To the investors and all supporters of ARK Invest (ARK):

Happy New Year! Thank you all very much for your support. As I outline in this article, we firmly believe that investors have every reason to remain optimistic! We hope you enjoy our discussion.

From an economic history perspective, we are at a pivotal moment.

A compressed spring ready to rebound

Despite three consecutive years of growth in the United States' real Gross Domestic Product (GDP), the core of the U.S. economy has undergone a 'rolling recession,' which has now transformed into a coiled spring poised for a robust rebound in the coming years.

To address the supply shock triggered by the COVID-19 pandemic, the Federal Reserve Fund rate surged from 0.25% to 5.5% within 16 months between March 2022 and July 2023, marking a historic 22-fold increase. This measure pushed sectors such as real estate, manufacturing, capital expenditures outside the artificial intelligence domain, and middle- and low-income groups in the U.S. into a state of recession, with specific data illustrated in the chart below.

Measured by existing home sales, the housing market plummeted 40% from an annualized rate of 5.9 million units in January 2021 to 3.5 million units in October 2023, a level last seen in November 2010, with persistent fluctuations over the past two years. More indicative of the 'degree of compression' is the fact that current existing home sales have dropped to levels seen in the early 1980s, when the U.S. population was approximately 35% smaller than it is today.

As measured by the U.S. Purchasing Managers' Index (PMI), the manufacturing sector has been in contraction for nearly three years. This diffusion index uses 50 as the dividing line between expansion and contraction, as shown in the chart below.

Meanwhile, capital expenditure, measured by non-defense capital goods excluding aircraft, peaked in mid-2022 and has now rebounded to that peak level, whether or not the technology sector is included. In fact, this capital expenditure indicator struggled to break through a bottleneck for more than 20 years following the bursting of the tech and telecom bubble in the 1990s, until supply shocks triggered by the COVID-19 pandemic in 2021 forced companies to ramp up investment in both digital and physical assets.

Today, with the maturation of artificial intelligence, robotics, energy storage, blockchain technologies, and multi-omics sequencing platforms, what was once considered an upper limit of spending now appears to be the new baseline. After experiencing two decades of stagnation around the 70 billion USD mark post the tech and telecom bubble of the 1990s, this indicator is now poised to enter the strongest capital expenditure cycle in history, as illustrated in the chart below.

In our view, the emergence of an artificial intelligence bubble is still several years away!

Moreover, according to data from the University of Michigan, consumer confidence among middle- and low-income groups in the United States has dropped to its lowest level since the early 1980s—when double-digit inflation and interest rates severely eroded purchasing power, pushing the U.S. economy into successive recessions. Notably, confidence among high-income groups has also declined in recent months, as shown in the chart below. In our view, consumer confidence is one of the most compressed “springs,” and it is now primed for a strong rebound.

Deregulation, combined with tax cuts, falling inflation, and declining interest rates.

Thanks to the combined benefits of deregulation, tax cuts (including tariff reductions), receding inflation, and falling interest rates, the United States' 'rolling recession' over the past few years is poised for a rapid and significant reversal in the coming year and beyond.

Deregulation is accelerating innovation across industries, with David Sacks, the first Artificial Intelligence and Cryptocurrency Czar, playing a leading role in the fields of artificial intelligence and digital assets.

Meanwhile, tax cuts targeting tips, overtime pay, and social security will deliver substantial tax refunds to American consumers this quarter, potentially boosting annualized real disposable income growth from about 2% in the second half of 2025 to approximately 8.3% this quarter.

Moreover, corporate tax rebates are set to surge significantly – accelerated depreciation policies for manufacturing facilities, equipment, software, and domestic R&D expenditures will reduce the effective corporate tax rate to around 10% (as illustrated in the chart below), placing it at a globally competitive low. For instance, any enterprise commencing construction of manufacturing facilities in the U.S. before the end of 2028 can enjoy full depreciation in the first year of operation, instead of the previous 30 to 40-year depreciation schedule; equipment, software, and domestic R&D expenditures are also eligible for 100% first-year depreciation. This cash flow incentive was made permanent through last year's budget bill and retroactively effective as of January 1, 2025.

Over the past few years, inflation measured by the Consumer Price Index (CPI) has stubbornly lingered in the 2%-3% range but is expected to drop to unexpectedly low levels (potentially even negative inflation) in the coming years, for reasons detailed in the charts below. First, West Texas Intermediate crude oil prices have fallen by approximately 53% from their post-COVID-19 high of about $124 per barrel on March 8, 2022, with the current year-over-year decline being roughly 22%.

Since peaking in October 2022, the sales prices of newly constructed single-family homes have fallen by approximately 15%; meanwhile, the inflation rate for existing home sales (based on a three-month moving average) has decreased from its post-COVID peak in June 2021 (approximately 24% year-over-year) to about 1.3%, as shown in the chart below.

In the fourth quarter of 2025, to clear an inventory of nearly 500,000 newly constructed single-family homes (the highest level since October 2007, just before the global financial crisis, as shown in the chart below), the three major homebuilders significantly reduced prices:$Lennar Corp (LEN.US)$A 10% price reduction, $KB Home (KBH.US)$ a 7% price reduction,$D.R. Horton (DHI.US)$and a 3% price reduction. These price cuts will gradually be reflected in CPI data over the coming years.

Finally, as one of the most powerful forces against inflation, non-farm productivity grew against the trend during the 'rolling recession,' rising by 1.9% year-on-year in the third quarter of 2025. Compared to the 3.2% increase in hourly compensation, the rise in productivity reduced unit labor cost inflation to 1.2%, as shown in the figure below. This data indicates that there is currently no cost-push inflation similar to that of the 1970s!

Further evidence of easing inflation is that the inflation rate measured by Truflation has recently dropped to 1.7% year-on-year, nearly 100 basis points lower than the CPI inflation rate calculated by the Bureau of Labor Statistics (BLS).

Productivity Boom

In fact, if our research conclusions on technology-driven disruptive innovation hold true, over the next few years, influenced by both cyclical and long-term factors, the year-on-year growth rate of non-farm productivity could rise to 4%-6%, further driving down unit labor cost inflation.

The integration of current major innovation platforms—artificial intelligence, robotics, energy storage, public blockchain technologies, and multi-omics—not only has the potential to push productivity growth to sustainable new highs but will also create enormous wealth.

Increased productivity may also alleviate significant geo-economic imbalances in the global economy. Companies can direct productivity gains into one or more of four strategic directions: expanding profit margins, increasing R&D and other investments, raising wages, or lowering product prices. In China, providing higher wages for more productive workers and/or improving corporate profit margins can help shift the economy from over-reliance on investment toward a more balanced structure. Since joining the World Trade Organization in 2001, China's investment-to-GDP ratio has averaged as high as 40%, nearly double that of the United States (as shown in the figure below). Raising wages would steer China’s economy toward consumption-driven growth, moving away from excessive reliance on commoditization, aligning with the goal of 'anti-involution.' Meanwhile, U.S. companies could increase investment and/or reduce prices, further enhancing their competitiveness relative to Chinese enterprises.

However, in the short term, technology-driven productivity gains may continue to slow U.S. employment growth, pushing the unemployment rate up from 4.4% to above 5.0%, and prompting the Federal Reserve to continue cutting interest rates. Subsequently, deregulation and other fiscal stimulus measures will amplify the effects of these rate cuts, driving an acceleration in GDP growth in the second half of 2026. Meanwhile, thanks to falling oil prices, housing prices, and tariffs, along with various technologies that boost productivity and reduce unit labor costs, inflation may decline further.

Notably, according to some benchmark data, the cost of artificial intelligence training is declining by 75% annually, while the cost of AI inference is dropping by up to 99% per year. The unprecedented decline in the costs of various technologies will lead to explosive unit growth. Therefore, against the backdrop of productivity growth of 5%-7%, labor force growth of 1%, and inflation ranging from -2% to +1%, it would not be surprising if nominal GDP growth in the U.S. were to remain in the 6%-8% range over the next few years.

The deflationary effects of artificial intelligence and four other major innovation platforms will gradually accumulate, creating an economic environment akin to the period of significant technological revolutions driven by the internal combustion engine, electricity, and the telephone during the 50 years leading up to 1929. During that time, short-term interest rates were roughly in line with nominal GDP growth, while long-term interest rates reacted to the underlying deflationary currents brought about by these technological booms, with the yield curve averaging an inversion of 100 basis points, as shown in the chart below.

Other New Year Reflections

Rising Gold Prices and Falling Bitcoin Prices

In 2025,$XAU/USD (XAUUSD.CFD)$Up 65%, while$Bitcoin (BTC.CC)$Prices fell by 6%. Although many observers attribute the surge in gold prices from $1,600 to $4,300 (a 166% increase) since the end of the U.S. stock market bear market in October 2022 to inflation risks, another interpretation is that the pace of global wealth creation (represented by a 93% rise in the MSCI World Index) has outpaced the annualized growth rate of approximately 1.8% in global gold supply.

In other words, incremental demand for gold may have exceeded its supply growth. Interestingly, during the same period, Bitcoin prices surged by 360% despite an annualized supply growth of approximately 1.3%. One notable aspect of this comparison is that gold miners and Bitcoin miners may react quite differently to price signals: gold miners can increase production in response to rising gold prices, whereas this is not possible for Bitcoin. According to mathematical algorithms, Bitcoin's annualized supply growth over the next two years will be approximately 0.82%, after which it will drop to around 0.41%.

A Historical Perspective on Gold Prices

Measured by the ratio of market value to M2, over the past 125 years, gold prices have only reached higher levels during one period—namely, the Great Depression of the 1930s, when gold was fixed at $20.67 per ounce while M2 plummeted by approximately 30%, as illustrated in the chart below. Recently, the ratio of gold to M2 has surpassed its previous peak (reached in 1980 when inflation and interest rates soared into double digits). In other words, from a historical perspective, current gold prices are at an extreme level.

Moreover, as can be seen from the chart below, a long-term decline in this ratio tends to be accompanied by robust stock market returns. According to research by Ibbotson and Sinquefield, since 1926, the compound annual return of stocks has been approximately 10%. Following two major long-term peaks of this ratio in 1934 and 1980, the Dow Jones Industrial Average (DJIA) delivered returns of 670% and 1015% over the subsequent 35-year period ending in 1969 and the 21-year period ending in 2001, with annualized returns of 6% and 12%, respectively. Notably, the average annualized returns for small-cap stocks were 12% and 13%, respectively.

For asset allocators, another key consideration is Bitcoin’s relatively low correlation with gold returns and with returns of other major asset classes since 2020, as shown in the table below. Notably, Bitcoin's correlation with gold is even lower than$S&P 500 Index (.SPX.US)$its correlation with bonds. In other words, for asset allocators seeking higher risk-adjusted returns in the coming years, Bitcoin should serve as an excellent diversification option.

Outlook for the US Dollar

In the past few years, a prevailing view has been that the era of "American exceptionalism" is over. One piece of evidence cited is the largest decline in the US dollar during the first half of the year since 1973, as well as its largest annual drop since 2017.

In 2025, measured by the trade-weighted US Dollar Index, the US dollar fell 11% in the first half of the year and 9% for the full year. If our outlook on fiscal policy, monetary policy, deregulation, and US-led technological breakthroughs holds true, the United States’Return on Investmentcompetitiveness will relatively improve compared to other regions globally, thereby driving the strengthening of the US dollar. The policies of the Trump administration bear a striking resemblance to those during the early period of Reaganomics in the 1980s — when the value of the US dollar nearly doubled. Details are illustrated in the chart below.

Artificial Intelligence Boom

As shown in the figure below, the artificial intelligence boom is driving capital expenditures to their highest levels since the late 1990s. In 2025, investments in data center systems (computing, networking, and storage equipment) are expected to grow by 47% to nearly $500 billion, and are projected to increase by another 20% in 2026, reaching approximately $600 billion. This far exceeds the long-term annual trend of $150 billion to $200 billion per year in the decade prior to the launch of ChatGPT. Such a scale of investment raises the question: 'Where will the returns from these investments come from, and in what form will they materialize?'

Beyond the semiconductor industry and major cloud service companies listed in public equity markets, the growth in investment and returns also benefits native artificial intelligence enterprises that have yet to go public. Artificial intelligence companies represent one of the fastest-growing groups of businesses in history. According to our research, consumer adoption of artificial intelligence is occurring at twice the speed of internet adoption in the 1990s, as illustrated in the figure below.

According to reports, by the end of 2025, the annualized revenues of OpenAI and Anthropic will reach 20 billion US dollars and 9 billion US dollars, respectively, increasing 12.5 times and 90 times from 1.6 billion US dollars and 100 million US dollars a year ago! There are rumors that both companies are considering an Initial Public Offering (IPO) within the next one to two years to raise substantial investments to fund the models supporting their products.

As Feiji Simo, CEO of OpenAI's application business, stated: "The potential of artificial intelligence models far exceeds the scope of most people's daily experiences. The core mission for 2026 is to bridge this gap. The leaders in the artificial intelligence field will be those companies capable of transforming cutting-edge research into indispensable products for individuals, businesses, and developers." This year, substantial progress is expected in this area through more purposeful, intuitive, and integrated user experiences. An early example is ChatGPT Health — a dedicated section within ChatGPT designed to provide health and wellness assistance based on users' personal health data.

In the corporate sector, many artificial intelligence applications are still in their early stages, progressing slowly due to factors such as bureaucracy, inertia, and the need to restructure and build the data infrastructure required for constructing AI applications. By 2026, enterprises may realize the necessity of training models on their own data and iterating quickly, or risk being overtaken by more proactive competitors. AI-driven application scenarios are expected to deliver instant and high-quality customer service, accelerate product launches, and enable startups to achieve more with fewer resources.

High Market Valuation

Many investors are concerned about stock market valuations — current valuations are already at the upper end of historical ranges, as shown in the chart below.

Our own valuation assumption is that the price-to-earnings (P/E) ratio will revert to its average level of 20 times over the past 35 years.

Notably, bull markets often form during periods of compression in valuation multiples. For instance, from mid-October 1993 to mid-November 1997, the P/E ratio of the S&P 500 Index dropped from 36 times to 10 times, while delivering an annualized return of 21%; similarly, from July 2002 to October 2007, the P/E ratio of the S&P 500 Index contracted from 21 times to 17 times, with an annualized return of 14%. Given our outlook for productivity-driven acceleration in real GDP growth and moderating inflation, a similar dynamic — potentially even more pronounced — is likely to unfold during this market cycle.

As always, heartfelt thanks to the investors and all supporters of Ark Invest, as well as Dan, Will, Katie, and Keith for assisting me in drafting this lengthy New Year outlook!

NiuNiu Celebrity Tracker: Five Featured Columns! Helping you not only "do your homework right" but also understand problem-solving strategies.

Editor/joryn