How much would the downside scenario reduce earnings?

  • J.P. Morgan modeled 11 electrical equipment and diversified industrial companies with relatively high exposure to data centers or electric utilities, focusing on the main impact in 2029–2030. In the downside scenario, earnings per share would average 26% and 38% below market expectations in the respective years.
  • The differences between companies are substantial: in 2030, scenario earnings per share for FPS, Vertiv and Eaton would be 77%, 64% and 57% below market expectations, respectively. The gaps for Regal Rexnord, Hubbell, Johnson Controls, Trane, SPX Technologies and Carrier would all be no more than 25%.
  • Even so, under the bank’s scenario, the group’s average compound annual earnings per share growth rate for 2026–2030 would remain positive, supported by companies with a higher share of business outside data centers and strong earnings growth in 2027–2028.

How do the downside assumptions flow through to earnings?

  • The model assumes organic revenue growth related to data centers slows sharply in 2028 from 2027 and declines substantially in 2029–2030. The bank assumes revenue from related electrical and HVAC equipment falls by roughly 30% year over year in each of those two years.
  • The 2027 estimates were left unchanged because order backlogs and announced hyperscaler capital spending plans have locked in sales for 2027 and part of 2028. Most equipment order backlogs typically cover less than 12 months, so revenue could fall relatively quickly once the downturn begins.
  • The assumed decline in electric utility revenue is shallower: its rise is more gradual, and some companies have large order backlogs or higher service revenue. The model also assumes no market share changes among peers, operating leverage of approximately 30% during the downturn, and unchanged base forecasts for other businesses in 2028–2030.

How expensive do valuations look after the earnings cuts?

  • Using the share prices in the report, the 11 companies’ average price-to-earnings ratios under the downside scenario are approximately 20 times for 2029 and 22.5 times for 2030. The latter is higher than the 2030 average of 16.5 times for other electrical equipment and diversified industrial companies unrelated to AI.
  • The comparison reflects different points in the cycle: J.P. Morgan believes many AI-related companies’ scenario earnings in 2030 could be at a trough, while other companies’ earnings are expected to reach new highs in 2030. The economic impact of a decline in AI investment could also change earnings expectations for the latter.
  • Differences between individual stocks are also pronounced: under the 2030 scenario, Regal Rexnord trades at 10.2 times earnings and Carrier at 15.2 times, while Vertiv and Eaton trade at 35.9 and 34.6 times, respectively.

Why does J.P. Morgan favor these three companies relatively?

  • GE Vernova’s exposure to electric utilities is much greater than its data center exposure, and it has a large order backlog: remaining performance obligations in fiscal 2026 were approximately 3.4 times fiscal 2027 sales. The bank believes it could have the highest compound annual earnings per share growth rate in the sample for 2026–2030 under the downside scenario.
  • Trane benefits from diversified end markets and higher service revenue. The bank expects its scenario earnings growth to be among the strongest as well, with the premium of its 2030 scenario price-to-earnings ratio over the group median narrowing from 2027.
  • nVent ranks among the top three in the bank’s fundamentals screen. Its liquid cooling business could increase equipment value per megawatt of data center capacity, partly cushioning a decline in capital spending. The constraint is that its 2030 scenario earnings per share would still be 50% below market expectations, while its price-to-earnings ratio would be 25.4 times, above the sample median of 21.4 times.