Cathie Wood, CEO of ARK Invest, recently released her investment outlook for 2026, likening the U.S. economy to a “coiled spring” poised for a strong rebound in the coming years. Although the U.S. real Gross Domestic Product (GDP) has grown continuously over the past three years, the economy has experienced a “rolling recession” internally, with significant pressure on some sectors. For instance, home sales dropped by 40% from January 2021 to October 2023, hitting the lowest level since 2010; meanwhile, the manufacturing sector has been contracting for three consecutive years according to the Purchasing Managers’ Index (PMI).
Wood pointed out that factors such as deregulation, tax cuts, declining inflation, and lower interest rates have collectively created conditions for an economic rebound. She particularly emphasized that, as manufacturing facilities, equipment, software, and domestic R&D accelerate depreciation, corporate income tax refunds will increase, potentially bringing the effective corporate income tax rate close to 10%. In response to the supply shocks triggered by the COVID-19 pandemic, the Federal Reserve significantly raised interest rates between March 2022 and July 2023, rapidly increasing rates from 0.25% to 5.5%. This move exerted pressure on real estate, manufacturing, capital expenditures in non-AI sectors, and low- to middle-income groups. Now, with the shift in the policy environment, the foundation for economic recovery is being laid.
The AI boom is driving capital expenditures to their highest level since the late 1990s. In 2025, investment in data center systems grew by 47% to nearly $500 billion, and is expected to increase by another 20% in 2026. Investment on such a scale raises questions about the sources of returns. In addition to the semiconductor industry and major cloud service companies (such as relevant entities in the public stock market), investment growth is also benefiting native AI companies that are not yet publicly listed. It is reported that, as of the end of 2025, OpenAI and Anthropic achieved annualized revenues of $20 billion and $9 billion, respectively, showing rapid growth. Rumors suggest that both companies are considering initial public offerings (IPOs) within the next one to two years.
In the corporate sector, the adoption of many AI applications remains slow, constrained by factors such as bureaucracy, inertia, and the reconstruction of data infrastructure. In 2026, companies may realize that they must train models based on their own data and iterate quickly, or risk being overtaken by competitors. AI-driven applications are expected to provide higher-quality customer service, accelerate product launches, and help startups achieve more with fewer resources.
AI and other innovative platforms have significant deflationary effects. According to some benchmark data, AI training costs are decreasing by 75% annually, while inference costs are declining by up to 99% per year. The rapid decline in technology costs will drive explosive growth in its applications. Against the backdrop of productivity growth of 5%-7%, labor force growth of approximately 1%, and inflation remaining in a low range, the U.S. nominal GDP growth rate is expected to remain within the 6%-8% range in the coming years. This technology-driven deflationary environment, potentially similar to historical periods of major technological revolutions, could provide enduring momentum for economic growth.
