IBM's Surprise Q2 Miss Triggers Its Worst Day Since 1987

The Surprise Release

IBM doesn't pre-announce. The company was scheduled to report Q2 on July 22, yet on the morning of July 14 it broke from that calendar and pushed out preliminary numbers — the corporate tell of a management team that couldn't wait for the room to be ready. They were bad. Revenue came in at $17.2 billion, up just 1% and roughly $660 million short of the ~$17.86 billion consensus. Operating (non-GAAP) EPS of $2.93 missed the $3.01 estimate, with GAAP EPS at $2.27.

The damage was concentrated exactly where IBM is supposed to be strongest. Infrastructure revenue fell 7% on a shortfall in mainframe Z systems and the high-margin Transaction Processing software that rides on top of them. CEO Arvind Krishna didn't reach for spin — he told investors the company 'faltered' and that 'numerous large deals failed to close', pinning part of the miss on clients who redirected late-June capital toward servers, storage, and memory to lock in supply ahead of expected price hikes. Whatever the framing, the market's verdict was blunt: shares tore toward their worst single session since the October 1987 crash.

The Perp Was Early

HIPERWIRE flagged the IBM HIP-3 perp yesterday, when it had slid to around $247 on a book turning over only a few million dollars a day. Cash IBM was still sitting near $288.50 at the time, so the perp was marked roughly 14% below spot with no headline attached — the textbook look of a thin-book dislocation, the kind of gap that usually snaps back toward cash. We framed it as noise to be settled at the July 22 print.

It settled early, and it settled the other way. The surprise release didn't drag the perp back up to cash; it dragged cash down through the perp. IBM shares fell roughly 22–23% into the low $220s, and the perp — now $219.90, down 24.97% over 23 hours — has essentially converged with spot. The 'dislocation' was never mispricing. On a thin HIP-3 book, a motivated seller had already walked price to almost exactly where the fundamentals were about to land. That's the live lesson for anyone trading these low-liquidity single-stock perps: a gap below spot is not automatically a mean-reversion trade. Sometimes the thin book gets there first.

Where the Bears Draw the Line

The sell-side had already started leaning bearish before the numbers even hit. HSBC's Abhishek Shukla cut IBM to Reduce from Hold and dropped his target to $191 from $231, arguing the stock's ~22x forward multiple was stretched against a ~17x sector median and that a 'synthetic' basket of IBM's peers offered better value at the same price. That $191 line matters here: with the perp at $219.90, the most bearish target on the Street still implies double-digit downside from current levels.

The real reckoning is July 22, when IBM has to put segment detail and forward guidance behind a preliminary release that already cost it decades of goodwill in one session. Between now and then, the perp is trading a full quarter of bad news that it, oddly, already front-ran on a few million dollars of volume. The open question is whether $17.2 billion was the floor or just the first data point.