International Business Machines (IBM)

The sleeper story is that IBM's own management is already discounting the two engines behind this quarter's 6% beat: distributed infrastructure decelerated from a 'best quarter on record' 37% to 32%, and Jim Kavanaugh said outright, 'I don't expect double digit performance in TP to keep going', right after TP posted +10%. The obvious headline, software accelerating to 9% with a raised 7%-8% guide, is genuine, but a quarter of that upside is just Q2's slipped deals finally closing, and when asked point-blank whether the long-term double-digit software target still holds, management punted: 'We'll spend a lot more time on this question here in January', a retreat from last quarter's 'complete conviction'. The binding constraint isn't demand, it's conversion timing on big ELA-linked mainframe deals, which management now admits can swing software growth by multiple points depending on when a handful of CFOs sign paper. My view changes if Q4 TP actually normalizes down as guided and distributed infrastructure stabilizes rather than keeps decelerating; if instead TP stays elevated or infra keeps sliding, the 'timing not destruction' thesis from Q2 starts to look like a one-quarter sugar high dressed up as vindication.