When: Close below $197 (July 23 lower Bollinger test)

Then: Exit residual position to zero

When: Securities inquiry escalates with material disclosure/indictment

Then: Exit residual position to zero

When: Q3 2026 misses with guidance reset or goodwill impairment

Then: Exit residual position to zero

HOLD

• Dividend of $1.68/share is backed by ~$11-12B annual FCF against $6.25B payout = 2x coverage ratio, not a yield trap
• Operating margin expanding from 13.5% to 18.5% over four years
• 2.88% dividend yield and 5.5% FCF yield provide defensive ballast
• 0.705 beta means IBM decays slower than average large-cap in risk-off tape
• July 14 volume spike (opened $250, hit $217, closed $215) was capitulation event, not institutional distribution; recovery from $197.77 to $238.42 with declining volume confirms basing pattern
• Together AI deal worth 0.07-0.12% of revenue is immaterial today but IBM consulting could capture multi-hundred-million revenue over 3-5 years if enterprise AI spend redirects
• FCF stable at $11.4B (last year), $11.7B (prior year), $12.1B (two years prior) — definition of stability
• ATR of $8.80 supports tighter $220 closing-basis stop (1.0 ATR from $229 entry) rather than aggressive analyst's $215 stop
• 2026-07-15 low of $209.70 is true panic level; $220 stop exits before second wave of panic
• Recommend trim of 30-50% toward the 30% end to preserve defensive hedge quality

→ vs aggressive: Aggressive analyst conflates price risk with income risk — dividend survives even if multiple compresses; 2x FCF coverage means income is not at risk even if stock declines further

→ vs aggressive: Aggressive analyst's interpretation of July 14 volume spike as institutional distribution is narrative bias — distribution happens on rallies, not capitulation bars; the open at $250 hitting $217 and closing $215 is a panic flush, not distribution

→ vs aggressive: Aggressive analyst's dismissal of AI catalysts as marketing is equally sloppy as bulls pricing them as $10B catalysts — the truth is in between; IBM consulting could redirect enterprise AI spend from Accenture and boutiques

→ vs neutral: Neutral analyst's waiting-for-confirmation approach is paralysis dressed as prudence — waiting for close above $240 or mainframe stabilization data requires accepting additional drawdown in the meantime; each confirmation trigger involves giving back more of the 24.7% drawdown

BUY (trim 40% to underweight with $2.15 residual stop, add back 25% on close above $2.40 with expanding volume and positive momentum)

• Revenue growth decelerated from 7%+ in FY25 to 1.1% YoY, a slope change not noise
• Operating margin compressed 160 basis points quarter-over-quarter
• Securities fraud inquiry is a binary overhang that cannot be priced
• July 14 67-million-share print opened at $2.50, hit $2.17, closed at $2.15 — consistent with forced liquidation/panic flush, not institutional distribution
• Dividend has grown 28 consecutive years and is covered 2x by free cash flow
• Q1 2026 trough: restructuring charges $359 million, revenue $15.9 billion; Q2 2026 recovered to $17.2 billion — sequential improvement on quarter-over-quarter basis
• FCF trajectory: $11.4 billion, $11.7 billion, $12.1 billion over three years — stable and cash generative
• Operating margin expanded from 13.5% to 18.5% over four years
• Beta of 0.705 confirms low-beta defensive profile
• ATR of $8.80 makes $2.20 stop (1x ATR) too tight; $2.15 stop (1.6x ATR) is appropriate for basing pattern volatility
• 2020-2026 low of $2.09 is the panic level
• Prediction market signals 85% probability of no Fed cuts in 2026 with only 8% recession probability — soft-landing with hawkish Fed backdrop favorable for defensive income compounder
• Re-entry trigger: close above $2.40 on expanding volume with MACD histogram positive and RSI pushing through 50

→ vs aggressive: Volume interpretation is wrong — July 14 print shows panic flush characteristics (wide range, heavy volume, gap down), not institutional distribution which happens quietly into bids on the way up

→ vs aggressive: Dividend trap framing conflates yield gap compression with income investor behavior — IBM dividend buyers prioritize growth track record and defensive beta, not absolute yield

→ vs aggressive: Treating one quarter of deceleration as structural break ignores sequential recovery: Q1 2026 was a trough with $15.9B revenue, Q2 2026 recovered to $17.2B

→ vs conservative: $2.20 stop is too tight at roughly 1x ATR for a name in a basing pattern — will get whipsawed out before structure resolves; $2.15 (1.6x ATR) is appropriate

→ vs conservative: 30% trim underweights the asymmetric downside signal — 40% captures roughly four-fifths of the downside while preserving 60% of defensive ballast