Year-end review of the top ten global economic hotspots in 2025: the year of the collapse of the capital order - the United States' asset belief is shaken, AI valuation transitions from "dream" to "debt" assessment.

Looking back on 2025, this year is destined to be a historic watershed for the reshaping of the global capital markets order in the midst of severe turbulence. In this year, the "American exceptionalism" encountered a crisis of disillusionment in the boomerang effect of "equal tariffs" and the political game of the White House pushing the Federal Reserve, while the 43-day federal government shutdown plunged the market into unprecedented "data fog." At the same time, the AI revolution entered deep waters, from the "algorithm equality" stirred up by DeepSeek to the "AI debt wave" carried by tech giants, to the global frenzy for storage chips supply chain anxiety, the battle between technological progress and capital returns is becoming increasingly thrilling. In the collision of macro-narratives and micro-reforms, the logic of asset prices underwent a fundamental reversal: gold and silver staged an epic surge amidst credit unrest, while Bitcoin was trapped in macro headwinds without any scandals in the background, and Tesla, Inc. and streaming media giants sought valuation reconstruction in a life-and-death battle. Standing at the intersection of the collapse of the old order and the emergence of a new order, 2025 witnessed ten key events that shook the global capital markets, replaying the annual drama that is crucial for wealth redistribution and power shift. 1. The "equal tariffs" policy triggers a seismic shock to the "American exceptionalism," and the global assets enter a year of "bipolar resilience" battle. On April 2, 2025, the U.S. government initiated a comprehensive "equal tariffs" policy, intending to reshape the global supply chain, but unexpectedly triggered market turmoil and raised deep doubts about the long-standing dominance of "American exceptionalism." This policy backfired on the U.S. economy like a "boomerang," causing the S&P 500 index to plummet by 4.84% on April 3, the Nasdaq to plummet by 5.97%, and the Dow Jones to drop by 3.98%. Over the course of April 3 to 4, the U.S. stock market posted two consecutive days of sharp declines, with a total market value evaporation of about $6 trillion, equivalent to the total GDP of Germany for the whole year. This seismic shock not only shattered the myth of the never-ending rise of the U.S. stock market but also sparked discussions internationally about the disillusionment of the "American exceptionalism" and the return of the discourse of "the East rises while the West falls." In addition, the challenge to the "American exceptionalism" in 2025 was also due to market concerns about the continued escalation of federal debt and stubborn inflation limiting the policy space of the Federal Reserve. Markets in Europe, Japan, and other regions outperformed U.S. stocks, and there was a significant diversification of global capital allocation for the first time in a decade as investors began to reassess the risk premium of U.S. assets. Compared to the internal challenges in the United States, Eastern economies like China demonstrated greater resilience. Despite the impact of the 84% tariff on China in the first quarter of 2025, China's economy still exceeded expectations in the first quarter, with the adaptability of the industrial chain and the effectiveness of policy support being verified. The global trade pattern did not stagnate due to unilateral U.S. actions; on the contrary, there was an acceleration of industrial chain diversification and regional cooperation. However, later, as the tariff war cooled down, and the explosive profitability of artificial intelligence technology landed, and as the Federal Reserve opened its liquidity support amid inflation pressure, U.S. stocks showed an astonishing "recovery power" after the shock, with the three major indices rebounding strongly and hitting new historical highs. Despite this rebound, the discourse of the "East rises while the West falls" has lost its momentum, and the "American exceptionalism" can no longer return to the blind worship of the past. Behind the double-digit gains in U.S. stocks lies record volatility and a "roller coaster" psychological reshaping, which fully exposes the long-term erosion of the foundation of U.S. stocks by policy uncertainty. It is worth noting that the ups and downs in Sino-U.S. economic and trade relations over the year saw the suspension of the year-long tariff suspension agreement in early October. Looking ahead to the future, the expiry date of the tariff extension on November 10, 2026, will be a significant potential event affecting global markets and U.S. stocks, with the core logic being the relationship between trade frictions escalating, maintaining the status quo, or easing. 2. The White House pressures the Federal Reserve: Trump attacks the rate cut delay, preparing for an early nomination to sideline Powell. In 2025, after President Trump was re-elected as President of the United States, the power struggle between the White House and the Federal Reserve escalated into an open "total war," becoming one of the most impactful financial events of the year. The core conflict centered on the divergence of monetary policy paths: Trump repeatedly criticized Federal Reserve Chairman Powell as a "stubborn cow" and "extremely incompetent" on social media, accusing him of "too little, too late" in cutting interest rates. Despite the Federal Reserve's three consecutive rate cuts in the second half of the year, lowering the benchmark interest rate to 3.5%-3.75%, Trump still advocated for a rate cut to below 1% to stimulate the economy and the real estate market. The strong performance of 4.3% in the annualized GDP growth rate of the United States in the third quarter, on the other hand, reinforced the Federal Reserve's cautious attitude towards further rate cuts, while the Trump administration saw it as an opportunity for rate cuts, arguing that maintaining a high interest rate was artificially suppressing market performance. To weaken Powell's influence, Trump adopted a "sidelining" strategy: Powell's term was due to expire in May 2026, but as early as the summer of 2025, Trump began planning on an early nomination for a successor. This move may aim to establish a "shadow leadership" loyal to the White House within the Federal Reserve and effectively sideline Powell for the rest of his term. The selection criteria emphasized "absolute loyalty" and a "rate cut" stance, with Trump setting a "litmus test" on December 23: anyone who disagrees with the view of a "good Hershey Company rate cut" has no qualifications to be Federal Reserve Chairman. Popular candidates included White House National Economic Council Director Kevin Hassett, former Federal Reserve Board Member Kevin Warsh, and current board member Christopher Waller, with Hassett openly criticizing the delay in U.S. rate cuts. Looking ahead to the selection of the next Federal Reserve Chair after Powell's term expires in May 2026, the three main candidates, Hassett, Warsh, and Waller, have clear positions: Hassett represents the "political intervention faction," advocating closely with Trump and criticizing the delay in rate cuts, although his betting odds have dropped significantly from a peak of over 80% to 51%, market concerns about his politicalization tendency could trigger a bond market turmoil; Waller stands out with his positioning as the "data-robust faction," advocating actively for rate cuts within the Federal Reserve, his voting record in July opposing maintaining rates and the favorable view on Wall Street resonating, he is seen as a potential safeguard for stable risk assets; Warsh, with his "reform experiment faction" posture, has created market uncertainty about his policy stance, leading to diverging market expectations. The final selection will depend on Trump's balancing act between policy independence compromise and internal Federal Reserve power balance, and the ultimate outcome of this critical personnel change will serve as a key observation point for global financial markets to assess the risk of U.S. policy uncertainty. 3. AI Valuation Reconstruction: DeepSeek's Miracle and the Reversal of Alphabet Inc. Class C's Ecosystem Global AI industry underwent a disruptive valuation logic reconstruction in 2025, becoming the most strategically significant technological transformation. This transformation was led by the "algorithm equality revolution" initiated by the Chinese company DeepSeek with the innovation of multi-token prediction optimization and sparse-mixed expert model architecture, their new-generation model compressed the inference cost to less than 10% of similar products in Silicon Valley, showcasing that algorithm optimization could partially replace the dependency on top-tier GPUs. This breakthrough made "high-performance AI at the thousand-dollar cost" a reality, shaking the industry consensus of "expensive models = the only way" held by OpenAI and leading to a significant drop in NVIDIA Corporation's stock price and market capitalization by billions of dollars, prompting the market to reassess the real value of algorithm monopolies. Facing the impact of algorithm efficiency, U.S. tech giants quickly launched the "infrastructure defense war": Microsoft Corporation restarted the Three Mile Island nuclear power plant project, Amazon.com, Inc. increased its data center land reserves, and Meta accelerated the construction of a million-chip-level computing power cluster. This competition essentially boils down to a bid for the duel of "electricity + land" even though algorithm efficiency has improved, training the next generation of super-intelligent models still requires absolute superiority in physical resources, and these tech giants are attempting to solidify their lead through scale barriers in energy and infrastructure. In this transformation, Alphabet Inc. Class C made a comeback with its vertical integrated ecosystem. In the second quarter, Alphabet Inc. Class C launched the Gemini 2.0 series, directly benchmarking against DeepSeek R1 in terms of token cost and inference efficiency, using their self-developed TPU chips to achieve end-to-end cost reduction from chips to models to applications, establishing a core moat against low-cost disruptions. By the fourth quarter, with the introduction of the Gemini 3 and more widespread adoption of TPU chips, Alphabet Inc. Class C's stock price surged in late November, nearing $4 trillion in market cap, and the AI market shifted from a monopoly held by NVIDIA Corporation to a dual technology system of "Alphabet Inc. Class C + NVIDIA Corporation." More importantly, Alphabet Inc. Class C deeply integrated AI into search, Android, and Workspace, building upon its massive user base to quickly monetize AI investments, proving the ultimate advantage of the "model-chip-end terminal" integrated stack. Looking ahead, the AI industry is moving away from a valuation system based on "computing power scale + model parameter volume," towards a three-dimensional evaluation system based on "algorithm efficiency + ecosystem penetration rate + data moat." DeepSeek demonstrates the possibility of technology democratization, while Alphabet Inc. Class C proves the ultimate advantage of full-stack integration in a world where algorithms are becoming increasingly standardized, being a "full-stack player" with autonomous control over underlying chips and a top-level application ecosystem will make them the biggest winners. 4. "Chip Rush" in the global storage chip market in 2025: AI-driven structural shortages become the focus In 2025, the global storage market experienced the most severe structural imbalance in history. Although the industry's overall impact was not as widespread as in 2021, under the driving force of explosive growth in AI computing power demand, the capacity battle in the high-end storage chip sector intensified suddenly, with price fluctuations reaching new heights in nearly five years. The deep logic of this transformation has shifted from the traditional "consumer electronics-driven demand" to a structural shortage driven by the demand for AI computing power, and its effects on industry restructuring are deeply reshaping the global technology industry landscape every link in the chain, from chip design and manufacturing to end applications, is undergoing a reassessment of value and strategic adjustment, with far-reaching effects beyond simple price fluctuations, becoming a key variable in defining the competitiveness of the tech industry over the next decade. The trigger for this "chip rush" lies in AI computing power's "siphoning effect" on storage. 2025 was hailed as the peak year for AI infrastructure construction, with AI servers requiring over 30 times more memory than traditional servers. To meet the demands of companies like NVIDIA Corporation, Samsung, SK Hynix, and Micron, the three major storage giants redirected a large portion of their DRAM production capacity to High Bandwidth Memory (HBM). Although manufacturers had initially planned to completely phase out DDR4 production by the second half of 2025, the surge in market prices and strong demand from traditional industries led Samsung and SK Hynix to extend DDR4 supply until the end of 2026. Despite this, as the new capacity was nearly entirely allocated to HBM and DDR5, the effective supply of traditional general-purpose chips remained severely insufficient, leading to persistent high costs for industries like PCs and automotive. In the development trajectory, this storage crisis underwent a dramatic evolution from "structural scarcity" to "global premium." In the first half of 2025, the increase in AI infrastructure expansion drove storage chip prices to moderate levels starting in April. However, in the second half of the year, the HBM (High Bandwidth Memory) production gap burst open, leading to a direct switch to a "surge" mode from September to October. Due to the almost monopolization of capacity by AI demand, tech giants like Alphabet Inc. Class C, Microsoft Corporation, and others increased their focus on "infrastructure defense." Microsoft Corporation restarted the Three Mile Island nuclear power plant project, Amazon.com, Inc. expanded its data center land reserves, and Meta accelerated the construction of a million-chip-level computing power cluster. This competition essentially reflects an attempt at a "double monopoly" of "electricity + land" despite the improvements in algorithm efficiency, training the next generation of super-intelligent models still requires a significant advantage in physical resources like energy and infrastructure, leading these tech giants to try to solidify their leading positions through scale barriers. In this evolution, Alphabet Inc. Class C staged a comeback with its vertical integrated ecosystem. In the second quarter, Alphabet Inc. Class C launched the Gemini 2.0 series, directly benchmarking against DeepSeek R1 in terms of token cost and inference efficiency, using their self-developed TPU chips to achieve end-to-end cost reduction from chips to models to applications, establishing a core moat against low-cost disruptions. By the fourth quarter, with the introduction of the Gemini 3 and more widespread adoption of TPU chips, Alphabet Inc. Class C's stock price surged in late November, nearing $4 trillion in market cap, and the AI market shifted from a monopoly held by NVIDIA Corporation to a dual technology system of "Alphabet Inc. Class C + NVIDIA Corporation." This transition marks the beginning of the end for the structural scarcity of the storage crisis, as AI-driven structural shortages have become the focus of the technology industry, pushing forward a new era of global technology. 5. Global capital markets experience an "AI debt wave," with concerns focusing on the AI bubble bursting In 2025, global capital markets experienced a milestone event known as the "AI debt wave," with tech companies globally issuing a record amount of bonds in the year. According to Dealogic, global tech companies issued a total of $428.3 billion in bonds in 2025 by the first week of December, marking a new high. Of this amount, U.S. companies accounted for a significant portion, with bond issuances reaching $341.8 billion. Of particular note was the burst of the "AI debt wave" toward the end of the year, triggered by the significant support for AI infrastructure that led to capital expenditures far exceeding corporate cash flows. To meet the demand for expanding large-scale data centers, purchasing high-end GPUs, and upgrading power systems, tech giants saw their annual capital expenditures soar to about $400 billion. Despite the robust company profits, the immediate investment in AI infrastructure went beyond internal operational cash flows, prompting these companies to turn to the bond market for financing. Additionally, with inflation easing in 2025 and a relatively stable interest rate environment, corporations hoped to lock in low funding costs. However, the large-scale debt issuances also posed various risks. By the end of September 2025, the median debt-to-EBITDA ratio of large tech companies had risen to 0.4, close to twice the level seen in 2020, indicating that debt growth was outpacing profit growth. Market concerns about high leverage were evident in credit default swap spreads (CDS); for example, Oracle Corporation saw its 5-year CDS spread nearly double in the past two months to 142.48 basis points, with doubts raised about the sustainability of its debt-driven data center construction model. Even developers related to AI data centers issued high-yield "junk bonds" for financing, further increasing market risk. The rise of the "AI debt wave" fueled intense debates and worries among the public about the bursting of the AI bubble. Critics pointed out that while AI investments were growing exponentially, the ability to monetize this investment was at best linear. If returns did not meet expectations, the large debt burden could strain balance sheets. Investor attitudes shifted from blind endorsement to cautious scrutiny, with companies like Oracle Corporation seeing significant fluctuations in their stock prices in the fourth quarter due to debt concerns, leading investors to sell off companies that had taken on large amounts of debt without demonstrating corresponding profit growth. In addition, the massive supply of tech bonds accounted for 10% of the U.S. corporate bond market, significantly squeezing the financing space for sectors like energy and utilities. This led to a widening of financing spreads in traditional infrastructure industries and heightened the overall credit cost for the corporate bond market, indirectly tightening the financing environment across society. In 2025, AI investments shifted from being "dream-driven" to "debt-driven," with active capital markets supporting rapid technological developments. However, the excessive debt and leverage risks associated with AI investments posed a significant threat, and 2026 was pegged as a critical year whether AI projects could generate enough cash flow to pay off the debt would determine whether the bubble would land smoothly or burst. 6. "Data Fog" Under the Shadow of Shutdown: The Fiat between the Federal Government and the 2025 U.S. Federal Government Shut Down Crisis In 2025, the U.S. federal government faced its longest shutdown in history lasting 43 days, which not only led to a large-scale shutdown of federal services but also created a "data fog" due to the interruption of key economic data, plunging the Federal Reserve and financial markets into a state of uncertainty. The crisis began on October 1, 2025, the first day of the 2026 fiscal year. The core conflict arose from the profound divisions between the two parties in the 119th Congress over federal spending levels, the withdrawal of foreign aid, and health insurance subsidies, among other issues. The shutdown lasted until November 12, breaking the previous record set during the 2018-2019 shutdown, which lasted 35 days. Around 800,000 federal employees were furloughed, with 700,000 working without pay, causing significant disruptions to services provided by national health research institutions, the Centers for Disease Control and Prevention, and women and children nutrition subsidy programs. More significantly, other core economic data sources like the Labor Statistics Bureau and Economic Analysis Bureau were shut down, leading to delays or cancellations of reports such as non-farm employment and CPI for October and early November, with some data permanently missing due to the inability to collect backups, making it difficult to interpret the November data without comparable data. Federal Reserve Chairman Powell likened the situation to "driving in dense fog," forced to slow down decision-making and conduct "preventative rate cuts" in the absence of precise data. Financial markets were thrown into turmoil due to the lack of information, with investors highly sensitive to policy uncertainties and risk aversion mounting. Ultimately, the crisis was resolved when the House of Representatives passed a temporary funding bill that was signed by the President on November 12. The Congressional Budget Office estimated