At Amazon, austerity is part of its corporate DNA, setting it apart from other major tech companies in Silicon Valley, where luxuries and perks are common. At the Seattle headquarters, employees appreciate simple details like free bananas, a symbol of a culture focused on efficiency and cost control. However, this philosophy does not prevent the company from occasionally making colossal investments. Currently, Amazon is undergoing one of its largest spending phases, dubbed “capexapalooza” by JPMorgan Chase. This year, Amazon plans to invest a record $200 billion in fixed assets, partly financed through bond issuance. The main goal is to strengthen Amazon Web Services (AWS), its cloud computing division, to lead in AI infrastructure, from data centers to energy supply. Additionally, the company has announced a potential investment of up to $50 billion in OpenAI, significantly surpassing Microsoft's commitment to the ChatGPT creator since 2019. Amazon justifies this spending by the growing demand: AWS recorded its highest growth rate in over three years in the last quarter. Andy Jassy, Amazon's CEO, claims the company sells cloud capacity as fast as it can build it, despite the $250 billion invested in the last three years. However, competition is fierce. Although AWS was a pioneer in cloud and AI, since the launch of ChatGPT in 2022, Microsoft's Azure and Google Cloud have grown faster, reducing Amazon's advantage. The focus on AI data centers and collaboration with OpenAI aims to regain ground against these rivals. Amazon has experience in making big bets. Jeff Bezos, its founder, advocated for a strategy of seeking “home runs” in business, which have sometimes yielded great results, such as the increased investment in logistics and AWS in 2016-2017, which improved profit margins. However, during the pandemic, excessive warehouse expansion proved unprofitable, forcing cost cuts in 2023, just as Microsoft ramped up its AI investment. AWS has competitive advantages, such as a large customer base and partnerships with leading labs like OpenAI and Anthropic. AI model developers are heavy consumers of cloud resources, and AWS, with about a third of the global market, could benefit if more companies adopt AI. However, the realization of these investments will take at least 18 months, during which demand could increase even more with the evolution of AI agents. Collaboration with OpenAI and Anthropic allows AWS to offer access to leading market models, in addition to its own Nova model and compatibility with Nvidia chips and its internal alternative, Trainium. Microsoft and Google also offer variety, but their core businesses—Office and search engine, respectively—are more profitable than cloud, which could limit resource prioritization for their cloud customers. At Amazon, AWS is the profit engine, benefiting its business users. On Wall Street, the massive investment strategy generates mixed opinions. Although Amazon's stock has outperformed Microsoft's this year, it has fallen 8% due to doubts about future profitability. Amortization costs will rise before new infrastructures generate revenue, and there are concerns about the impact of high AI costs on AWS margins in the long term. Another challenge is the rise of AI agents capable of making purchases for users, which could affect Amazon's core business. However, the company has managed to curb initiatives like OpenAI's “Instant Checkout,” and its assistant Rufus has generated $12 billion in additional annualized sales. While competitors like Google advance in e-commerce, Amazon maintains its leadership. This “capexapalooza” represents a risky bet, but not necessarily a strategic mistake.