Market Overview - Asian stocks experienced a decline as investors expressed concerns over technology valuations and reacted to disappointing economic data from China [1] - Chinese shares fell due to bondholders of distressed developer China Vanke rejecting a bond payment extension proposal, highlighting structural weaknesses in the economy [2] - The Japanese market also tumbled, with tech shares following their U.S. counterparts lower amid valuation concerns [4] Economic Data - China's industrial production grew by 4.8 percent year-on-year in November, missing forecasts of 5.0 percent and down from 4.9 percent in October [2] - Retail sales in China rose by 1.3 percent annually, falling short of the expected 3.0 percent increase and down from 2.9 percent in the previous month [3] - Fixed asset investment in China decreased by 2.6 percent year-on-year, missing forecasts for a loss of 2.4 percent [3] Sector Performance - In Japan, the Nikkei average dropped by 1.31 percent, while the broader Topix index saw a slight increase of 0.22 percent [4] - Major decliners included Advantest, which plunged by 6.4 percent, and SoftBank Group, which lost 6 percent [4] - In South Korea, the Kospi average fell by 1.84 percent, with Samsung Electronics and SK Hynix losing 3.8 percent and 3 percent, respectively, due to concerns over AI margins [5] Commodity and Currency Movements - Gold prices surged nearly 1 percent to a seven-week high as the U.S. dollar struggled ahead of key economic data releases [1] - Oil prices increased slightly due to supply disruptions in Venezuela, despite concerns over oversupply in the new year [1] Investor Sentiment - U.S. stocks ended significantly lower as investors took profits from high-flying AI-related stocks amid valuation concerns [7] - Higher Treasury yields negatively impacted market sentiment, with the tech-heavy Nasdaq Composite dropping by 1.7 percent [8]

Investment Rating - The report maintains a cautious outlook on the smartphone industry, with a preference for component suppliers over OEMs due to rising memory prices and supply challenges [5][8]. Core Insights - The smartphone industry is expected to face challenges in 2026, with limited unit growth forecasted at +1.0% YoY in 2026 and +2.0% YoY in 2027, following a +3.5% YoY growth in 2025 [2][8]. - The UBS Evidence Lab 4Q25 Smartphone Survey indicates a moderate increase in purchase intent, with 40% of respondents planning to buy a smartphone in the next 12 months, up from 36% in 2Q25 [3][19]. - Rising memory prices are projected to significantly impact Bill of Material (BOM) costs, potentially accounting for approximately 14% of flagship smartphone BOM costs and 34% of mid-range/lower-end smartphone BOM costs by 4Q26 [4][101]. Summary by Sections Industry Outlook - The smartphone market is anticipated to grow moderately, with unit sell-in forecasts slightly raised to 1.27 billion units in 2025 and 1.28 billion units in 2026, reflecting a +3.5% and +1.0% YoY growth respectively [50][55]. - The average age of the smartphone installed base has decreased to 22.0 months, indicating a normalization in the replacement cycle [30][44]. Purchase Intent Trends - Purchase intent for the iPhone 17 series is strong, with 66% of respondents interested in this model, up from 61% for the iPhone 16 series [3][89]. - Retention rates for Apple and Samsung remain stable at 87% and 75% respectively, while Chinese OEMs have seen declines in retention rates [59][61]. Component Supplier Preference - The report favors component suppliers such as ASE, SK Hynix, and TSMC, while maintaining neutral ratings on major OEMs like Apple, Lenovo, and Xiaomi, and a buy rating on Samsung Electronics [5][8]. - The memory supply shortage is expected to create significant challenges for OEMs, particularly smaller ones, as they may struggle to secure adequate memory supply [4][110].

Accessible version We hosted a memory expert call on 12 Sep. The guest speaker was Avril Wu of TrendForce. Wu was not bullish on HBM for 2026 given her assumptions for bit growth (only +27% YoY vs +94% in 2025) and ASP (down 10% vs +21%) are more conservative vs our current forecasts (bit growth +51%, ASP -6%). Wu also expects Samsung Electronics' HBM4 to be more successful (vs 12-hi HBM3e) given superior tech specs (1c node, 4nm logic die) vs peers. However, we still assume Hynix's market-dominant position ...

Summary of Global Memory Technology Conference Call Industry Overview - **Memory Market Growth**: The global memory market is projected to experience significant growth, with DRAM sales expected to reach **US$196 billion** in 2026, representing a **51% year-over-year increase**. NAND sales are anticipated to hit **US$114 billion**, reflecting a **45% increase** [1][10]. - **ASP Increases**: Average Selling Prices (ASP) for DRAM and NAND are expected to rise by **33%** and **26%** year-over-year, respectively, contributing more to revenue than bit growth, which is projected at **14%** for DRAM and **15%** for NAND [1][10]. Key Points 1. **2026 Super-Cycle**: The memory market is expected to enter a super-cycle similar to the 1990s, driven by AI advancements and a prolonged upturn, with sales growth continuing into **2027** [1][10]. 2. **HBM Demand Growth**: High Bandwidth Memory (HBM) demand is projected to grow by **50-100% annually**, driven by new GPU and ASIC technologies, while wafer capacity growth is expected to remain below **50%** [2]. 3. **Supply Constraints**: High capital expenditures (capex) are anticipated, with SK Hynix's capex estimated at **W23 trillion** in 2025 and **W35 trillion** in 2026. However, productivity is expected to be low, leading to a production volume increase of less than **20%** year-over-year for DRAM and NAND [3]. 4. **Sustainable High ASP**: Despite spot price corrections, conventional DRAM prices are expected to stabilize between **US$10-15**, indicating a potential upside of over **50%** for contract prices [4]. 5. **Top Investment Picks**: SK Hynix is identified as the top pick in the global memory sector, with Samsung Electronics and Nanya Tech also recommended. EPS estimates and price objectives for all three companies have been raised due to favorable pricing conditions [5][6]. Company-Specific Insights - **SK Hynix**: Projected operating profit for 4Q25 and 2026 is **W16 trillion** and **W77 trillion**, respectively. The new price objective is set at **W900,000**, up from **W800,000**, with a **23%** increase in EPS estimates based on higher DRAM ASP [6]. - **Samsung Electronics**: Expected operating profit for 4Q25 and 2026 is **W15 trillion** and **W76 trillion**. The new price objective is **W150,000**, up from **W140,000**, with a **17%** increase in EPS estimates [6]. - **Nanya Tech**: Benefiting from legacy DRAM shortages, the new price objective is **NT$200**, up from **NT$178**, with a **25%** increase in EPS estimates [6]. Additional Insights - **Inventory Levels**: As of December 2025, both DRAM and NAND inventories are at **3-4 weeks**, lower than the normal **1-2 months**, indicating tight supply conditions [12]. - **Utilization Rates**: Memory chipmakers' fabs are fully utilized, excluding old idle capacity, suggesting strong demand [14]. - **Sales Mix**: The sales mix for DRAM is increasingly driven by server and smartphone applications, with servers expected to account for **55%** of DRAM sales by 2027 [17]. Conclusion The global memory market is poised for substantial growth driven by technological advancements and increasing demand across various applications. Investment opportunities are favorable for key players like SK Hynix, Samsung Electronics, and Nanya Tech, with a focus on ASP increases and supply constraints shaping the market dynamics.

ab 11 December 2025 Global Research Memory Semis Monthly December '25 Edition: Going into overdrive 4Q25 and 1Q26 DDR contract pricing could be much stronger than expected With 4Q25 memory contract pricing negotiations now completed, we revise up our forecast for DDR contract pricing to +35% QoQ (was +21%), and for NAND +20% (was +15%). We believe customers are trying to secure 1Q26 contract pricing in earnest, with further potential upside. We now forecast blended 1Q26 DDR contract pricing to increase 30% ...

Core Insights - The AI infrastructure buildout is accelerating across semiconductors, cloud computing, and software applications, leading to increased competition and concentration risk for investors [2] Group 1: Fund Overview - The Global X Artificial Intelligence & Technology ETF (NASDAQ:AIQ) provides broad exposure to 86 AI companies, holding $7.0 billion in assets with a 0.68% expense ratio [3][8] - AIQ aims to offer comprehensive AI participation without concentrating capital in a few companies, combining investments in AI infrastructure and applications [4] Group 2: Portfolio Composition - No single position in the AIQ portfolio exceeds 4.5%, with Alphabet (NASDAQ:GOOGL) at 4.44%, Samsung Electronics at 3.92%, and Advanced Micro Devices (NASDAQ:AMD) at 3.63% [5] - Information technology constitutes 52.3% of holdings, while communication services and consumer discretionary sectors add another 16% [5] Group 3: Performance Metrics - AIQ has achieved a return of 26.29% over the past year and 30.89% year-to-date through December 12, 2025, with an annualized return of 17.91% since inception in May 2018 [6][8] - The fund has experienced recent volatility, with a drawdown of 12% from $53.76 to $47.33 before recovering to $50.52 [6] Group 4: Cost Considerations - The 0.68% expense ratio is higher than broad market index funds by approximately 0.65 percentage points annually, leading to significantly higher fees over long-term investments [7]

Group 1: OKLO and Nuclear Power - Oklo Inc. Class A stock saw over 25% weekly gains after entering an ATM equity distribution agreement to issue up to $1.50 billion in common stock for financing its small modular reactor technology, particularly for AI data centers [1] Group 2: Natural Gas Market - The ProShares Ultra Bloomberg Natural Gas ETF (BOIL) gained more than 22% in the last week as U.S. natural gas prices surged to a three-year high due to intense winter cold and increasing export flows [2] Group 3: Cannabis Industry - The AdvisorShares MSOS Daily Leveraged ETF (MSOX) achieved approximately 21% weekly return, driven by a major acquisition announcement in Europe and positive sentiment regarding potential U.S. federal policy changes in the cannabis sector [3] Group 4: Palantir Technologies - Palantir Technologies' stock price increased significantly, supported by enthusiasm for its AI product suite and strong growth in both government and commercial sectors [4] Group 5: Semiconductor Sector - The Direxion Daily Semiconductor Bull 3x Shares (SOXL) recorded over 12% returns last week, benefiting from AI momentum and a stronger-than-expected decline in inflation expectations [5] Group 6: Quantum Computing - The Defiance Daily Target 2X Long IONQ ETF (IONX) was driven by renewed momentum in the quantum computing sector, with notable commentary from Nobel laureate John Martinis on China's advancements in this field [6] Group 7: Transportation Sector - The Direxion Daily Transportation Bull 3X Shares (TPOR) returned approximately 11.6% last week, fueled by optimism around stabilizing trade policies, improving freight rates, and strong e-commerce demand [7] Group 8: Tesla and AI Chips - The Direxion Daily TSLA Bull 2X Shares (TSLL) featured approximately 11% returns as Tesla's stock jumped following CEO Elon Musk's announcement about plans to manufacture AI chips at unprecedented volumes [8] Group 9: S&P 500 Performance - The Direxion Daily S&P 500 High Beta Bull 3X Shares (HIBL) provided 3x leveraged exposure to high-beta stocks, with the S&P 500 gaining due to investor expectations of an interest rate cut by the Federal Reserve and strong individual stock performances [9] Group 10: South Korean Market - The Direxion MSCI Daily South Korea Bull 3X Shares (KORU) saw significant performance due to global enthusiasm for AI and advanced chips, benefiting major companies like Samsung Electronics and SK Hynix [10]

Market Overview - Asian stocks experienced a rally following a less hawkish outlook from the U.S. Federal Reserve, with the U.S. dollar remaining stable and gold prices near a seven-week peak [1] - Oil prices increased due to concerns over potential supply disruptions from Venezuela [1] Chinese Market - Chinese shares mostly rose, with the Shanghai Composite increasing by 0.41% to 3,889.35, supported by a commitment from Beijing to maintain a "proactive" fiscal policy aimed at stimulating consumption and investment [2] - The Hang Seng index in Hong Kong surged by 1.75% to 25,976.79, driven by expectations of further growth initiatives from Chinese authorities [2] Japanese Market - Japanese stocks rallied, with the Nikkei average climbing 1.37% to 50,836.55, influenced by reports of SoftBank Group and Nvidia discussing a $1 billion investment in Skild AI [3] - The broader Topix index rose by 1.98% to 3,423.83, while SoftBank shares increased by 3.9% [4] South Korean Market - Seoul stocks rebounded with the Kospi average surging 1.38% to 4,167.16, following strong earnings from Broadcom [5] - SK Hynix's shares rose by 1.1% despite a warning advisory from the Korea Exchange [5] Australian and New Zealand Markets - The S&P/ASX 200 index in Australia rose by 1.23% to 8,697.30, led by gains in financials and mining sectors [6] - New Zealand's S&P/NZX 50 index finished slightly higher at 13,406.91, following a survey indicating slight expansion in the manufacturing sector [6] U.S. Market - U.S. stocks ended a volatile session mostly higher, with the Dow increasing by 1.3% and the S&P 500 rising by 0.2%, both reaching new records [7][8] - The Nasdaq Composite, however, slipped by 0.3% due to Oracle's disappointing revenue and guidance, raising concerns about AI overspending [8]

Joshua Meyers - Specialist Sales - US TMT AC (1-617) 310 0767 joshua.meyers@jpmorgan.com J.P. Morgan Securities LLC North America Specialist Sales 09 December 2025 J P M O R G A N H200s, Hynix, MRVL, TER, ANET, STX, ALAB, AVGO/CIEN Buyside Bars Joshua Meyers +1 617 310 0767 joshua.meyers@jpmorgan.com Yesterday Semifor reported that the Trump administration would approve H200 shipments to China, which by the evening it appeared to confirm (with a 25% US Treasury revenue share), with Trump saying Xi "responde ...

Core Insights - The demand for AI is shifting from training workloads to inference, with a significant increase in token processing, indicating a growing need for computing power and chips [1][16] - Hyperscaler capital spending is rising, with global data center capex increasing by 53% year-over-year in Q1 2025, driven by persistent demand for AI workloads [2] - The semiconductor market is projected to grow by 15% in 2025, reaching a total value of $728 billion, with strong growth expected in the Americas and Asia Pacific [2] Group 1: AI Demand and Inference - AI demand has transitioned towards inference, where trained models process new data, leading to increased token generation and associated costs [1] - Google reported processing 480 trillion tokens in April 2025, a 50-fold increase from the previous year, highlighting the surge in AI model usage [1] - The launch of new reasoning models is enhancing AI's ability to tackle complex problems, increasing the demand for computational resources during inference [3] Group 2: Capital Expenditure and Infrastructure - Major tech companies like Amazon and Meta are significantly investing in data center infrastructure, with Amazon planning to invest at least $20 billion in Pennsylvania and $13 billion in Australia [2] - Meta is expanding its capital spending to build multi-gigawatt data center clusters to support its AI initiatives, with the first facility expected to be operational next year [2] - The global sales of semiconductors reached $60 billion in June 2025, marking a 20% year-over-year increase, driven by the expansion of data centers [2] Group 3: Custom AI Chips and Technology - Hyperscalers are increasingly adopting ASICs for AI workloads, which are more efficient and cost-effective compared to traditional GPUs [5] - Google introduced its seventh-generation Tensor Processing Unit (TPU) designed specifically for inference workloads, expanding access to enhance cloud business growth [5] - The custom computing device market is projected to grow to $55.4 billion by 2028, indicating a strong trend towards specialized AI hardware [5] Group 4: High-Bandwidth Memory (HBM) Technology - HBM technology is expected to capture over 50% of the DRAM market by 2030, driven by the increasing computational demands of AI [6] - SK Hynix, a major HBM supplier, anticipates a 30% annual growth in the global HBM market through 2030 [6] Group 5: Semiconductor Industry Performance - Nasdaq's PHLX Semiconductor Index (SOX) delivered a total return of 96% over the past three years, outperforming other semiconductor indices [7][18] - Nvidia, the largest constituent of SOX, achieved a 52% return over the past year and became the first company to reach a $4 trillion market valuation [13] - Broadcom, another key player, generated an 85% return over the same period, dominating the AI ASIC market and engaging with major hyperscalers [14]