COMP · Compass, Inc. · Stock research

Current Thesis

The binary resolved on 2026-08-04 and it resolved in the bulls' favour — revenue $4.306B against $4.081B consensus, adjusted EBITDA $363M above the guided $310–350M, the Year-1 Anywhere cost-synergy target actioned five months early and raised from $300M to $330M. Twelve sessions later the price is lower: the 2026-08-07 close was $12.58, the 2026-08-21 close $11.75, leaving the shares 13.5% under the $13.58 52-week high with RSI(14) at 50.0 versus 59.9 on 2026-08-07. What changed in between is not the company, it is the volume series the company sells into. NAR existing-home sales for July, released 2026-08-11, fell 1.7% month-over-month to a 4.06M annualised rate — a second consecutive monthly decline after June's 2.4% drop. NAR pending home sales for July, released 2026-08-18, printed an index of 71.2, down 2.3% month-over-month and 2.2% year-over-year, with the West down 4.7% sequentially and 7.1% from a year earlier. Contract signings lead closings by roughly 30–60 days, so July pendings land inside the quarter Compass has already guided to $3.85–4.05B. The narrative leg still on offer is the ~$170M of the three-year $500M synergy plan that has not yet been banked, converting to EBITDA while pro-forma transaction growth stays positive. The leg that has visibly weakened is the transaction count that growth rests on.

The narrative is saturated. Dating it: the best fundamental news of this cycle — a beat above the top of guidance, a synergy raise, and a Q3 guide set above the $3.745B street revenue number — arrived on 2026-08-04, drew six sell-side actions in seven sessions spanning Goldman Neutral $12 to Oppenheimer Outperform $18 plus a Zelman downgrade to Neutral $13.50 on 2026-08-06, and has been followed by a lower price and a 6.9% single-session decline on 2026-08-20. Coverage is broad and the fundamental print was good; the marginal buyer did not show up for either. That is late-cycle behaviour in a story that is well understood, and it is a step on from the maturing read carried since the 2026-08-09 update.

Bullish and bearish views on Compass, Inc.

The model's bull view on Compass, Inc. (COMP), in brief: Q2 2026 (2026-08-04): revenue $4.306B vs $4.081B consensus; adjusted EBITDA $363M vs a $310–350M guide; GAAP net income $92M. The bear view: The leading housing indicator turned on 2026-08-18: July pending home sales index 71.2, -2.3% month-over-month, -2.2% year-over-year, with the West — Compass's highest-GTV geography — down 4.7% sequentially and 7.1% year-over-year. Both cases follow in full.

Bull Case

• Q2 2026 (2026-08-04): revenue $4.306B vs $4.081B consensus; adjusted EBITDA $363M vs a $310–350M guide; GAAP net income $92M. Both headline lines cleared the range management set on 2026-05-05.
• Synergy capture ahead of schedule: the full Year-1 $300M was actioned five months early and the target raised to $330M; the 2026 realized goal went from $200M to $220M ($150M P&L, $70M capex). The three-year $500M target is unchanged, leaving roughly $170M unbanked.
• Pro-forma organic momentum stripping the 2026-01-09 Anywhere close: Brokerage GTV $155.2B, +15.9% year-over-year on 153,009 transactions, +7.4%. Management put Q2 market outperformance at 1,000bps versus 580bps in Q1.
• Price per transaction is not the problem: the July NAR release (2026-08-11) put the median existing-home price at $434,100, up 2.0% year-over-year and the 37th straight month of increases, and NAR's Lawrence Yun noted year-to-date sales up 2.4%.
• Balance sheet built without equity issuance: Q2 operating cash flow $191M, free cash flow $180M, cash $694M (+$210M QoQ). Management stated intent to redeem the $500M 9.75% notes when first callable in Q2 2027.
• Agent economics firmed: principal agent retention 95.5% in Q2 versus 94.1% in Q1, on 2,816 gross additions.

Bear Case

• The leading housing indicator turned on 2026-08-18: July pending home sales index 71.2, -2.3% month-over-month, -2.2% year-over-year, with the West — Compass's highest-GTV geography — down 4.7% sequentially and 7.1% year-over-year. Yun attributed it to "the highest mortgage rates of the year" hitting mid-summer.
• Closings are already rolling: existing-home sales fell 1.7% in July to a 4.06M rate (2026-08-11), the second consecutive monthly decline. That was the exact observable named as the volume risk in the prior update, and it printed.
• The Q2 report was not clean below the revenue line: EPS $0.11 versus a $0.12 estimate, and FY2026 non-GAAP opex guidance was raised to $2.75–2.80B from $2.70–2.75B, $35M of it from a July M&A closure.
• Agent count fell sequentially to 83,184 from 84,187 in Q1 on separation of low-productivity agents; the retention improvement and the roster decline describe one action.
• Franchise economics keep diluting: net royalty per side of $505 in Q2, down 14.6% year-over-year, against Franchise GTV of $120.0B (+11.7% pro forma).
• The New York Attorney General's antitrust inquiry, opened 2026-06-03, remains open with no timetable and no publicly disclosed civil investigative demand.
• The listing-access fight with Zillow has no scheduled resolution: a federal judge denied Compass's preliminary-injunction request in February 2026, and the shares fell 11.8% on 2026-02-23 to about $9 on that ruling.
• Third-party valuation work is now leaning against the move — GuruFocus published on 2026-08-20 that the $11.92 price sat 33.5% above its $8.93 GF Value estimate.

Setup & Price Structure

• Reference close 2026-08-21: $11.75. The 52-week high of $13.58 is 13.5% above; the shares are still up 39.9% over three months, so the pullback is inside a trend rather than a broken one.
• Momentum has bled out without a capitulation: RSI(14) 50.0 on 2026-08-21 against 59.9 on 2026-08-07. The single largest recent down-session was 2026-08-20, a 6.9% decline to $11.92, two sessions after the July pendings release.
• The structure that matters is the $11 area — the shelf built through the July consolidation above the June trough near $9.78 that followed the 2026-06-03 AG headline. Losing it on a weekly close puts that June low back in play; holding it keeps the post-Anywhere uptrend intact.
• Positioning observables, stated as observed: Investor's Business Daily's SwingTrader model was reported buying the stock on 2026-08-13 and selling it on 2026-08-17 — a four-session round trip in a retail-facing momentum list. Sell-side targets bracket spot unusually wide ($12 Goldman Neutral to $18 Oppenheimer Outperform, actions dated 2026-08-05/06), which is what dispersion looks like when the fundamental case and the macro case point opposite ways.
• No company-specific print inside the next 30 days. Q3 results are not expected until roughly 2026-11-03, so the tape between now and then is set by housing data and rates, not by disclosure.
• Insider flow is scheduled rather than discretionary: CEO-affiliated trusts sell under a Rule 10b5-1 plan adopted 2026-03-03, most recently 23,456 Class A shares at a $12.5056 weighted average on 2026-07-01. No new Form 4 has been filed since.

Catalyst Calendar (next 30 days)

• 2026-09-10 — NAR existing-home sales, August 2026 (per NAR's published release schedule). A third consecutive monthly decline would extend the series that the Q3 revenue guide implicitly assumes stabilises.
• ~2026-09-16 (est.) — FOMC decision. Yun tied July's weak signings to the year's highest mortgage rates; the rate path is the transmission channel into Q4 closings.
• 2026-09-17 — NAR pending home sales, August 2026. The cleanest forward read on Q4 sides and the first chance to show July's -2.3% was a rate-driven single month.
• ~2026-11-03 (est.) — Q3 2026 results. Outside this window, but it is the event that grades the $3.85–4.05B revenue and $275–305M adjusted EBITDA guide and whether the $330M actioned synergy figure reaches EBITDA rather than being absorbed by the raised $2.75–2.80B opex plan.

What Would Change Our Mind

The structural break is the July–August shelf near $11 built above the June AG-dip low of $9.78: a weekly close below $11 ends the post-Anywhere advance as a trend and reopens that low, and given the shares are 13.5% below the $13.58 high with RSI already back at neutral, the cushion between the two is thin. The secondary break is macro and dated — a third consecutive monthly decline in existing-home sales at the 2026-09-10 NAR release, or a second consecutive decline in pending home sales on 2026-09-17, would confirm that the volume series moved against the $3.85–4.05B Q3 guide rather than dipping for one rate-driven month. A disclosed civil investigative demand or a new legal-proceedings item naming the New York AG in the Q3 10-Q would restore the June gap risk on top of that.

What would push the other way and argue the saturated label was called early: August pendings printing positive on 2026-09-17, the shares reclaiming the $13.58 high, or the Buy-side cohort (Barclays $16, UBS $17, Oppenheimer $18) raising numbers on evidence of Q3 sides holding rather than on synergy arithmetic already disclosed.

Correlation Notes

• The name now trades as a housing-transaction volume proxy, not as a merger-integration story. The two moves that mattered since the last update were both NAR releases (2026-08-11, 2026-08-18), and neither was company-specific.
• Rate sensitivity runs through mortgage rates to contract signings; NAR's July commentary explicitly named the year's rate high as the cause of the -2.3% pendings print.
• Sector context is mixed rather than uniformly weak: 24/7 Wall St reported on 2026-08-18 that Opendoor is down 42% in 2026, while Compass has held a large three-month gain — so the proptech complex is not moving as one block, and a Compass drawdown is not automatically a read-through from peers.
• Revenue is volume-driven. Median price at $434,100 and 37 straight months of year-over-year price gains support GTV per side, but the transaction count is the swing variable in both the guide and the tape.