$1.4 trillion. This is the total price of the Stargate computing power layout presented to investors by OpenAI CEO Sam Altman at the end of 2025. Fourteen months later, this figure was slashed to $600 billion.

According to a report by The Information on March 16, OpenAI has significantly restructured the Stargate computing infrastructure project, abandoning plans for self-built data centers and fully shifting to renting computing power from cloud service providers such as Microsoft Azure, Oracle, and Amazon AWS. Stargate has been divided into three functional teams, managed uniformly by former Intel Chief Technology and AI Officer Sachin Katti.

The direct reason for the shift is not complicated. Stargate was publicly launched at the White House in January 2025, announcing a joint venture with SoftBank and Oracle to build large data centers, with an initial investment of $100 billion and a total investment of $500 billion over four years. However, more than a year after the project was launched, not a single employee was hired, and no substantial development of a data center took place. According to CNBC, lenders were unwilling to provide billions of dollars in construction financing for a company still facing significant operational losses. Earlier this month, OpenAI also withdrew from negotiations to expand Oracle Stargate in Abilene, Texas.

More than a year, zero employees, zero construction. The "self-built" path of Stargate never really took off.

According to detailed breakdown data shown in investor materials, Altman's $1.4 trillion total commitment is distributed among seven suppliers. According to a venture capital analyst Tomasz Tunguz's analysis of the investor materials, Broadcom accounts for $350 billion, Oracle $300 billion, Microsoft $250 billion, NVIDIA $100 billion, AMD $90 billion, and AWS and CoreWeave a combined $60 billion.

In February 2026, CNBC reported that this figure was reset to approximately $600 billion (by 2030), a reduction of 57%. The same report provided a slightly different but directionally consistent figure, stating that OpenAI expects to spend $665 billion on cloud servers by 2030.

$600 billion is still a number that requires anchoring to comprehend. According to internal forecasts from OpenAI, the company’s revenue target for 2030 is $280 billion, which implies a five-year cumulative spending to revenue ratio of about 2:1. Meanwhile, internal financial data cited by ainvest indicates that the company's expected loss for 2026 is $14 billion, with a gross margin reported by multiple media outlets at only 33% (Note: gross margin reflects the product’s earning ability, while net loss is the final result after accounting for all costs such as R&D and management, and both can coexist).

Putting OpenAI’s spending targets into the broader context of the Big Tech computing power arms race clarifies the proportional relationships.

According to financial reports and public guidance from each company, Amazon plans capital expenditures of $200 billion in 2026, Alphabet $180 billion, Meta $125 billion, and Microsoft approximately $120 billion. The expenditures of these four companies have generally multiplied by 2-3 times within two years, totaling over $650 billion, with about three-quarters going towards AI infrastructure.

OpenAI's $600 billion is a five-year cumulative target, annualized at about $120 billion, comparable to Microsoft's annual capital expenditure. The difference lies in Microsoft's annual revenue of over $240 billion, while OpenAI’s annualized revenue just hits $25 billion and does not expect to achieve positive cash flow before 2030.

The restructuring of Stargate signifies more than just a change in budget figures; the adjustments in organizational structure reveal a deeper directional shift.

The restructured Stargate is divided into three lines. The Epic Business Partnerships team is led by OpenAI veteran and former Deloitte manager Peter Hoeschele, managing cloud contracts with Microsoft, Oracle, and Amazon, as well as deals with chip manufacturers, including multi-year contracts with AMD (utilizing up to 6 gigawatts of chips, costing up to 10% of AMD common stock) and agreements with chip startup Cerebras Systems.

The Technical Engineering and Design team is co-led by former Meta and Google engineer Chris Malone and former Microsoft engineering director Adrian Caulfield, responsible for redesigning the AI server clusters used by OpenAI. The Physical Facilities Operations team is led by former Google data center director Nick Saddock, taking over from Keith Heyde, who left a few weeks earlier.

The semiconductor team led by former Google chip executive Richard Ho is not under Katti's jurisdiction and reports directly to OpenAI President Greg Brockman. This team is collaborating with Broadcom to develop in-house chips, with OpenAI hoping these chips will ultimately reduce the inference costs of running products like ChatGPT.

The name "Stargate" still exists, but what it represents has completely changed. In January 2025, it was a joint venture project with SoftBank and Oracle to build data centers. By March 2026, it has become OpenAI's broad strategy for bringing gigawatt-level server capacity online. The focus has shifted from "I want to build my own power plant" to "I want to sign the best lease." The total planned capacity of all sites remains close to 7 gigawatts, with the total investment over three years still exceeding $400 billion. OpenAI is shifting its computing direction towards NVIDIA's Vera Rubin platform, aiming to achieve the first gigawatt-level capacity online in the second half of 2026.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。