Please Run ARPU Sensitivity Analysis
ARPU Sensitivity — Dollars Of EPS, FCF, And Fair Value Per 100 bps Of ARPU Change
Bottom Line
Charter's Q1 2026 print did not destroy a quarter of the company because volume cracked — it destroyed it because management told the tape that monthly residential broadband ARPU of $70.72 would stay flat for the rest of 2026, the first time in 20+ years that the cable price escalator stopped escalating. The dollar consequence is mechanical, not speculative. At ~100% drop-through (the right assumption for price moves on a residual base, since plant, headcount, programming, and bad debt do not flex), every 100 bps of residential broadband ARPU change is worth ~$1.16 of GAAP EPS, ~$184M of after-tax FCF, and ~$9.5 of fair value at constant 5.5× EV/EBITDA. Applied to the broader residential relationship metric ($118.44/mo, the figure that actually drove the print), the same 100 bps is worth ~$2.07/share of EPS and ~$17/share of fair value. Post-Cox close (197M FDSO), per-share sensitivities roughly halve. Crucially, the headline -1.4% Q1 26 ARPU print is ~170 bps of accounting reallocation on top of company-disclosed underlying +0.3% — so the run-rate EBITDA damage is materially smaller than the reported decline implies, and the market reaction is repricing the forward growth trajectory (loss of the historical +2.6% per-cycle pace), not the in-quarter dollar EBITDA hit.
$/EPS per 100 bps of broadband ARPU
$/EPS per 100 bps of residential blended ARPU
Fair value $/sh per 100 bps broadband ARPU (5.5x)
Fair value $/sh per 100 bps residential ARPU (5.5x)
Q1 26 broadband ARPU ($/mo) — guided flat 2026
Q1 26 residential ARPU ($/mo) — -1.4% YoY
FY25 diluted shares (M)
Post-Cox FDSO (M)
The conversion ratio in one sentence. $1.16 of GAAP EPS per 100 bps of residential broadband ARPU on the $23.8B Internet revenue base (100% drop-through × 77.3% after-tax × 86.5% after-NCI ÷ 137.7M FY25 diluted shares); $2.07/share when the same 100 bps moves the residential relationship ARPU on the $42.6B residential revenue base. The post-Cox math is the same dollar EBITDA divided by 197M shares — roughly half the per-share sensitivity, before crediting Cox's incremental EBITDA contribution.
1. Why ARPU Drop-Through Is ~100%, And Sub-Loss Is Only 70–90% — The Load-Bearing Distinction
The single most important modeling decision in this tab is the drop-through ratio. The companion EPS-per-1M-subs tab uses 70–90% because a departing customer saves some variable opex (modem returns, marketing pull-back, bad-debt save, programming saves on attached video). When ARPU moves on the residual base, none of that flexes. The customer is still there; the plant is the same; the bad-debt rate is unchanged; the technician headcount, programming contracts, and call-center hours are all already paid for. Every dollar of ARPU change flows essentially 1:1 to EBITDA — and from EBITDA the only leakage is federal/state tax and A/N's noncontrolling slice.
The takeaway. Of every $1 of revenue moved by an ARPU change, roughly 67 cents lands as NI to Charter shareholders — versus ~53 cents for a sub-loss-driven revenue change. The 14-cent gap is the variable cost save that does not exist when price moves but the customer stays. This is why a 100 bps of ARPU and ~280K of sub-loss are roughly equivalent in dollar-EPS terms, despite the optical asymmetry (1% feels small; 280K subs feels large). The market reaction to Q1 26 — collapse on flat 2026 ARPU guidance more than on the -120K sub print — suggests investors finally internalized this.
2. The Per-100-bps Conversion Table — Four ARPU Framings Side By Side
Each ARPU metric Charter discloses maps onto a different revenue base. The cleanest way to underwrite an ARPU view is to know which metric the market is reacting to and apply the right base.
Read this band, not the point. The defensible per-100-bps EPS sensitivity ranges from $0.06 (mobile) to $2.07 (residential blended) — a 35× spread driven entirely by which metric moves. A press headline that says "Charter ARPU down X%" tells the PM almost nothing without specifying which ARPU. The two figures the market reacts to are the residential blended ($118.44, the headline 10-K metric) and the broadband line ($70.72, the metric management gave 2026 guidance on). SMB and mobile are second-order.
3. The Headline Sensitivity Grid — ARPU Change × Revenue Base
The spine of the tab. Rows are ARPU change scenarios (-3% to +3%); columns are dollar-EPS impact by revenue base. Each cell uses 100% drop-through (30% for mobile), FY25 137.7M diluted shares, 22.7% tax, 13.5% NCI.
How to read this. A +1% residential blended ARPU year (roughly the Charter historical pace through FY25, before the Q1 26 break) is worth ~$2.07 on a $36.21 FY25 EPS base — ~6%. The "flat ARPU 2026" management commentary, against a "+1.5% consensus had pencilled in" baseline, costs ~$3.10 of EPS opportunity ($2.07 × 1.5) before any other moves — roughly 7% of the $41.91 FY26 consensus. A -2% residential blended print (worst-case persistent damage) would compress FY25 EPS to $32.07 and FY26 consensus to $37.77 — still inside the $34.85–$46.03 published estimate range, but at the low end. A +2% recovery would push FY26 consensus to $44.21 — at the high end of the panel.
4. The Reported-vs-Underlying Reconciliation — Why Q1 26 Was Not As Bad As It Looked
The single most important interpretive call: the Q1 26 -1.4% residential ARPU print was not -1.4% of real pricing pressure. Charter's disclosure isolates ~170 bps as an accounting reallocation — programmer-streaming-app costs netted against video revenue — and what's left is a small positive. Pricing the reported number flows into a $2.90/share-looking annualized EBITDA hit that mostly does not exist as EBITDA.
The optical hit vs. the EBITDA hit. Applying the reported -1.4% to the $42.58B residential revenue base × 100% DT × after-tax × after-NCI ÷ 137.7M shares implies a $2.90/share annualized EPS drag. But ~170 bps of the -1.4% is the programmer-app reallocation, which is EBITDA-neutral by construction (revenue down, programming cost down by the same amount netted into the video line). The real run-rate EBITDA drag from underlying pricing is the +0.3% — which is positive for EPS by about $0.62/share. The market reacted to a print that, in EBITDA terms, was much closer to flat than the -1.4% headline. The legitimate bear interpretation is not "Q1 was bad on a run-rate basis" — it is "the 2026 forward outlook (flat broadband ARPU for the year) signals loss of the historical +2.5–3.0% per-cycle rate-hike pace."
What changes for the Q2 print. The $218M Q1 26 reallocation will be in the prior-year base for Q2, Q3, and Q4 comparisons. Reported residential ARPU should mechanically improve ~85–170 bps on the YoY comparison even if underlying pricing is unchanged. A Q2 print of -0.5% reported would actually mean roughly -1.5% to -2.0% of underlying deterioration — much worse than Q1's +0.3% underlying — and would be a genuine bear signal. A Q2 print at or above 0% reported would mean underlying remains roughly flat to slightly positive — consistent with mgmt's "flat broadband ARPU 2026" framing but no worse. This is the asymmetry the next print resolves.
5. The Combined ARPU × Sub-Loss Matrix — The Missing Piece
The companion EPS-per-1M-subs tab quantifies volume risk in isolation; this section closes the loop by pairing both axes in a single 2D matrix. A coherent bear scenario layers both — and the brief's bear case ("subs fall AND price falls") cannot be read off either grid alone without double-counting or under-counting.
The PM-decisive read. A "-1M subs / -1% ARPU" stress (roughly: FY26 print at the brief's central bear scenario) costs $5.37/share — versus $3.30/share if you read only the sub-loss grid and $2.07/share if you read only the ARPU grid. A "-2M subs / -2% ARPU" sustained bear lands at $10.72/share, which on $41.91 FY26 consensus is a 25.6% EPS hit (pro-forma $31.19). That's the single cell most worth memorizing — it's the path that takes the print materially below the $34.85 FY26 low estimate and forces a multiple discussion. Conversely, a "Flat subs / +1% ARPU" relief case (the bull setup if Q2 prints clean) lifts EPS by $2.07/share, taking FY26 consensus to ~$43.98 — squarely into the high-end of the panel. The grid spans roughly $24/share of EPS optionality between the worst and best cells, which is most of what the equity is debating.
6. Peer Wedge — Charter Flat vs. Comcast +2.6%, And What It Costs Over 3–5 Years
Comcast Q3 2025 residential broadband ARPU growth was +2.6% YoY (Trefis/Fierce), and Q4 2025 was +1.1%. Against Charter's FY25 +0.3% residential blended print and management's 2026 flat broadband outlook, the relative gap is the visible margin of pricing-power impairment the equity is repricing on. The dollar opportunity cost of trailing Comcast by 100–300 bps compounds quickly.
The 5-year math anchors the stock-price reaction. If Charter sustainably trails Comcast by ~260 bps (the Q3 25 wedge) over 5 years on the broadband line, that's ~$124/share of fair value foregone at constant multiple — comparable to the $89/share Charter actually lost in May 2026. The market priced most of a 5-year structural wedge in 30 days, which is either (a) the right discount of a permanent regime shift or (b) a meaningful overshoot if Q2/Q3/Q4 2026 prints normalize. Historical reference: Charter's last two rate-hike cycles (Oct 2022 and Dec 2020) were each ~$5/month rack-rate moves estimated by New Street Research at ~+2.6% annualized broadband ARPU — the same ~2.5–3.0% per-cycle pace Comcast is currently running. Missing one full rate cycle is roughly $3/share of EPS forgone.
7. Share-Count Regime Overlay — The Cox-Close Divider
ARPU sensitivity is mechanically inversely proportional to share count. Charter's denominator is in motion: standalone buyback math, Q1 26 as-converted (179M with Liberty/A/N exchange), or post-Cox FDSO (197M). Per-100-bps EPS sensitivity roughly halves once Cox closes.
The Cox close changes the per-share math, not the dollar math. Same $159M (broadband) or $285M (residential) NI delta to Charter holders per 100 bps — but the share count denominator is ~43% larger post-Cox (137.7M → 197M), so per-EPS-per-100-bps sensitivity drops ~30% from $1.15 → $0.81 (broadband) and $2.07 → $1.45 (residential). This grid is standalone math on a standalone share count. Post-Cox the EBITDA base also grows (Cox brings ~$5B+ of incremental EBITDA on its own ~6M residential Internet base and ~$75/mo broadband ARPU), and the $413M/yr Cox preferred dividend (6.875% × $6B notional) layers in above common EPS. The catalysts and long-term-thesis tabs own the combined-pro-forma exercise.
8. From EPS To FCF And Fair Value — The Multiple Translation
EPS is the metric consensus prints against; FCF and EV/EBITDA are what the equity actually trades on. Since ARPU drop-through is ~100%, the EBITDA delta tracks revenue 1:1, FCF tracks EBITDA at the after-tax rate (capex and net interest unchanged for a price move), and fair value at constant multiple is mechanical.
The $238 → $149 drawdown decoded. $89/share of equity destruction × 137.7M ≈ $12.3B of market cap erased in 30 days. At 5.5× EV/EBITDA, that implies $2.23B of EBITDA expectation cut — which would correspond to roughly a sustained -940 bps of broadband ARPU OR -520 bps of residential blended ARPU OR (more plausibly) a permanent multiple re-rate from ~6.5× to ~5.5× layered on top of a smaller EBITDA mark-down. The mechanical bridge says the actual in-year underlying ARPU damage from Q1 26 was negligible (+0.3%); the market reaction is repricing the long-term growth trajectory, not the in-quarter print. For a PM, the question is which path Q2/Q3/Q4 2026 prints validate: a return to +0.3% to +1.0% underlying residential ARPU prints would unwind much of the multiple compression; a slide toward -1% to -2% underlying would confirm the bear thesis and the $124/share fair-value-loss math above.
9. Calibration — Does The Bridge Reproduce Q1 2026 Actuals?
The calibration confirms the central interpretive point. Reading the -1.4% reported residential ARPU print at face value and flowing it through the mechanical bridge produces a $2.90/share annualized EPS drag and ~$15–17/share fair-value impairment. That number is wrong as a run-rate EBITDA signal because ~170 bps of the print is the programmer-app reallocation (revenue-down/cost-down, EBITDA-neutral). The underlying +0.3% × residential base produces a +$0.62/share positive EPS contribution — i.e., underlying pricing actually helped, marginally. The market reaction (-$89/share) is rational only if Q2/Q3/Q4 2026 prints validate a forward trajectory step-down of -100 to -200 bps vs the historical +2.5–3.0% per-cycle pace — not if the in-quarter -1.4% print were the run-rate damage. The Jul 24 Q2 print is the first test.
10. Watch Items — What Q2/Q3/Q4 2026 ARPU Prints Mean In Dollars
11. Pitfalls, Limitations, And What This Tab Does Not Try To Answer
12. Decision Frame
The bridge is intentionally simple: drop-through (~100% for ARPU on the residual base), tax (22.7%), NCI (13.5%), share count (137.7M today / 197M post-Cox) — in that order of importance. A PM who walks away with the two central conversion ratios ($1.16/share per 100 bps of broadband ARPU; $2.07/share per 100 bps of residential blended ARPU), the combined ARPU × sub-loss matrix from §5, and the reported-vs-underlying distinction from §4 can convert any forward ARPU view into an EPS and fair-value position in real time. The honest read is that the equity at $140 is pricing a 5-year sustained ~260 bps trailing-Comcast wedge on the broadband line — a strong claim about the durability of the FWA $50 anchor that will be resolved (in either direction) by the next two to four ARPU prints, beginning with Q2 2026 on July 24.