辉达（Nvidia）正将其资产负债表转化为战略武器，透过为客户提供融资担保来推动AI基础设施的扩张。该公司近期与六家华尔街金融机构达成框架协议，为客户提供约五千亿美元的融资，辉达担保其中最高四分之一的交易额；同时为OpenAI在俄亥俄州的新数据中心提供高达一千零五十亿美元的后盾支持。穆迪目前仍维持其信用评级，认为这些担保尚未削弱辉达「优异的财务状况」。（关键数字：500, 105）
执行长黄仁勋主张辉达的晶片已被广泛使用，足以构成一种新的资产类别，金融机构可以放心以此作为贷款担保。然而，这一新兴资产类别缺乏信用分析师所依赖的长期历史数据。当前极端的供不应求局面——SpaceX甚至声称不到一年便能透过出租AI设施收回数据中心的全部建设成本——掩盖了供需逆转时可能出现的风险：一旦市场低迷，违约客户的闲置产能未必能迅速找到接手者，更高效的新一代晶片也可能令耗电量大的旧款晶片失去运行价值。（关键数字：100, a year）
除了信贷工具的策略运用，辉达正积极拓展新客户群，包括全球快速增长的新型云端公司（neocloud）、主权客户及企业客户，协助它们建设自有AI数据中心，而非依赖美国大型云端平台。辉达计划从这些新数据中心的营收中抽成，实质上以「轻资产伪云端」模式深入云端市场。再加上开源AI模型的发布及针对医疗等垂直市场的完整AI解决方案，辉达正以资产负债表为后盾，全力推进AI热潮的下一阶段。
Nvidia is strategically deploying its balance sheet to fuel the AI infrastructure boom, extending financing guarantees to customers while limiting its own exposure. A landmark framework agreement with six Wall Street firms will channel roughly $500 billion to Nvidia's customers, with Nvidia guaranteeing up to a quarter of certain transactions, and a separate $105 billion backstop supports a new OpenAI data centre in Ohio. Moody's has affirmed the company's credit rating, stating that these commitments have not weakened Nvidia's excellent financial profile.
CEO Jensen Huang contends that Nvidia's chips are now so ubiquitous that they constitute a new asset class suitable for collateralised lending, yet this asset class lacks the deep historical track record that credit analysts typically require. The current severe supply shortage—exemplified by SpaceX's claim that it can recoup the full cost of its AI data centres in under a year through rental income—creates conditions that may mask downside risks. Should supply and demand rebalance, defaulted capacity may find few willing takers, and newer, more efficient chips could render older, power-hungry models like the A100 economically unviable despite full depreciation.
Beyond credit strategies, Nvidia is cultivating a new customer base of neoclouds, sovereign entities, and enterprises worldwide that prefer to operate their own AI data centres rather than rely on US hyperscalers. By taking a revenue share from these facilities, Nvidia is positioning itself as what Stifel analysts call an 'asset-light pseudo-cloud,' extending deeper into the cloud market without accumulating heavy infrastructure assets. Coupled with the release of its latest open-source AI model and the development of turnkey AI solutions for vertical markets such as healthcare, Nvidia is leveraging its financial strength to power the next phase of the global AI expansion.