AI总结
核心结论 (Core Conclusion)
- 中文：电力是AI发展的核心限制因素。亚马逊凭借存量优势保持领先，而谷歌通过创新策略及清洁能源布局，正在快速缩小差距。
- English: Power is the primary constraint for AI development. Amazon maintains an incumbent advantage, while Google is rapidly closing the gap through innovative strategies and clean energy integration.
结构逻辑 (Structural Logic)
- 电力瓶颈：定义AI扩张的核心制约因素为电力供应。
- 竞争格局：对比四大 hyperscalers（亚马逊、微软、谷歌、Meta）的存量与预测产能。
- 战略差异：亚马逊侧重成本与可靠性（自建为主）；谷歌侧重清洁能源与灵活性（租赁与垂直整合）。
- 未来趋势：行业正探索核能、离网天然气及空间计算等多元化能源方案。
关键信息 (Key Information)
- 数据中心电力：亚马逊约9GW；微软与谷歌各5GW；Meta约4GW。
- 发展速度：预计至2030年，亚马逊将增加最多总量，但谷歌增速最快。
- 关键概念：自建 vs. 租赁模型、离网电源、AI电力消耗、核能协议。
- 主要公司：Amazon (AMZN), Alphabet/Google (GOOGL), Microsoft (MSFT), Meta (META), SpaceX.
问题与回答 (Key Questions)
- 问题：谁能在AI能源竞赛中胜出？
- 回答：短期内亚马逊领先；长期看，拥有清洁能源布局和高效产能扩充策略的谷歌具备强劲竞争力。
正文
An Amazon Web Services data center in Aldie, Virginia.
An Amazon Web Services data center in Aldie, Va. Lexi Critchett/Bloomberg News
Amazon’s AMZN -3.10%decrease; down pointing triangle chief executive has said that power is the “single biggest constraint” in its cloud and artificial-intelligence business. This means investors tracking hyperscalers’ AI race will also want to keep close tabs on how they plan to get their hands on more electricity.
Here, Amazon has an incumbent advantage. It is the world’s largest cloud provider and has been building a lot of data centers over the past two decades. The company’s operating, self-built data centers in the U.S. consume up to roughly 9 gigawatts of power, according to Aterio, a data provider. That is comparable with the generation capacity of North Dakota.
By comparison, Microsoft MSFT -3.46%decrease; down pointing triangle and Alphabet’s GOOGL -0.46%decrease; down pointing triangle Google each have self-built data centers that use up to about 5 gigawatts of power, while Meta Platforms ’ META -2.65%decrease; down pointing triangle data centers have a roughly 4 gigawatt capacity. So far, most of the hyperscalers’ data center capacity is self-built, rather than rented from data-center operators.
So which companies are best positioned to win going forward? Based on estimates from Aterio, which tracks company announcements, utility filings, building permits and satellite data, Amazon is expected to add the most data center and power capacity in the U.S. through 2030. But Google is expected to add capacity at the fastest rate. In fact, including leased capacity from third-party data center owners, Google will have significantly closed its gap with Amazon by 2030, according to Aterio.
Given the constraints on the existing grid and the supply chain, hyperscalers have to pick their priorities: speed, cost, reliability or environmental impact. Thus far, Amazon’s strategy seems to emphasize cost and reliability, while Google appears focused on clean energy.
Amazon has a “really good understanding of building capacity” and has a methodical approach to it, said Sergio Toro, founder of Aterio. Scale and experience come with longstanding relationships with utilities and suppliers of scarce power equipment. It was earlier than peers on certain energy opportunities, announcing the first power purchase agreement with an existing nuclear power plant. Since then, Microsoft, Google and Meta have signed their own deals with mothballed nuclear-power plants or those slated to be shut down.
Amazon plans to build out most of its own capacity, while Google is expected to rely more heavily on leases. Based on Aterio’s data, about a quarter of Google’s expected data center capacity in 2030 is expected to come from leases. Self-built can take longer but is the cheaper option over the long term.
Google stands out for “doing everything it can” to avoid building data centers that rely on fossil fuels, according to Michael Thomas, founder of data firm Cleanview. It is the only hyperscaler with an in-house, renewables-focused developer, having bought Intersect Power earlier this year. The company also seems focused on finding ways to make data centers’ power needs more flexible: It was the first tech giant to start studying how its data centers might use less power during hours when demand is high on the grid.
Also, Google has proved that it can get speed with clean energy. At least three of its planned Texas data centers will be able to skip the long queue to connect to the grid because they are being built next to solar and wind projects, according to a report from Cleanview. In two of these data centers, Google will build the solar and wind capacity through its Intersect Power subsidiary. Texas’ power market rules allow faster grid connection if the data center is co-located with a new source of power. In all three cases, the solar and wind capacity well exceeds that of the data center.
But access to speedy power means accepting near-term trade-offs, even for a green-conscious company. Google will be renting computing capacity from SpaceX, which built out its data centers in record time because it used off-grid, natural-gas-fired power equipment. SpaceX’s subsidiaries were sued earlier this year over air pollution concerns from its use of those turbines.
Across the country, Americans are turning against proposed data centers in their communities, citing concerns about power, water and noise. The Journal tracked the grassroots and legislative push against these large-scale data centers and found over 170 communities with moratoriums against them. Photo: Lexi Critchett/Bloomberg News
Others have more direct plans for off-grid, natural-gas-dependent data centers, most of which use smaller, industrial-scale power equipment. Microsoft, for example, struck a 20-year agreement with Chevron to power its AI data center in Texas with an off-grid, natural-gas-fired power plant. Both Meta and Amazon have plans for such projects, according to data compiled by Cleanview. Amazon hasn’t publicly confirmed its involvement, but its data center in Fayette County, Ohio, is the only planned source of large power demand near a permitted off-grid, natural-gas-power project, according to Cleanview’s Thomas. In most cases, hyperscalers eventually want these data centers to connect to the grid.
All the hyperscalers have also placed bets on novel technology providers that promise to eventually supply cheap, clean, reliable and fast energy. These range from small modular nuclear reactors, advanced geothermal and novel batteries to solar energy beamed in from space. SpaceX and Google are even eyeing data centers in space.
The ultimate winner of the AI race will depend in part on where political winds shift—today’s environment is permissive of natural-gas power, but it might not stay that way forever. Some new energy technologies might reach a breakthrough faster than others. For now, though, Amazon and Google have a clear lead.
Appeared in the June 25, 2026, print edition as ‘Amazon and Google Take Lead in AI Power’.