Summary of Conference Call on Zhipu and Minimax Company Overview - **Zhipu**: Founded in 2019, leveraging Tsinghua University technology for large model algorithm research, focusing on AI model capabilities and applications in various industries [8][10] - **Minimax**: Established in 2022, aims to advance AI technology and achieve AGI, emphasizing efficiency and accessibility of AI models for a broader user base [29][30] Key Points on Zhipu - **Investment Highlights**: - Core value derived from model capabilities, with a focus on coding models [3] - Recent model releases (GN5, GN4.5) show significant performance improvements, positioning Zhipu as a leading model company in China [4][11] - **Revenue Growth**: - Anticipated acceleration in revenue growth in 2026 and 2027, with comparisons to Anthropic's revenue projections [5][10] - API-related revenue is rapidly increasing, with token consumption expected to rise significantly [6][15] - **Market Position**: - Positioned as a key player in the AI model market, with a focus on B2B applications and local enterprise clients [17][19] - Expected to capture a significant share of the cloud and API market by 2026 [14] Key Points on Minimax - **Model Capabilities**: - Focus on multi-modal integration, with advancements in text, video, and audio models [30][31] - Competitive edge through rapid model iteration and a strong organizational structure [35][36] - **Market Trends**: - The global generative AI market is projected to grow from $92.9 billion in 2023 to $1.6 trillion by 2030, with a CAGR of 37% [34] - Minimax is positioned to maintain a leading role in this evolving landscape [34][35] - **Strategic Focus**: - Emphasis on maintaining technological leadership and leveraging first-mover advantages in the AI space [35][38] - Plans to enhance user experience through improved model capabilities and integration into various applications [46][52] Additional Insights - **Zhipu's R&D and Talent**: - Strong emphasis on R&D with over 70% of employees in research roles, contributing to a robust talent pool [28] - **Minimax's Competitive Landscape**: - The competition is consolidating among top players, with Minimax aiming to differentiate through innovative model architectures and user-friendly pricing strategies [37][48] - **Future Outlook**: - Both companies are expected to capitalize on the growing demand for AI solutions across various sectors, with a focus on enhancing productivity and user engagement [49][54] Conclusion - The conference highlighted the significant growth potential for both Zhipu and Minimax in the AI industry, driven by advancements in model capabilities and strategic market positioning. The emphasis on R&D, user experience, and multi-modal integration will be crucial for maintaining competitive advantages in a rapidly evolving landscape.

Summary of Conference Call on Hailuo Group and Hailuo Chuangye Company and Industry Overview - The conference call discusses Hailuo Group's acquisition of a 10.61% stake in Hailuo Chuangye, which will make Hailuo Group the largest single shareholder and effectively grant it actual control over Hailuo Chuangye [1][2] - Hailuo Chuangye is transitioning towards a state-owned enterprise structure, with Hailuo Group's increased stake indicating a clear move towards nationalization [2] Key Points and Arguments - Hailuo Group's acquisition is strategic, as it not only increases its control over Hailuo Chuangye but also indirectly enhances its control over Hailuo Cement, in which Hailuo Chuangye holds a 17.8% stake [2][3] - The market undervalues Hailuo Chuangye's asset value, particularly its indirect stake in Hailuo Cement, which is seen as being priced at a significant discount (64% undervalued) compared to the market [3][10] - Hailuo Chuangye's operational efficiency is improving, with increasing free cash flow due to declining capital expenditures [4][17] - The company’s main business, waste incineration, is expected to see growth in capacity and operational efficiency, with a projected increase in waste processing capacity to 5.1 million tons by 2026 [11][12] Financial Metrics and Projections - Hailuo Chuangye's current market capitalization is approximately HKD 234 billion, with a price-to-book (PB) ratio of 0.44 [9] - The expected profit for Hailuo Chuangye in 2025 is projected to be over RMB 800 million, translating to around HKD 900 million [14] - The company anticipates a significant increase in free cash flow, with projections of reaching positive cash flow of HKD 600 to 800 million in 2025 [16][17] Additional Important Insights - Hailuo Group's dividend payments to Hailuo Chuangye were only HKD 400 million in 2024, indicating a need for external financing to support capital expenditures [15] - The increase in metal prices is positively impacting the value of by-products from waste incineration, contributing to revenue growth [12][13] - The overall market sentiment suggests that Hailuo Chuangye's valuation will improve as the market recognizes the true value of its assets and operational improvements [18]

Summary of the Conference Call on China's AI Industry Industry Overview - The Chinese AI industry is transitioning from a "hundred models war" phase to a stage where commercial viability, model innovation, and global expansion are key determinants of success [1][18] - The global AI market is projected to reach $1.4 trillion by 2030, with the B2B application market estimated at $1.1 trillion and the B2C application market at $300 billion [1][18] Key Companies Zhipu AI (智谱) - Zhipu AI is recognized as a leading independent large language model developer, focusing on private deployment and API services for B2B clients [1][31] - The company has established a solid customer base in regulated industries in China, which is expected to provide long-term demand [31] - Revenue growth is projected at a compound annual growth rate (CAGR) of 127% from 2026 to 2030, with profitability expected by 2029 [1][32] MiniMax - MiniMax is another prominent independent large language model developer, with over 70% of its revenue coming from international markets [1][35] - The company has a balanced business model across consumer applications, generative media, and B2B APIs, supporting a projected CAGR of 138% from 2026 to 2030, with profitability also expected by 2029 [1][35] Investment Ratings - Both Zhipu AI and MiniMax have been initiated with an "Overweight" rating, with target prices set at HKD 400 (44% upside) for Zhipu and HKD 700 (36% upside) for MiniMax [1][4] Market Dynamics - The AI market in China is rapidly consolidating, with the number of capable model developers decreasing from over 200 to less than 10 [1][3] - The competitive landscape features a dual-track system: large tech giants with scale advantages and independent innovators like Zhipu and MiniMax [1][3] Financial Projections - By 2030, Zhipu is expected to generate $6.009 billion in revenue and $1.217 billion in adjusted net profit, while MiniMax is projected to achieve $6.078 billion in revenue and $1.475 billion in adjusted net profit [1][39] - Both companies are valued at a 30x P/E ratio based on expected earnings in 2030, reflecting their high growth potential compared to other Chinese internet companies [1][38] Risks - Key risks include export controls and geopolitical tensions, particularly for Zhipu, which has been placed on the U.S. Bureau of Industry and Security's entity list [1][41] - MiniMax faces a lawsuit from major entertainment companies over copyright issues, which could impact its operations in the U.S. [1][42] - Both companies are in a capital-intensive phase, requiring ongoing investment in model training and infrastructure, with potential cash flow challenges until they reach profitability [1][46] Conclusion - Zhipu and MiniMax are positioned as leading players in the rapidly evolving AI landscape, with significant growth potential driven by their innovative models and global strategies [1][25][35]

Summary of Microsoft Corp. Investor Meeting Takeaways Company Overview - **Company**: Microsoft Corp. (MSFT.O) - **Market Cap**: US$2,958,816 million [5] Key Industry Insights Copilot Adoption - **Growth Driver**: Copilot is now the primary growth driver for M365 Commercial, surpassing E5 [2] - **Adoption Progress**: Significant improvements in quality and usage over the last 3-6 months, aided by Agent Mode [2] - **Competitive Advantage**: WorkIQ provides a strong competitive moat due to its deep M365 data indexing, making it hard for external AI applications to replicate [2] Azure Capacity and Growth - **Capacity Constraints**: Azure growth is currently limited by capacity, but there is a clear allocation framework in place [1] - **Future Capacity**: Data Center capacity is expected to double by FY27, with a faster pace of new capacity coming online [3] - **Long-Term Margins**: Azure margins are expected to remain sustainable due to efficiency gains from Maia silicon, software improvements, and in-house LLMs [1][8] Financial Insights Capital Expenditure (CapEx) - **CapEx Trends**: CapEx trajectory is expected to follow demand trends, with a focus on long-term ROI [3] - **Investment Strategy**: The shift to short-lived spending reflects Microsoft's early investment in long-dated CapEx due to visibility from the OpenAI partnership [3] Valuation and Target Price - **Target Price**: The one-year target price for Microsoft is set at US$635, based on a ~30x PE multiple to FY28 EPS of US$22.23 [9] - **Valuation Context**: Current share prices are at decade-low valuations and are trading at a discount to the S&P 500 on forward P/E [1] Risks and Challenges - **Investment Risks**: Potential downside risks include slower-than-expected GenAI adoption, competition in the hyperscale cloud market, and the impact of large acquisitions on earnings [10] - **CapEx Concerns**: Aggressive CapEx investment could hurt long-term growth if GenAI adoption does not meet expectations [10] Strategic Focus Prioritization of Resources - **1P Apps and R&D**: Microsoft is prioritizing investments in first-party applications and R&D, particularly in productivity and security [4][7] - **Long-Term Focus**: The company is focusing on free cash flow and gross profit in the near term while investing in compute capabilities [8] Conclusion - **Investment Recommendation**: The overall sentiment from the meetings supports a constructive long-term view on Microsoft, with a reiteration of a Buy rating due to its strong positioning in AI-driven enterprise software and attractive valuation metrics [1]

Summary of Kingdee (0268.HK) Conference Call Company Overview - **Company**: Kingdee International Software Group Company Limited (0268.HK) - **Industry**: Enterprise Software and Cloud Services Key Points Revenue Guidance and Profitability - Kingdee provided a revenue guidance for 2H25 of Rmb3.76 billion to Rmb3.86 billion, aligning with the analyst's estimate of Rmb3.85 billion [1] - The company expects to achieve profitability in 2025 with a net profit forecast of Rmb60 million to Rmb100 million, a significant recovery from a net loss of Rmb142 million in 2024 [1] - The growth is attributed to increasing cloud-subscription revenues and the commercialization of enterprise AI solutions [1] AI Strategy and Product Development - Kingdee is implementing a "Subscription-first, AI first" strategy, which is expected to expand its AI client base and accelerate the commercialization of customized solutions [2] - The company has launched 30-40 AI agents across various applications, including Chat BI and Quality testing agent, and is integrating upgraded AI models into its Cosmic platform [2] Earnings Revision - Earnings estimates for 2025, 2026, and 2027 have been revised down by 47%, 8%, and 6% respectively due to higher-than-expected operating expenses [3] - Despite the revisions, there is optimism regarding rising subscription revenues and improving operational efficiency [3] Financial Projections - The updated financial projections for 2025E show revenues of Rmb7,044 million, with a net income revision down to Rmb94 million from Rmb178 million [4] - The gross profit margin is expected to remain stable at 66.7% for 2025E, while operating profit margin is projected to be -1.6% [4] Valuation and Price Target - The target price for Kingdee is revised down to HK$18.21 from HK$18.98, implying a P/S ratio of 6.8x for 2026E [9] - The valuation methodology includes a blend of DCF and P/S methodologies, with a WACC of 11.5% and a terminal growth rate of 2% [17] Risks and Challenges - Key risks include weaker-than-expected demand for ERP digitization, slower product development, and rising talent/labor costs [17] Market Position - Kingdee's market capitalization is approximately HK$38.9 billion, with a current share price of HK$11.16, indicating a potential upside of 63.2% to the target price [18] Additional Insights - The company is focusing on improving operational efficiency and scaling up its cloud services, which is expected to enhance free cash flow in the coming years [9] - Kingdee's annual recurring revenue is projected to grow significantly, reaching Rmb5.527 billion by 2026E, which is 63% of total revenue [4] This summary encapsulates the essential insights from the conference call, highlighting Kingdee's strategic direction, financial outlook, and market positioning.

Summary of OmniVision (603501.SS) Conference Call Company Overview - **Company**: OmniVision (603501.SS) - **Market Cap**: Rmb139.4 billion / $20.2 billion - **Enterprise Value**: Rmb134.9 billion / $19.5 billion - **Current Price**: Rmb115.82 - **Target Price**: Rmb138.30 - **Rating**: Downgraded to Neutral from Buy Key Industry Insights - **CIS Market Position**: OmniVision holds a leading position in the global CMOS image sensor (CIS) market, benefiting from a shift towards higher-end CIS products and innovative applications such as AI/AR glasses, drones, and action cameras [1][2] - **Product Diversification**: The company is expanding its product lines beyond traditional markets, including display driver ICs (DDIC), analog ICs, high-speed data transmission SerDes chips, microcontrollers (MCU), and liquid crystal on silicon (LCOS) [2] Financial Performance and Projections - **Revenue Growth**: Projected revenues for 2026 are Rmb39.28 billion, a decrease of 5% from previous estimates due to lower expectations in smartphone, automotive, and surveillance CIS segments [19][20] - **Earnings Forecast**: EPS for 2026 is revised down to Rmb5.12, reflecting a 9% reduction from earlier forecasts [19][20] - **Gross Margin Adjustments**: Gross margins for smartphone CIS are expected to decrease by 0.1 percentage points in 2026 and 0.4 percentage points in 2027 due to price sensitivity and competition [19] Market Challenges - **Smartphone Outlook**: Global smartphone shipment growth is expected to decline by 6% YoY in 2026, influenced by rising memory costs and price sensitivity among consumers [3][18] - **Impact of Memory Costs**: The increase in memory costs is anticipated to affect brand customers' choices regarding camera specifications, leading to uncertainties in demand [1][18] Earnings Revisions - **2026/27E Revenue Cuts**: Revenues for 2026 and 2027 have been reduced by 5% and 8%, respectively, due to headwinds in end demand and rising costs [19][20] - **Operating Expenses**: Operating expenses are expected to rise due to increased R&D spending as the company develops new products for innovative applications [19] Valuation and Risks - **Target P/E Ratio**: The target P/E ratio has been adjusted to 27.0x, down from 33.0x, reflecting slower earnings growth [23] - **Investment Risks**: Key risks include slower product expansion, fluctuations in automotive CIS growth, and potential impacts from trade tensions [30] Conclusion - OmniVision's strategic diversification into new product lines and applications is positive, but the company faces significant challenges from rising memory costs and a declining smartphone market. The downgrade to Neutral reflects a cautious outlook amid these uncertainties [1][24]

Summary of Microsoft Corp. (MSFT) Conference Call Company Overview - **Company**: Microsoft Corp. (MSFT) - **Market Cap**: $3.0 trillion - **Enterprise Value**: $2.9 trillion - **Industry**: Americas Software Key Points and Arguments 1. Maia 200 Inference Accelerator - Microsoft announced Maia 200, an updated custom accelerator for AI inference, on January 26, 2026, indicating significant progress in its internal silicon strategy [1] - Initial benchmarks suggest that Maia's performance is now comparable to competitors like Amazon's Trainium and Google's TPUs, enhancing Microsoft's price/performance for AI compute services [1][2] - Limitations noted include the lack of performance statistics from Maia in full production runs and the need for a robust software ecosystem to support it [1] 2. Diversification of Silicon Footprint - Diversifying Microsoft's silicon footprint is crucial for achieving better gross margins and ROI in AI compute, with expectations that AI compute gross margins will approach those of CPU-based compute over time [2] - Merchant solutions are expected to retain the majority share of AI accelerators due to the rapidly evolving AI model development landscape [2] 3. Financial Projections - Revenue projections for the next four fiscal years are as follows: - FY 2025: $281.7 billion - FY 2026: $328.6 billion - FY 2027: $387.0 billion - FY 2028: $456.3 billion [3] - EBITDA and EBIT are also projected to grow significantly, with EBITDA reaching $280.7 billion by FY 2028 [3] 4. Competitive Landscape - Microsoft is diversifying its GPU supplier base, with AMD's GPU share expected to grow from 5% in 2025 to 8% in 2028 [17] - Groq's LPUs are emerging as competitors, offering significant advantages in AI inference speed and energy efficiency [17] - Microsoft claims that Maia 200 offers 30% better performance per dollar compared to the latest generation hardware in its fleet [17] 5. Performance Metrics - Key performance metrics for Maia 200 compared to competitors: - FP4 TFLOPS: 10,145 (Maia 200) vs. 2,517 (AWS Trainium 3) - FP8 TFLOPS: 5,072 (Maia 200) vs. 4,614 (Google TPU v7) [18] 6. Risks and Valuation - The company maintains a Buy rating with a 12-month price target of $600, based on a 28x P/E multiple of adjusted net income [24] - Key risks include less-than-expected revenue from the OpenAI partnership, longer ramp-up for internal silicon, and potential leadership changes [24] 7. Cash Flow and Financial Health - Free cash flow is projected to increase from $71.6 billion in FY 2025 to $106.1 billion in FY 2028 [14] - The company maintains a strong balance sheet with a net debt/EBITDA ratio of (0.3) [3] Additional Important Insights - The integration of inference software engines is critical for optimizing AI performance and cost, with Microsoft looking for improved industry feedback on this front [25] - Nvidia is expected to maintain its leadership in the accelerator market due to its rapid pace of innovation and significant R&D spending [25] This summary encapsulates the key insights from the conference call regarding Microsoft's strategic initiatives, financial outlook, competitive positioning, and associated risks.

Summary of Knowledge Atlas Technology JSC Ltd (Z.ai) Conference Call Company Overview - **Company Name**: Knowledge Atlas Technology JSC Ltd (Z.ai) - **Industry**: AI Foundation Models - **Market Position**: Leading player in the global AI foundation model market with strong academic roots from Tsinghua University [2][24] Key Points Revenue Growth and Projections - Z.ai is expected to experience rapid revenue growth, forecasting an increase from Rmb769 million (~US$100 million) in 2025 to Rmb4.2 billion (~US$600 million) in 2027, and reaching Rmb30 billion (~US$4.4 billion) by 2030 [3][8] - The company anticipates a favorable mix shift from on-premise deployment (85% of revenue in 1H25) to cloud-based deployment, which is expected to account for 66% of revenue by 2027 and over 80% by 2030 [3][64] Product Development and Market Position - Z.ai's latest model, GLM-5, ranks among the top global AI foundation models, particularly in coding and agentic capabilities, and has seen a price increase of approximately 60% compared to previous models [2][50] - The company has a strong focus on productivity scenarios, which enhances its product-market fit and allows for higher pricing power [27][55] Financial Metrics and Valuation - The company has been initiated with an Overweight rating and a price target of HK$560, implying a 2027 estimated price-to-sales (P/S) ratio of 53x [4][11] - Key financial metrics include projected revenues of Rmb4.2 billion in 2027 and an operating profit breakeven expected by 2029 [4][29] Risks and Challenges - Z.ai faces significant geopolitical risks due to its limited exposure to the US market, as it is on the Entity List, which restricts its ability to operate in the largest AI market [3][28] - The company is also exposed to competition from global players like OpenAI and Google, which have superior resources and broader datasets [27][28] - Supply chain risks are notable, as Z.ai relies heavily on high-performance GPUs, with over 50% of its procurement coming from the top five suppliers [27][89] Market Dynamics - The global foundation model market is projected to grow significantly, with a forecast CAGR of 80.7% from US$10.7 billion in 2024 to US$206.5 billion by 2029 [74] - Z.ai's cloud revenue mix is expected to increase, driven by rising AI cloud adoption in China, despite historical downward trends in cloud gross margins [64][74] Investment Positives - Z.ai's strong foundation model leadership and advanced capabilities position it well for future growth, particularly in the burgeoning Chinese AI market, which is expected to grow at a CAGR of 63.5% from 2024 to 2030 [27][31] - The company's early commercialization efforts and extensive customer base provide a solid foundation for revenue visibility and growth [72][56] Conclusion - Z.ai is positioned as a leading player in the AI foundation model industry with strong growth prospects, although it must navigate significant risks related to geopolitical factors, competition, and supply chain vulnerabilities [24][28][89]

Summary of Conference Call Notes Industry Overview - The report focuses on the Chinese IP retail and toy industry, specifically highlighting companies such as Pop Mart, Miniso, and Bloks, along with their recent performance and market strategies [1][6][7][8]. Key Points Pop Mart - **Sales Growth**: Pop Mart's online sales growth in China accelerated in January, attributed to increased supply on platforms like Tmall and Douyin. January recorded the highest sales volume over the past year [1][11]. - **Product Launches**: Successful new product launches, particularly the Twinkle Twinkle's Crush On You series and Skullpanda's My Little Pony series, positively impacted sales. The Crush On You series maintained a secondary price premium of approximately 40% [1][9]. - **Market Confidence**: The company's share buyback program helped restore market confidence [1]. - **Sales Data**: Combined sales on Tmall and Douyin flagship stores increased over 300% year-on-year in January, compared to a 53% increase in December [9]. Miniso - **Partnerships and Marketing**: Miniso partnered with the 2026 CCTV Spring Festival Gala, launching a co-branded IP "马优优". However, initial sales were muted, with less than 50 units sold by February 8 [1][50]. - **Sales Growth**: US credit card sales growth was approximately 65% in January, down from 75% in December [26][28]. - **IP Development**: The CEO highlighted plans to develop 100 proprietary IPs over the next decade and expand the Miniso Land format globally [50]. Bloks - **Product Launches**: Bloks experienced a slowdown in new product launches in January compared to December, consistent with seasonal trends [1][60]. - **Sales Performance**: Overall sales on Tmall and Douyin were down 10% year-on-year in January, attributed to timing distortions around the Chinese New Year shopping window [10]. Financial Forecasts and Valuations - **Earnings Forecasts**: The earnings forecasts for the three covered IP retailers were updated, with target prices lowered for all [2][55][56]. - **Pop Mart**: Expected low 30% year-on-year growth in topline and bottom line for 2026, with a new target price of HK$298 [57][59]. - **Miniso**: Adjusted target price lowered to US$25/HK$49 per ADR/H-share, with a focus on mid-high teens growth in net profit [56]. - **Bloks**: Target price revised down to HK$63, reflecting a slower growth outlook [60]. Market Sentiment and Consumer Confidence - **US Market**: Consumer confidence in the US declined sharply, while sentiment improved in the Euro Area, Malaysia, and Indonesia. The US credit card sales growth for Pop Mart and Miniso showed signs of deceleration [24][25][47]. - **Geopolitical Tensions**: Issues related to Pokemon and Detective Conan in China were noted amidst geopolitical tensions between China and Japan [49]. Additional Insights - **Secondary Market Trends**: Prices for Labubu's plush toys in the secondary market showed a downward trend, with discounts deepening from low single digits to 20% [9][32]. - **Online Sales Trends**: The online sales run-rate for plush toys in February remained broadly on track compared to January, indicating stable demand [1][9]. This summary encapsulates the key insights and data points from the conference call, providing a comprehensive overview of the current state and outlook of the Chinese IP retail and toy industry.

Summary of Cisco Systems Inc. (CSCO) Earnings Call Company Overview - **Company**: Cisco Systems Inc. (CSCO) - **Market Cap**: $340.8 billion - **Enterprise Value**: $351.5 billion - **Industry**: Networking Equipment and Solutions Key Financial Highlights - **F2Q26 EPS**: $1.04, beating guidance of $1.01-$1.03 and consensus of $1.02 [16][18] - **F2Q26 Revenue**: $15.3 billion, up 10% year-over-year, exceeding guidance of $15.0-$15.2 billion [18] - **Networking Revenue**: $8.3 billion, up 21% year-over-year, beating consensus of $7.7 billion [18] - **Security Revenue**: $2.0 billion, down 4% year-over-year, missing consensus of $2.2 billion [18] - **Gross Margin**: 67.5%, down 120 basis points year-over-year, below guidance of 67.5%-68.5% [18][15] - **F3Q26 Guidance**: Revenue of $15.4-$15.6 billion, Non-GAAP EPS of $1.02-$1.04 [17] Core Business Insights - **Networking Orders**: Grew 18% year-over-year, with significant contributions from campus networking refresh and Wi-Fi 7 upgrades [15] - **AI Infrastructure Orders**: Increased to $2.1 billion from $1.3 billion in F1Q26, with 60% from systems and 40% from optics [15] - **Security Segment Challenges**: Continued decline in legacy products and transition to cloud subscriptions affecting revenue [15] - **New Product Growth**: Newer offerings like Secure Access and AI Defense represent 1/3 of the security segment, expected to drive future growth [15] Strategic Actions and Outlook - **Gross Margin Pressure**: Expected to remain under pressure due to higher commodity costs and business mix [15] - **Memory Price Management**: Strategies include price increases and updated contracts to manage memory costs [15] - **Long-term Growth Framework**: EPS growth of 6-8% expected, with revenue growth of 4-6% [15] Risks and Opportunities - **Upside Risks**: Include hybrid work trends, multi-cloud architecture adoption, and increased edge computing use cases [24] - **Downside Risks**: Competition from lower-cost providers and potential margin degradation due to supply chain issues [24] Investment Thesis - **Market Position**: Cisco is a market leader in networking solutions, facing challenges from lower-cost competitors but maintaining a comprehensive product offering [25] - **Revenue Visibility**: Strong product backlog and recurring revenue provide visibility even in a challenging macro environment [25] Valuation - **Target Price**: $75, based on a P/E ratio of 16x [21]