"AI burning money" narrative is in a big retreat! As long as the "circular AI prosperity" can continue, the logic of the bull market in US stocks is impeccable.

As Anthropic is on the brink of profitability for the first time and the "Circular AI Boom" is spreading from the North American market to every corner of the globe, the unexpected windfall reaped by the Asian AI computing power industry in recent years is starting to support the funding ecosystem of those "hyperscale tech giants" in the North American market (such as Meta, Amazon.com, Inc., and Alphabet Inc. Class C). The flourishing global development of artificial intelligence has brought unprecedented cash flows to Asian hardware manufacturers in the AI computing power industry, with some of the funds flowing back into US dollar assets, indirectly supporting a financial environment conducive to unprecedented AI capital expenditures by large tech companies - the very low borrowing cost environment that Trump has been longing for. Leading Asian manufacturers in the AI computing power infrastructure chain, such as SK Hynix, Samsung Electronics, Taiwan Semiconductor Manufacturing Co., Ltd. Sponsored ADR, Hon Hai Precision Industry Co., Ltd., and Wistron Corporation, have earned massive windfalls in the global economy, reflecting the flow of funds within the AI ecosystem globally. The AI boom is creating an unprecedented influx of cash for smaller Asian economies like South Korea and Taiwan, which is leading to the creation of a narrower version of Asian savings surplus. At the same time, these leaders in the Asian AI computing power industry are growing into a key force in the long and healthy bull market in the US and global stock markets; previously, in the late 1990s and early 2000s, Asian savings surplus had kept US benchmark borrowing costs at historic lows. This is the viewpoint of one of the world's leading think tanks, the Oxford Institute for Economic Research, in a research report that examines how technology-producing economies convert massive export revenues into external surpluses, with the speed of this conversion outpacing the ability of domestic investment and consumer power to absorb it. As the AI intelligence wave sweeps the globe, the main investment focus in AI computing power is shifting from the "competition around GPU-centric computing power" to the "full-stack computing power system driven by AI intelligence agents." The next round of excessive alpha returns will not be limited to the strongest players in the AI GPU/AI ASIC field, but will systematically expand to the full-stack AI computing power infrastructure layer, including data center CPUs, DRAM/NAND/HBM storage, AI PCBs, liquid cooling systems, data center optical interconnection systems, ABF substrates/glass substrates, and various wafer foundries. These foundational layers, which are crucial for the ambitious AI projects of US tech giants, are mostly concentrated in the Asian market. On April 30th, three cloud computing supergiants - Microsoft Corporation, Alphabet Inc. Class C, and Amazon.com, Inc. - all delivered impressive results on the same night, highlighting the unexpectedly explosive growth of their cloud computing businesses fueled by the AI wave. The latest research report from Morgan Stanley's analyst team projects that the combined capital expenditures of the five hyperscale tech giants (Amazon.com, Inc., Alphabet Inc. Class C, Meta, Microsoft Corporation, Oracle Corporation) will reach approximately $800 billion in 2026 and are expected to surpass $1.1 trillion in 2027, an increase from the previous prediction of $950 billion. Morgan Stanley's analysts emphasize that the core logic behind these massive capital investments is to first reinvest and build capacity, followed by generating scale-based business revenue and ROIC returns based on AI computing power resources; the sharp increase in backlogged orders for cloud computing and AI application tokens is the most direct evidence that this logic is working. The unexpectedly rapid growth of the cloud computing businesses of these tech giants is prompting Wall Street to reevaluate the commercial returns on AI. In the "Circular AI Investment Boom," the logic driving the bull market in US stocks is more solid than simply "tech giants burning money to buy GPUs," as it is creating a global closed loop - US AI application and cloud companies are creating immense demand and generating strong AI-related revenue, while Asian chip/storage/server supply chains are benefiting from export dividends, and the surplus from Asia is flowing back into US dollar assets, thereby significantly supporting US tech giant's financing conditions and AI capital expenditures. The surplus from the AI computing infrastructure is circulating again! As the Circular AI Boom goes global, Asian AI dividends are feeding back into US tech giants Louise Loo, head of Asian market economic research at the Oxford Institute for Economic Research, writes in the report that part of the capital is flowing back into US dollar-denominated equity and debt assets. These investments may indirectly support funding conditions that contribute to large tech companies' capital expenditures - these large tech companies are often referred to as hyperscale tech giants, including Alphabet Inc. Class C's parent company Alphabet Inc., Meta Platforms Inc., Microsoft Corporation, Amazon.com, Inc., as well as Tesla, Inc. and Oracle Corporation. As shown in the chart above, the large tech giants in the US are planning to significantly increase their capital expenditures to over $700 billion by 2026 - these four leading global tech companies have raised their AI computing infrastructure spending plans after the first quarter. Note: The 2026 data represents company performance guidance, with Meta's data being the midpoint of the estimated range. Loo writes in the research report, "This Asian savings recycling framework into US assets echoes Ben Bernanke's surplus framework from 20 years ago." She is referring to a theory promoted by former Federal Reserve Chair Ben Bernanke. However, "the narrower range and greater concentration of Asian surpluses that are now closely related to AI computing infrastructure are also more pronounced." For most of Asia, the soaring export of AI hardware such as chips has outpaced the drag caused by the rise in energy prices due to the Iran war. This is driving Taiwan to achieve its fastest economic growth in decades and boosting Japan and South Korea, whose economies are mainly based on the export of tech products. This is also continually creating massive windfalls for manufacturing giants such as Taiwan Semiconductor Manufacturing Co., Ltd. Sponsored ADR. Taiwan Semiconductor Manufacturing Co., Ltd. Sponsored ADR is the core chip manufacturer for NVIDIA Corporation and Apple Inc.; Samsung Electronics' semiconductor division saw its first-quarter profit surge by 48 times. At the macroeconomic level, this prosperity is creating significant imbalances. Data released on Wednesday shows that Taiwan's current account surplus in the first quarter skyrocketed by a record 111% year-on-year to $62.5 billion. Meanwhile, there are no signs of a slowdown in the demand for AI chips; export data from South Korea in the first 20 days of May showed that exports surged by nearly 53% year-on-year after adjusting for workday differences. ASML Holding NV ADR, the Dutch lithography machine giant, has predicted that the booming global semiconductor market will face long-term supply shortages in the foreseeable future and has even forecasted that the global semiconductor market could reach a staggering $15 trillion by 2030. ASML Holding NV ADR's CEO Christophe Fouquet stated on Wednesday that "the strong demand for AI is so robust that the market will remain supply-constrained for a considerable period of time." The strong demand for ASML Holding NV ADR's EUV equipment, which is crucial for advanced chip manufacturing, is being fueled by the near insatiable demand for AI GPUs/ASICs and HBM/DRAM storage chips from advanced chipmakers such as Taiwan Semiconductor Manufacturing Co., Ltd. Sponsored ADR and Intel Corporation, helping this Dutch semiconductor equipment manufacturer become the most valuable company in Europe. The excess profits accumulated by emerging economies in Asia now far exceed the surpluses accumulated by Arab economies in the Gulf; the latter was once the core pillar of the petrodollar system, which relied on oil exports priced in dollars. According to a report by Gavekal Research, "the trend toward de-dollarization is more likely to depend on how Northeast Asia manages its surpluses, rather than a reduction in the recycling of Middle Eastern petrodollars." However, the surplus is concentrated in a few upstream Asian tech economies, and the final demand is concentrated in US hyperscale cloud computing companies, creating a certain level of potential vulnerability. Loo warns in the report, "This financial recycling loop is very powerful but may be exposed to shifts in the US-led AI cycle, foreign exchange pressures, and rising risks to domestic institutions' balance sheets." The largest US tech companies are currently planning capital expenditures of up to $725 billion this year, mainly for high-performance equipment for AI data centers. Asia's role in these investment plans is crucial. Loo points out that over 55% of advanced technology products imports into the US come from Asian economies, reflecting a production advantage that other regions find difficult to replicate. According to the Oxford Institute for Economic Research, while the unprecedented investment boom around AI could drive significant benefits in exports and growth for Asia, the effect on the wider economy will be limited due to factors like dependence on imported equipment, strong foreign ownership, capital intensity of production, and geographical concentration. As South Korea and Taiwan accumulate massive current account surpluses, the IMF predicts that external balances will gradually stabilize in the coming years. From Claude's profit turning point to the Asian chip super surplus, the narrative of "burning money" seems to be outdated Loo's forecast data shows that every $100 increase in AI-related hardware imports into the US could increase the GDP of major Asian suppliers by at least $35. Loo indicates that the main beneficiaries are China, South Korea, and Vietnam. In a report last week, an economist team led by Andrew Tilton, chief economist at the Goldman Sachs Group, Inc., described the surge in technology exports as creating what they termed an "AI-driven super surplus," which could see South Korea's current account surplus exceed 10% of GDP in 2026. The analysts and economists at Goldman Sachs Group, Inc. stated that so far, South Korea's onshore fiscal surplus has mainly been recycled into overseas equities markets, while Taiwan's surplus has mainly flowed into foreign exchange deposits, but pressure for appreciation is building up. The Oxford Institute for Economic Research's Loo reached a similar conclusion. Loo writes, "Policy interventions and portfolio outflows can indeed manage currency appreciation pressures for a period, but as external debt increases relative to GDP, the tension between external market fundamentals and current exchange rate settings will become increasingly difficult for policymakers to ignore." The latest developments in the "Circular AI Boom" and Anthropic's expected doubling of revenues in Q2, along with its first achieved operating profit, underscore the shift in the AI industry chain from a "burning money narrative" to a "cash flow cycle narrative". Recent market trends show that Anthropic, the AI company that stunned the global software stocks in February with a series of powerful AI agents, is projected to increase its Q2 revenue from $4.8 billion in the first quarter to $10.9 billion, with an operating profit of approximately $559 million, indicating that cutting-edge AI applications are not just consuming computing power but are transforming enterprise programming, intelligent workflow, network security, and data analytics into high-value token returns. The profit inflection point of Anthropic driven by Claude and the surge in demand for AI tools is approaching, with the cost of computing power per $1 of revenue decreasing from about 71 cents in the first quarter to about 56 cents in the second quarter, indicating that economies of scale and reasoning efficiency are improving the economic model of AI applications. Against the backdrop of the Asian chip super surplus, the profits of the Asian AI computing power chain are directly linked to the logic supporting the long bull market in US stocks From an industrial chain perspective and an equity market perspective, the profitability inflection point of Anthropic and the "super surplus" of the Asian AI computing power chain form a closed loop: US AI application companies are generating huge revenues, profits, and boosting their valuations on the stock exchange, Hyperscalers are increasing AI capital expenditures, Asian chip manufacturing chains, HBM, servers, and advanced packaging suppliers are reaping export dividends, and the surplus from Asia, flowing back in the form of US dollar assets, foreign exchange deposits, and overseas investments, is supporting the American Financial Group, Inc. environment, especially the "low rates" that Trump has been keen on. At the same time, the Asian AI computing power chain is gradually becoming a critical force in the long and healthy bull market in the US stock market. Therefore, the Asian AI computing power infrastructure chain is not only the hardware base on which US tech giants can implement their AI infrastructure projects, but it could also help absorb the financing pressure brought by the US AI capital expenditures through profit reinvestment, capital allocation, and bond purchases. For example, the surplus or cash earnings earned by Asian chip manufacturers such as SK Hynix, Samsung, and Taiwan Semiconductor Manufacturing Co., Ltd. Sponsored ADR from the AI frenzy are flowing through the global economy, forming a fund chain similar to the "Circular AI Boom". If the surplus from the Asian AI industry chain is converted into investments in US dollar assets, US corporate bonds, US tech bonds, and long-term bonds, it could indeed help absorb the US AI financing wave, thereby lowering credit spreads, improving the financing conditions for tech giants, and supporting the long-term bull market in US stocks. For every $100 increase in AI-related hardware imports into the US, the GDP of major Asian suppliers could increase by $35, driving Asian economies to accelerate the formation of an "AI-driven super surplus" - which means that the AI boom is no longer just a story of expanding valuations in the application layers of OpenAI/Anthropic, but a global fund cycle of "US AI application profit expansion - large cloud company Capex - Asian AI computing power infrastructure-class hardware exports - US dollar asset feedback".