“Data Centre Seven（DC7）”包括 Comfort Systems、Emcor、Vertiv、Quanta、Caterpillar、Generac 和 Cummins，过去十二个月涨幅分别为 +328%、+107%、+246%、+130%、+175%、+122% 与 +117%。作者将2025年营收以100为基准并对2030年做共识预测，结果显示历史增速与预期增速都维持高位。这些公司未来市盈率也大幅抬升；Comfort 与 Vertiv 目前约为48倍与42倍，几乎是去年的两倍。
Caterpillar 在其中最不对称。其发电机业务贡献了第一季 142 亿美元中约40%，约56.8亿美元，且在剔除库存去化因素后增长约32%。在能量销售加快与订单积压加深后，公司将2024至2030年的目标提高到6%到9%的年复合增长率；而 Morgan Stanley 分析师 Angel Castillo 对2030年能源业务预期年增12%，并将评级从低配调至均衡。Caterpillar 目前约以36倍远期市盈率交易，意味着估值已隐含高于指导值的增长预期；作者认为股价可能已跑在公司可持续兑现能力之上，而且若 AI 热潮降温，冲击将向大科技之外的工业生态外溢。
First-quarter U.S. earnings remain broadly strong. If the labor market stays stable and oil stays around US$106 per barrel, far below the cited US$150 ceiling, the case for continued risk-on performance is compelling. The piece argues that investors who only bought Nvidia missed the real AI beta; the deeper exposure is in firms that build and operate AI data centres, especially construction, power, cooling and generator operators.
The “Data Centre Seven” (DC7) cluster is identified as Comfort Systems, Emcor, Vertiv, Quanta, Caterpillar, Generac and Cummins, with one-year total returns of +328%, +107%, +246%, +130%, +175%, +122% and +117% respectively. Re-basing 2025 revenue to 100 and extending consensus projections to 2030 shows both historical and expected growth remain elevated. Forward valuations have also re-rated sharply: Comfort and Vertiv are now around 48x and 42x earnings, nearly double a year ago.
Caterpillar is the standout outlier. Its generator segment contributes about 40% of Q1 revenue, roughly US$5.68bn of the company’s US$14.2bn, and recorded 22% growth, or about 32% after adjusting for inventory destocking. After faster energy sales and a deepening backlog, management raised energy growth guidance to 6%–9% CAGR through 2030, while Morgan Stanley’s Angel Castillo models 12% annual growth to 2030 and upgraded the stock from underweight to equal weight. Caterpillar now trades near 36x forward earnings, implying a growth curve above management guidance; the article notes the shares may already be ahead of sustainable execution, and a broader AI slowdown could transmit shocks beyond large-cap tech into this industrial chain.