Baker Hughes (BKR)

The sleeper story here isn't the record IET orders or the AI power headline everyone will write about tomorrow; it's that 33% of GTE/CTS backlog now converts in 2027 versus 46% of comparable backlog a year ago, meaning the record RPO ($39.4B, +23% YoY) is becoming a slower-burning asset, not a near-term accelerant, exactly as management conceded: 'we expect more modest growth in 2027 with a greater share of revenue converting beyond next year'. The obvious headline (raised Horizon 2 target to >$50B, data center orders up ~500% YoY) is real demand, but the binding constraint isn't customer appetite, it's capacity and conversion timing, reinforced by management's own words that 'the key constraint is not demand for compute capacity, but how quickly customers can secure reliable power'. Layer on a leverage spike to 2.5x with a quietly softened deleveraging target and a Chart synergy program where only $3M of $70M 'executed' savings has actually hit the P&L, and this reads like a quarter of strong optics sitting on top of slower-than-headline underlying momentum. My view changes if next quarter shows realized Chart synergies jumping meaningfully above the $3M run-rate and IET margin re-accelerating back above guidance midpoint; absent that, I'd fade the 'record backlog' bull case into 2027.