Exchange Ranking
Top Exchanges
24h Volume Ranking
Popularity Ranking
Exchange BTC Balance
Proof of Reserves
Decentralized Exchanges
Funding Rate
Funding Heatmap
Liquidation Data
Max Pain
Long/Short Ratio
Whale L/S Ratio
Binance/Okex/Huobi L/S
Bitfinex Margin L/S
ETF Tracker
Solana ETF
XRP ETF
Hong Kong ETF
Bitcoin Treasuries
Crypto Reversal
Ethereum Reserves
HyperLiquid Wallet Analysis
Hyperliquid Whale Watch
Large Transactions
On-chain Movement
Bitcoin ROI
Stablecoin Market Cap
Options Analysis
News
Articles
Economic Calendar
Features
Wallet
Contract Calculator
Security
Collections
Watchlist
FollowingGoogle has invested heavily in nuclear power production capacity, and the AI race has quietly evolved into a "power plant race"
On Tuesday, Google announced that it has received 3590 megawatts of electricity from Constellation Energy. This capacity, equivalent to 3.6 gigawatts, is connected to the PJM Interconnected Grid, the largest power grid in the United States. As artificial intelligence data centers consume huge amounts of energy every hour, this large-scale power guarantee has become an inevitable requirement for the development of the industry.
It is reported that about a quarter of the total power supply will come from new nuclear power plants. Constlerson will invest more than $4.3 billion to improve its power generation capacity. Among them, the most representative part is nuclear energy cooperation: Google signed a 20-year agreement with Constlerson to obtain 890 megawatts of nuclear power from multiple facilities. To this end, relevant reactors will be upgraded to generate additional electricity, and new production capacity is expected to be put into operation one after another in 2028. Within PJM jurisdiction, the company plans to deploy a total of 11 nuclear generators in Illinois, Pennsylvania and New Jersey.
In addition, an additional 2700 MW of electricity comes from another 15-year contract, whose specific source of power has not yet been clarified. This has become a common trend among large technology companies. Microsoft has previously supported Constlerson in restarting the Three Mile Island nuclear power plant in Pennsylvania to power its data center, while Google has also reached an agreement with NextEra Energy to restore a nuclear power plant in Iowa.
The artificial intelligence race is quietly turning into a power plant race, because expensive GPUs will be useless if power is interrupted.
PJM电网面临扩容压力，巨头寻求自主供能方案
交易公布后，康斯特勒森股价周二上午盘中飙升逾13%，报收于304.04美元。这家总部位于马里兰州的企业是美国最大的核电站运营商，并在今年早些时候收购Calpine后，拥有了天然气电站和地热资产。
谷歌全球能源与电力主管阿曼达·彼得森·科里奥表示：“公司希望通过投资于更清洁、更可靠的供应来推动增长。”然而，在企业平滑的宣传辞令之下，压力正在积聚——数据中心的上线速度已快于电网的处理能力。
在PJM电网内部，自2024年以来，容量价格已上涨超过十一倍，数据中心扩张在推高价格方面发挥了重要作用。此外，电网还面临预期的能源短缺。管理层提出了一项非传统的解决方案：巨型数据中心将被要求自备电力，或在高需求时段面临远程断网的风险。
谷歌与康斯特勒森的这笔交易，部分正是对这一提案的反应。两家公司并未依赖公用事业公司在其他地方寻找额外容量，而是直接将新增需求与增加供给相连接。双方还制定了一套可推广的框架，涵盖新发电、电池储能以及需求响应机制，即在电网承压时，让大型客户减少或调整用电需求。
基础设施蓝图与土地保护争议并存
Google CEO Joe Domingos said the agreement provides a blueprint for technology company energy suppliers to invest in infrastructure expansion in the context of digital economic growth. However, the success of this concept depends on whether actual production capacity can actually be connected to the grid.
Meanwhile, the environmental group The Wilderness Society called on Monday for a moratorium on the construction of AI data centers on federal land, citing their huge consumption of electricity, water resources, and the impact on wildlife habitats and public access. The move follows the Trump administration's review of permits for projects under the Bureau of Land Management (BLM). Donald Trump has been advocating accelerating the construction of AI infrastructure and expanding the use of public land in industrial areas such as mining and oil and gas extraction. However, as data centers occupy a large amount of land, water and electricity, local residents are increasingly opposed to the industry.
The Wilderness Association hopes that policymakers will keep federal land focused on conservation and recreational uses when considering these large-scale AI facilities.
