哪怕投资者砸崩股价，2000亿美元也必须“烧下去”，亚马逊要“重整旗鼓”

Core Insights - Amazon is initiating its largest capital expenditure plan in history, aiming to regain momentum in the AI sector and defend its leading position in cloud computing against strong competition from Microsoft and Google [1][4] - CEO Andy Jassy announced that the company's capital expenditure will rise to $200 billion this year, surpassing the spending levels of Google and Microsoft, with approximately 75% allocated to its cloud computing division, AWS [1][4] - Despite the significant investment, investor concerns about the slow conversion of spending into returns have led to a more than 20% decline in Amazon's stock price since its peak last November [1] Capital Expenditure Plan - Amazon plans to invest $200 billion, primarily focused on expanding AWS, while total spending by Microsoft, Google, and Oracle is expected to approach $400 billion this year [4] - The company aims to increase data center capacity significantly, adding nearly 4 gigawatts by 2025, equivalent to the annual energy consumption of over 3.2 million U.S. households, with plans to double this capacity by 2027 [4] - Jassy has restructured the organization to unify chip, model, and advanced research teams under a single leadership structure, emphasizing a customer-centric and agile operational approach [4] Competitive Anxiety - Despite generating nearly $130 billion in sales last year and contributing over 60% of Amazon's total profit, AWS faces predictions that Microsoft's cloud business will surpass AWS within the next three years due to the surge in demand for AI-driven cloud services [5] - Internal concerns have emerged regarding AWS's ability to secure enterprise AI contracts, with employees expressing that the company was unprepared for the rapid developments in the market [5] Strategic Partnerships and Contracts - Microsoft secured an exclusive cloud computing contract with OpenAI early on, while Amazon only signed a $38 billion agreement after OpenAI allowed for company restructuring, which pales in comparison to Microsoft's $250 billion contract [6] - Amazon's investment of $8 billion in Anthropic for data center construction came after Google had already supported the startup, indicating a reactive rather than proactive approach [6] Chip Strategy and Performance - To reduce reliance on Nvidia products and improve profit margins, Amazon is promoting its self-developed chips, with projected annual revenues exceeding $10 billion from Graviton and Trainium chips [7] - Amazon is negotiating to join OpenAI's latest funding round to ensure the adoption of its semiconductor products, although competition remains fierce, with Google successfully attracting Anthropic to purchase its TPUs [7] AI Model Development - Amazon is investing in its AI model named "Nova," positioned as a low-cost alternative to competitors, but independent benchmarks show that Nova lags behind models from OpenAI, Google, Meta, and Anthropic [8] - Some AWS employees have informally referred to Nova as "Amazon Basics," reflecting concerns about its performance, while many engineers prefer using Anthropic's Claude model over Nova [8] - The pressure to regain ground in the AI competition is palpable among employees, with fears of stagnation and decline echoing sentiments expressed by founder Jeff Bezos in 2018 [8]