A股算力芯片概念午后拉升，消息面上，OpenAI讨论扩大云服务协议以纳入亚马逊Tranium芯片。此 外，英伟达2026财年第四季度财报，营收、净利润及下一季度指引全面超越市场预期，提振了市场信 心。 东莞证券指出，从海内外已披露业绩或业绩预告的企业情况看，半导体行业整体景气度上行但细分领域 分化，其中AI对算力芯片、存储芯片、晶圆代工等细分领域的需求端拉动作用明显，而非AI相关细分 则呈温和复苏态势，部分细分领域如消费类电子，因AI需求挤占存储等资源导致成本上升，景气整体 承压。展望后续，建议继续关注AI带来的高景气细分领域的投资机会，如算力、存力、先进封装、先 进制程晶圆代工、半导体设备与材料等。 截至2026年2月26日 13:45，上证科创板100指数(000698)强势上涨，成分股安路科技上涨13.11%，华曙 高科上涨13.05%，华丰科技上涨11.26%，云天励飞，信科移动等个股跟涨。科创100ETF鹏华(588220) 上涨1.22%，最新价报1.58元。 科创100ETF鹏华(588220)，场外联接(A：019861；C：019862；I：022845)。 MACD金叉信号形成，这些股涨势 ...

Core Insights - Nvidia's CEO Jensen Huang announced at the GTC conference that the company has up to $500 billion in AI GPU infrastructure orders for the years 2025 to 2026, driven by the Blackwell architecture and the upcoming Rubin architecture [1][3][4] - Huang's statement has sparked significant interest among investors, as it suggests a strong growth outlook for Nvidia amidst ongoing debates about an "AI bubble" in the market [1][2][6] Financial Performance - Nvidia's revenue has grown nearly 600% over the past four years, indicating strong market demand for AI chips [2] - Analysts expect Nvidia's Q3 earnings per share to be $1.25, with revenue around $54.9 billion, reflecting a year-over-year growth of 56% [7] - The anticipated revenue for Nvidia in 2026 is projected to be approximately $286.7 billion, significantly higher than previous Wall Street estimates [7][6] Market Dynamics - The demand for AI computing power is surging globally, supported by substantial investments from the U.S. government and major tech companies [2][4] - Nvidia's market share in the AI GPU sector exceeds 90%, but competition is increasing from companies developing custom AI chips [11][12] - Major clients like Amazon and Google are increasingly promoting their own AI ASIC chips, which could impact Nvidia's market position [11] Strategic Partnerships - Nvidia has made significant investments, including a potential $100 billion investment in OpenAI, which will purchase millions of Nvidia GPUs [10] - The company has also invested $5 billion in Intel to enhance collaboration between their CPU and GPU technologies [10] - Nvidia's recent $1 billion investment in Nokia aims to integrate its AI GPU infrastructure into Nokia's cellular network hardware [10] Future Outlook - Huang expressed confidence in achieving the $500 billion revenue target, indicating a "clear visibility" of this growth [8] - Analysts believe that the ongoing AI capital expenditure trend reflects a persistent demand for AI computing resources, despite concerns about potential overinvestment [10][12] - The potential for the Chinese market remains uncertain, with Nvidia's previous AI chip sales being restricted, but analysts see it as a future opportunity [12]

Group 1 - TSMC has further expanded its dominance in the semiconductor foundry sector, with Q3 revenue reaching $33.1 billion, driven by strong demand for 3nm processes and high utilization rates of 4/5nm processes [1] - Apple is the main driver behind TSMC's increased 3nm production, while Nvidia and AMD continue to push high demand for 4nm and 5nm chips, keeping production capacity fully loaded [1] - Major cloud computing companies, including Google's TPU, Amazon AWS's Tranium chip, and Meta's MTIA accelerator, are also increasing demand for TSMC's services [1] Group 2 - Intel expects its foundry customers' wafer commitments to begin mass production in 2026, with major clients anticipated to ramp up production between 2026 and 2027 [2] - Intel has adjusted its foundry strategy to be customer commitment-oriented rather than speculative capacity building, ensuring capacity expansion is directly linked to confirmed demand [2] - Samsung's advanced process utilization and wafer consumption increased in Q2 2025, with expectations for this trend to continue, driven by smartphone chips based on 2nm technology [2] - The future of Samsung's advanced nodes largely depends on the success of its 2nm chips, with collaborations, particularly with Tesla, being crucial for attracting more customers and securing additional orders [2]