欧洲AI算力缺口:仅占全球5%,2030年数据中心缺口或达14GW

Rohan Paul · @rohanpaul_ai · X·2026-09-11 15:09·36分钟前
AI 导读

据Mario Draghi引用的ECB情景分析,AI普及可为欧洲全要素生产率年增速贡献0.3至0.4个百分点,而该指标自2022年以来一直接近零。但欧洲AI算力严重不足,仅占全球不到5%,美国占75%,其数据中心缺口或从当前约3GW增至2030年的14GW。Draghi认为数据是欧洲最可控的资产,包括公共卫生和机器可读工业记录,但美国云厂商的所有权给欧洲敏感系统带来法律与政治风险。

Rohan Paul@rohanpaul_ai
39AI 编辑部评分,满分 100

欧洲AI算力缺口:仅占全球5%,2030年数据中心缺口或达14GW

2026-09-11 15:09· 36分钟前
AI 导读

据Mario Draghi引用的ECB情景分析,AI普及可为欧洲全要素生产率年增速贡献0.3至0.4个百分点,而该指标自2022年以来一直接近零。但欧洲AI算力严重不足,仅占全球不到5%,美国占75%,其数据中心缺口或从当前约3GW增至2030年的14GW。Draghi认为数据是欧洲最可控的资产,包括公共卫生和机器可读工业记录,但美国云厂商的所有权给欧洲敏感系统带来法律与政治风险。

ECB scenarios cited by Mario Draghi shows AI adoption could add 0.3% to 0.4% points to annual total-factor productivity growth.

That measure has remained near zero since 2022, so the projected gain would materially change Europe's growth path.

But, unlike the US, Europe is not building many data centres; it has far too little computing capacity for its AI demand. The EU holds less than 5% of global AI compute, vs 75% for US, and its data-centre shortfall could grow from about 3GW today to 14GW by 2030.

Draghi identifies data as Europe's clearest controllable asset, including public health and machine-readable industrial records.

But, Europe increasingly wants control over both its data and the machines processing it. US companies can provide local cloud capacity, but their US ownership creates legal and political risks for sensitive European systems.

来源:Rohan Paul· x.com