Recent statistical data reveals that overall corporate earnings in emerging markets have, for the first time in four years, exceeded analyst consensus expectations. This provides investors with a new and exceptionally solid reason to believe that the "emerging market super-cycle" driven by the AI computing infrastructure investment theme is just beginning. It increasingly highlights that the global investment frenzy around AI computing power is "expanding from a singular narrative focused on AI chips to a comprehensive re-rating of the full-stack AI computing infrastructure layer." This layer encompasses numerous elements, including CPUs, HBM/DRAM/NAND memory chips, and high-speed data center interconnection systems dominated by optical links.
Simultaneously, recent research reports from Wall Street financial giants like Goldman Sachs, Morgan Stanley, and JPMorgan indicate that the global stock market bull run led by the AI computing supply chain is far from over. As the latest signs of the U.S. and Iran nearing an end to a new round of Middle East conflict fuel market optimism, the risk premium associated with the Iran conflict has diminished. Traders have consistently resumed pouring into the highly popular pre-conflict trading themes, such as short-term U.S. Treasuries, Asian currencies, and the AI computing supply chain.
Research suggests that stronger emerging market earnings data could persuade more traditional asset managers on Wall Street and in Europe to shift asset allocations towards emerging market equities. A shift of just 5% in portfolio weight from the U.S. market could potentially lead to an increase of approximately 30% in emerging market allocation size. The AI computing supply chain, led by companies like Nvidia, AMD, ARM, SK Hynix, and Micron, represents the most robust investment theme in this pre-conflict trading playbook. Within the stock market, stocks directly tied to AI computing infrastructure—the "AI computing power super-group" spearheaded by Nvidia, SK Hynix, and AMD—are typically the most sensitive, quickest to move, and experience the largest gains during overall market and tech stock rebounds. The core logic behind their vanguard rebound is exceptionally "hardcore": it is directly tied to the tech giants' record-breaking trillion-dollar AI capital expenditures, not merely storytelling.
A recent report from the globally renowned research firm IDC shows that the world's highest market-cap company—the AI chip superpower Nvidia (NVDA.US)—has for the first time become the leading supplier in the global data center Ethernet switch market by revenue. IDC's latest findings align with views from Wall Street giants like Morgan Stanley, Goldman Sachs, and Bank of America, indicating that leaders in the AI computing supply chain, such as Nvidia, are expanding their reach from "single-point control of GPU/AI chips" to a system-level "AI factory" closed-loop encompassing GPU cabinet clusters, networking, DPUs, optical interconnect systems, software ecosystems, and the data center power chain.
Morgan Stanley states that the AI computing arms race has entered a system-level expansion phase. The firm has significantly revised its 2026 capital expenditure forecast for U.S. mega-cap tech giants from $433 billion a year ago to $805 billion. For 2027, capital expenditures are expected to reach $1.1 trillion, up from a previous forecast of $950 billion. The latest projections from Morgan Stanley, Goldman Sachs, and other major Wall Street banks highlight that supply chain bottlenecks at the AI computing infrastructure level have expanded from "massive purchases of GPUs/ASICs" to "striving to simultaneously address the entire AI data center delivery chain." This includes data center power equipment, liquid cooling, data center CPUs, DRAM/NAND/HBM, data center optical communication/interconnects, high-performance Ethernet network infrastructure, transformers, and gas turbines.
Furthermore, Morgan Stanley predicts that by 2028, nearly $3 trillion in AI-related infrastructure investment will flow through the global economy, with over 80% of the spending still ahead. The earnings surprise from emerging markets further fuels the AI bull market narrative: Asian AI supply chain leaders are driving earnings, with fundamentals taking over from liquidity-driven trading.
According to the latest compiled statistics from Bloomberg Intelligence, the "average annual earnings" reported by constituent companies in the MSCI Emerging Markets Index have, for the first time since April 2022, exceeded the consensus expectations set by analysts a year ago. Asian AI computing supply chain leaders, such as Taiwan Semiconductor Manufacturing Company (TSMC), Hon Hai Precision Industry (Foxconn), SK Hynix, and Samsung Electronics, are driving this outcome. However, overall profits in other market sectors are also improving, such as Indian refiners and Brazilian power companies.
A MarketWatch report also notes that TSMC, Samsung Electronics, and SK Hynix collectively account for approximately 28% of the MSCI Emerging Markets benchmark index. It states that the year-to-date gains of South Korea's benchmark KOSPI index (around 115%) and Taiwan's weighted index (around 59%) clearly constitute the main engine of this emerging market super-cycle and earnings surprise.
Asia hosts numerous AI computing infrastructure manufacturing companies, including core chipmakers like TSMC, as well as Foxconn, SK Hynix, Samsung, MLCC giants like Murata and Taiyo Yuden, and PCB leaders like Wus Printed Circuit and Shengyi Technology. Therefore, in the view of Wall Street analysts, Asia's AI computing infrastructure supply chain is poised to become the biggest winner of the "AI disrupts everything" trend.
The strongest theme within the current AI investment narrative is undoubtedly the AI computing infrastructure manufacturing/foundry segment, characterized by "supply constraints and extremely high technical barriers." This includes advanced process foundry, advanced packaging, HBM/high-end server memory, critical AI server components, data center power, and liquid cooling/thermal management equipment. This is because they translate the unit economics of AI training/inference systems into "computing power and energy consumption per token," and these segments are largely concentrated in Asia.
As the MSCI Emerging Markets equity benchmark has surged nearly 30% year-to-date, evidence of healthy profit growth is sending a strong signal to bulls: this rally is built on solid fundamentals, not a speculative AI bubble. This has prompted major Wall Street investment institutions like Morgan Stanley and JPMorgan to predict that the gains will soon spread beyond AI computing-related stocks.
"This is a real inflection point," said Archie Hart, a senior analyst at Ninety One UK Ltd. "The market is finally getting fundamental validation, rather than running ahead of fundamentals." Over the 12 months through May, the MSCI-reported weighted average earnings per share for the MSCI EM Index was 95.1 index points, higher than the 94.6 blended forward estimate analysts made a year ago.
Stronger earnings data could convince more traditional Wall Street asset managers to shift funds into emerging market equities, helping propel this rally into its next phase. According to a calculation from a research report led by Hart, due to differing market sizes, a 5% shift in portfolio weight from the U.S. would translate to roughly a 30% increase in emerging market equity allocation. He also emphasized that emerging market tech companies still trade at a significant valuation discount compared to their U.S. peers while generating faster earnings growth, which is another core bullish argument.
For instance, the valuation of a U.S. semiconductor equipment manufacturer index covering giants like Applied Materials, Lam Research, and KLA exceeds 46 times forward 12-month expected earnings. In contrast, the broader tech-focused MSCI Emerging Markets Information Technology Index, which includes Tokyo Electron (a competitor to Applied Materials), is valued at just 12.3 times.
Earnings exceeding expectations are a key component of the emerging market recovery taking shape through 2025. Supported by larger-scale AI spending and Chinese economic stimulus measures, profits have improved markedly since last year. Before this, emerging market profits fell by about 25% between 2022 and 2024 as higher interest rate expectations and U.S. Treasury yield curves weighed on growth.
Among the super-giants closely tied to AI computing, South Korea's SK Hynix reported Q1 profits 43% above expectations, Samsung Electronics exceeded by 16%, and TSMC beat by 5.7%. Among other top performers, Indian Oil Corporation exceeded by 33%, while Brazilian power producer Eneva SA surpassed by 44%.
"Performance across regions may continue to differ, but the direction of travel for almost all of them is now very positive, and that has not been the case for much of the past decade," said Jitania Kandhari, deputy chief investment officer for Morgan Stanley Investment Management. However, she also noted that the dominance of the AI computing trade theme is raising concerns about concentration risk in market index weightings.
Compared to other areas of emerging markets, Asian companies with the highest exposure to AI computing infrastructure are beating expectations by a wide margin, while the magnitude of beats in other emerging market sectors is more modest. Energy companies began exceeding earnings consensus this quarter, while financials are expected to cross that threshold by the end of 2025. Traditional commodity and industrial companies reported profits close to expectations. But in other market segments, missing expectations remains the norm. Consumer staples and discretionary companies are the biggest laggards in earnings, while healthcare, real estate, and utilities also largely underperformed.
"Internally, there's still narrowness," said Ashish Chugh, portfolio manager at Loomis Sayles. "Most of the EPS growth will come from the AI-related tech sector." Other factors support the case for a broader earnings recovery. As the Chinese economy shows resilience, investors see room for a wider industrial recovery. Additionally, equity issuance is slowing, and buybacks are increasing, thereby boosting earnings per share.
Hart noted that since 2010, the Chinese stock market has been flooded with massive new share issuance, depressing EPS growth by as much as 6 percentage points even when underlying businesses performed well. He emphasized this trend is now reversing. JPMorgan Asset Management expects emerging market demand to continue its strong revival pace, supporting profit growth.
Anuj Arora, chief investment officer for emerging market equities at JPMorgan Asset Management, stated that with China's economic recovery and accelerating inflation, the investment backdrop should become more favorable for sectors like AI, industrials, defense, and commodities. "A weaker U.S. dollar, sustained deficit spending by major economies, and the multi-year AI application penetration and AI computing infrastructure capital expenditure cycle continue to create a constructive backdrop for emerging markets," Arora said. "The 'AI Computing Arms Race Narrative' Expands from GPUs to CPUs, Memory Chips, Ethernet Switches, and More."
The core of emerging markets' earnings outperformance is not only the strong fundamental validation following the MSCI EM Index's near-30% rise but also highlights how AI capital expenditures are transmitting profits from a few GPU/AI ASIC suppliers to a broader range of players. These include Asian semiconductor manufacturing, optical interconnect component makers, memory, foundries, and the power industrial chain.
MarketWatch recently pointed out that TSMC, Samsung, and SK Hynix collectively account for about 28% of the MSCI EM Index, with AI chip demand driving strong gains in South Korean and Taiwanese stock markets. Some macroeconomists are even beginning to emphasize that the Asian AI computing investment boom is improving overall corporate profits, wages, consumption, and fiscal revenues in South Korea and Taiwan. This indicates AI investment is transitioning from a "tech stock trade" to a "regional economic re-acceleration variable."
Nvidia becoming the leading supplier in the data center Ethernet switch market by revenue for the first time is a key piece of evidence for this diffusion of investment. As the IDC report shows, leaders in the AI computing supply chain like Nvidia are expanding their reach from "single-point control of GPU/AI chips" to a system-level "AI factory" closed-loop encompassing GPU clusters, networking, DPUs, optical interconnect systems, and software. This reveals that the bottleneck for AI training and inference clusters is no longer just "how many GPUs to buy," but whether GPUs can synchronize, transfer parameters, schedule inference, and isolate multi-tenancy through a low-latency, high-bandwidth, low-packet-loss, high-utilization network.
IDC data shows Nvidia's data center Ethernet switch revenue surged 192.7% year-over-year to $2.1 billion, with Spectrum-X representing the explosive demand for AI factory network infrastructure. This means DPUs, Ethernet switches, optical modules, CPO, cables, data center optical interconnects, and networking software are no longer peripheral accessories to GPUs but core capital goods determining cluster utilization and inference costs.
In Goldman Sachs' view, the global bull market around the AI computing chain is far from over. The market's main theme has evolved from the long-standing post-2008 "valuation expansion driven by programming/code-based software and light assets" to a "re-rating of AI computing infrastructure centered around a series of physical assets." Goldman Sachs' latest model projects global AI capital expenditures to grow from an annual $765 billion in 2026 to $1.6 trillion annually by 2031, with cumulative spending from 2026-2031 estimated at around $7.6 trillion. U.S. data center power demand is expected to rise from 31 GW in 2025 to 66 GW in 2027. This will directly spill AI computing infrastructure investment over into server CPUs, DRAM/NAND/HBM, advanced packaging, liquid cooling, power equipment, transformers, gas turbines, grid connection equipment, data center REITs, and engineering construction.
In other words, from Goldman Sachs' perspective, the AI bull market is far from declaring its end. It has moved from the "AI chip purchasing frenzy" into the second stage of "large-scale AI factory construction." The next round of excess alpha returns will no longer belong solely to the strongest leaders in the AI GPU/AI ASIC field but will systematically diffuse across the full-stack AI computing infrastructure layer of the "AI factory." This includes data center high-performance CPUs, DRAM/NAND/HBM memory, AI PCBs, liquid cooling systems, data center optical interconnect systems, ABF substrates/glass substrates, MLCCs, electronic fabrics, and a wide range of foundry services.
Nvidia CEO Jensen Huang stated last Wednesday that AI infrastructure could revitalize U.S. factories, with artificial intelligence potentially ushering in a new era of U.S. manufacturing and industrial growth. In the view of the analyst team at Wall Street giant Bank of America, AI computing infrastructure is entering a more sustained and broader capital expenditure cycle.
Almost simultaneously, a research report from another Wall Street giant, Morgan Stanley, indicated the AI computing arms race has entered a system-level expansion phase. Demand for AI infrastructure is exhibiting a rare "inelastic" trend—regardless of the cost curve, tech giants continue to ramp up construction of AI data centers. This "demand inelasticity" is expected to persistently strengthen U.S. economic resilience and overall S&P 500 earnings growth. Morgan Stanley predicts that by 2028, nearly $3 trillion in AI-related infrastructure investment will flow through the global economy, with over 80% of the spending still ahead.
Comments