2026-02-11 15:40Summary of Conference Call Notes Industry Overview - Industry: US Semiconductors and Cloud Capital Expenditure (Capex) - Key Players: Major cloud vendors including META, GOOGL, AMZN, MSFT, and others Key Points and Arguments Capital Expenditure Growth - Capex Projections: - CY26 capex is projected to reach $748 billion, a 56% increase YoY - CY27 capex is projected at $869 billion, a 16% increase YoY - Previous estimates were $141 billion for Q4, now revised to $148 billion, reflecting a 14% QoQ and 66% YoY increase [1][11] Demand and Supply Dynamics - Demand for AI Semiconductors: - Strong demand for data center and AI semiconductors is expected to persist, with compute demand outpacing supply throughout the year [1] - Upcoming launches of advanced models (e.g., Blackwell-trained models) are anticipated to enhance sales and return on investment (ROI) significantly [1][19] Company-Specific Capex Updates - META: Increased CY26 capex outlook to $125 billion from $105 billion [2] - GOOGL: Aggressively raised CY26 capex guidance to $180 billion, nearly 100% increase YoY [2] - AMZN: Forecasting record capex of $200 billion for CY26, up 50% YoY [2] - MSFT: Maintained prior guidance of a 58% YoY growth rate for FY26 capex [2] Financial Metrics and ROI - Capex as a Percentage of Operating Cash Flow: - Expected to rise to 95% in CY26, compared to 73% in CY25 and 52% in CY24 [3][13] - Hyperscalers are expected to maintain free cash flow (FCF) to support these investments [3] Capacity and Contracting - Contracted Capacity: - Most GPU-related expenditures are already contracted for the majority of their useful lives (2-6 years), indicating that current capex is demand-driven rather than speculative [4] - Companies like GOOGL and AMZN are ramping up capex to meet customer demands [4] Depreciation Concerns - Depreciation Schedules: - Concerns have arisen regarding the extension of depreciation schedules for AI infrastructure assets, now commonly set at 4-6 years compared to the historical 3-4 years [23][24] - This change raises questions about the sustainability of reported profits and justifications for heavy IT equipment purchases [24] Future Outlook - AI Infrastructure Investment: - The AI infrastructure investment is expected to have more significant upside potential compared to traditional telecom infrastructure buildouts [14][15] - The overall capex intensity for cloud services is projected to rise towards 50% of total sales in CY26/27E [14] Additional Important Insights - Market Dynamics: - The competitive landscape for AI models is shifting, with new models expected to outperform older generations, particularly with the introduction of Blackwell-based models [19][20] - Investment Recommendations: - Analysts maintain a "Buy" rating on companies like NVDA, AVGO, CRDO, and AMD, while raising price objectives for ALAB, LITE, COHR, and ARM due to their exposure to AI capex growth [1] This summary encapsulates the critical insights from the conference call, highlighting the robust growth in capital expenditures within the semiconductor and cloud sectors, the demand dynamics for AI technologies, and the financial implications for major players in the industry.