will reshape the global economy in the coming years