PEGA › analyze
PEGA · Analyze
Full analysis · 2026-08-10 · $33.01 · Technology / Software–Application
One line: A business whose pricing model is structurally immune to the AI-seat problem, trading at 10.9x EV/FCF — and whose unit volume is collapsing anyway. The cheap multiple is real; so is the deceleration. WATCH, conviction 5.5.
0. Knowledge-base check (Step 0)
kb.py find PEGA → no matches. First analysis of this name.
Knowledge/Themes/software-saas.md (live, recheck 2026-09-15) governs the field and supplied the
governing diagnostic — can the company still take price? — which is 2-for-2 (ADBE ✗, TEAM ✓).
Pitfalls applied before trusting any vendor figure:
[[pattern-net-margin-above-operating-margin-is-a-tripwire]] · [[pitfall-tax-valuation-allowance-round-trip-breaks-eps]] ·
[[pitfall-roicai-free-plan-caps-history-at-two-years]] · [[pitfall-vendor-forward-eps-is-stale-on-the-day-of-a-guidance-cut]].
Three of the four fired.
1. 🚩 Data integrity — the headline P/E is void
The tripwire fired immediately: FY2025 net margin 22.5% sat above operating margin 16.3%. Under [[pattern-net-margin-above-operating-margin-is-a-tripwire]] the trailing multiple does not qualify until the below-the-line source is named. It is named, and it is the tax line.
| FY | Pretax income | Tax provision | Net income | Effective rate |
|---|---|---|---|---|
| 2022 | −$161.80M | +$183.79M (expense on a loss) | −$345.58M | n/m — allowance ESTABLISHED |
| 2023 | $95.44M | +$27.63M | $67.81M | 29.0% |
| 2024 | $142.64M | +$43.45M | $99.19M | 30.5% |
| 2025 | $280.63M | −$112.81M (BENEFIT) | $393.44M | −40.2% |
The identity closes exactly: $280.63M + $112.81M = $393.44M. Confirmed three independent ways —
the Q4'25 tax provision of −$129.62M, the cash-flow deferred-tax line of −$171.05M, and the
balance sheet, where non-current deferred tax assets jumped $3.97M (Sep-25) → $175.47M (Dec-25).
The Q2'26 10-Q attributes it to "the absence of a valuation allowance on substantially all of the
Company's U.S. and U.K. deferred tax assets."
This is the third confirmed instance of [[pitfall-tax-valuation-allowance-round-trip-breaks-eps]] (after QCOM and DT) and the cleanest yet — both ends of the round trip are visible in the same four-year window. FY2022's loss was inflated by the charge; FY2025's profit by the release. Any net-income CAGR spanning 2022–2025 is noise.
Consequence for valuation:
| Reported | Normalized (21–24% tax) | |
|---|---|---|
| FY2025 net income | $393.4M | ~$211–222M (−47%) |
| TTM net income | $324.0M | ~$142–148M |
| TTM diluted EPS | $1.77 | ~$0.83 |
| Trailing P/E | 18.8x | ~40x |
The $129.6M Q4'25 benefit sits inside the current TTM window and rolls off after the Q3 2026 print — the reported P/E will roughly double on its own, with no change in the business.
Other vendor traps found:
- PE(fwd) 12.52 is a non-GAAP construct and is not comparable to either figure above. It also
predates the Q2 miss — [[pitfall-vendor-forward-eps-is-stale-on-the-day-of-a-guidance-cut]].
- Yahoo FY2025 operating income $284.39M vs as-filed 10-K XBRL $263.10M. Prefer the filing.
- Short interest 17.8% of float is ~9.6% of shares outstanding — Trefler's ~46% stake means float is
only ~54% of the company. Any founder-controlled name needs this adjustment before short interest is
compared cross-sectionally. (New pitfall candidate.)
- Headcount unreconciled: Yahoo 5,598 vs Revelio ~4,896.
- roic.ai returned only 2 annual periods and refused quarterly — [[pitfall-roicai-free-plan-caps-history-at-two-years]].
2. Fundamentals
The business signal is ACV, not GAAP revenue
PEGA recognizes term-license revenue ~100% upfront at near-100% margin, so quarterly GAAP revenue and operating income are close to useless as a business read. Q1'26 showed revenue −10% and GAAP net income −62% while ACV grew +12% and FCF grew +2%. Read ACV and FCF; ignore the quarterly top line.
| Q4'25 | Q1'26 | Q2'26 | |
|---|---|---|---|
| Total ACV | +17% | $1,622M, +12% | $1,620M, +7% |
| Pega Cloud ACV | +33% | $906.7M, +29% | $926.3M, +22% |
| Backlog / RPO | — | $2,011M, +16% | $2,019M, +10% |
🔴 Total ACV declined sequentially, $1,622M → $1,620M, and reported ACV growth halved in two quarters. Backlog growth halved too. This is the single most damaging fact in the file, and it is company-specific — not a sector alibi.
Health scorecard (TTM through Q2'26, my computation)
| Metric | Value | Read |
|---|---|---|
| Revenue TTM | $1,736.4M | H1'26 −1.1% YoY ($850.7M vs $860.1M) |
| Operating income TTM | $196.5M (11.3%) | ⚠️ down from 16.3% FY25; H1'26 op income −63% YoY |
| FCF TTM | $492.5M (28.4%) | ✅ held flat through the worst two quarters |
| SBC TTM | $159.1M (9.2% of rev) | ⚠️ rising in dollars (+8.8% FY25, +5.0% H1'26) |
| SBC-adj FCF TTM | $333.4M (19.2%) | ✅ genuinely positive, unlike most of the field |
| SBC / FCF | 32% | ✅ passes the March-cohort filter that killed ZS/HUBS/MDB/PD |
| Debt / Assets | 5.2% | ✅ $72.0M, all capital leases — zero financial debt |
| Net cash | +$289.9M | ✅ converts fully repaid in 2025 |
| Diluted shares | 184.10M → 171.77M (−6.7% in 4 qtrs) | ✅ real per-share tailwind |
FCF CAGR (framework requires 5–8yr; only 4 years available from vendors, flagged): −$13.0M (2022) → $201.0M (2023) → $338.2M (2024) → $490.7M (2025). The ramp is the back end of the perpetual-license → subscription transition (cash collection catching up), not repeatable operating leverage. Do not extrapolate it.
Capital allocation FY2025 — genuinely good
FCF $490.7M deployed into: buybacks $517.2M (105% of FCF), debt repayment $467.9M (the converts), dividends $15.4M. Funded partly by drawing the securities portfolio ($560M of sales). Debt went $549.7M → $76.0M; the company emerged debt-free with net cash. TTM buybacks are $629.5M = 128% of FCF, retiring ~11.6% of the company.
⚠️ Equity is being consumed: $787.4M (Dec-25) → $560.4M (Jun-26). Cash+securities $425.8M → $361.9M.
This pace is not indefinitely fundable from a $492M FCF base.
⚠️ Unexplained: Dividends Payable jumped $5.11M (Mar-26) → $83.96M (Jun-26). PEGA's regular
dividend is ~$5M/quarter. Not verified — could be an accrued-buyback reclass or a declared special
dividend. Check the Q2'26 10-Q before relying on any per-share cash figure.
3. Moat — and the pricing-power test
🎯 PEGA passes the pricing test, and it does so by inverting it
PEGA bills per CASE — a completed workflow — not per seat.
Trefler: "tying our economics directly to the business value delivered rather than to users or seats." Trefler (Q2'26): "Pega does not charge per token… monetization strategy is based on the business value that clients create."
2026-06-08, Pega Infinity '26: clients pay a flat price per completed case regardless of AI consumption, and PEGA absorbs the token cost. From the Q2 call: "case price uplift for advanced AI-powered runtime capabilities, including innovations such as Agentic Process Fabric."
That is a price increase — the direct analogue of Atlassian's +15% Data Center raise in Feb 2026, and the opposite of Adobe's deferred Creative Cloud increases. No deferred or cancelled increase was found anywhere in the record.
And the structural point is stronger than the price action. The bear case that broke the per-seat cohort — AI agents replace knowledge workers, so a per-seat base shrinks with headcount — does not apply to a case-metered vendor. The cases still get processed; arguably more of them. Mechanically, agent-driven volume is revenue-accretive here.
🔍 But ACV growth still halved. So the diagnostic returns a pass and the business missed anyway. PEGA is the case that proves a company can pass the pricing test and still fail — because its binding constraint is unit volume and sales execution, not price. That is a genuine refinement to the 2-for-2 framework and the most valuable methodological output of this analysis.
Management named both causes, and the second is the damaging one:
CFO Stillwell: "unprecedented change in the software market created significant buyer uncertainty… The market entered a token-maxing mindset." — exogenous. Stillwell: "we didn't execute well enough on our go-to-market change to drive deeper and broader engagement… Our progress in the first half was slower than we anticipated." — self-inflicted.
⚠️ Unresolved: a third-party vendor-pricing site quotes PEGA list at $150–$300 per named user per month, contradicting the "not users or seats" claim. Most likely legacy on-prem/term contracts retain named-user metering while new Pega Cloud is case-metered. The ACV split between seat-metered legacy and case-metered cloud is not disclosed, and it is the number that would settle this definitively. Treat the pricing pass as probable, not certain.
Switching costs — the strongest evidence in the file
Pega Cloud ACV +22% to $926.3M (~57% of total ACV) while total ACV grew +7%. The installed base is migrating deployment models rather than leaving — the hardest test of switching cost there is, and PEGA is passing it. FCF held +1% in H1 through the two worst bookings quarters: revenue that keeps paying while new business stalls is the signature of contractual lock-in. Long-term Cloud target ~75% of ACV.
Adversarial stress-test
The squeeze is two-sided. Microsoft Power Platform commoditizes the simple end on price and Office distribution; ServiceNow — far larger, better distributed, stronger AI narrative — takes the enterprise platform-consolidation budget. Appian attacks the same BPM heritage; UiPath from RPA; Camunda developer-first. PEGA's defensible middle — very complex, very regulated, very long-lived case management in financial services, insurance, healthcare payers and government — is real but is the smallest of the three slices.
Does GenAI commoditize low-code workflow? Both effects are live and they hit different layers: - Erodes the authoring layer. Blueprint's own existence proves an LLM can generate a workflow spec. PEGA has effectively conceded application creation is no longer scarce — which is precisely why it moved monetization onto execution. - Deepens the runtime layer. Trefler: "if clients generate code directly with tools like Claude or Codex, where is the architecture? How is the business logic captured and understood?" Deterministic, auditable, replayable execution of a long-lived regulated case is what a stochastic model is worst at. This is the strongest argument in the file and it is not rhetoric. - The unhedged risk: if ServiceNow or an LLM-native entrant ships a governed, auditable agentic runtime, PEGA's remaining differentiation is the installed base and 20 years of accumulated customer process logic — a defensible but decaying asset.
⚠️ [S, unverified] Secondary sources report Pega is no longer a Leader in the 2026 Gartner Enterprise LCAP Magic Quadrant (previously a Challenger; Leaders now Microsoft, Mendix, OutSystems, ServiceNow, Appian, Salesforce). The primary MQ was not obtained. If true it is a meaningful moat-erosion signal; verify before relying on it.
Moat rating: 6.0/10 — genuine switching costs and a structurally advantaged pricing meter, against a shrinking defensible niche and a two-sided competitive squeeze. Evergreen: no. The runtime-governance thesis is credible for 5 years and speculative at 10.
4. Valuation
Dividend models (DYT/DDM) are N/A — the 0.37% yield is incidental. Graham's IV computes to $8.12 (normalized EPS $0.83, BVPS $3.41) against a $33.01 price, but book value has been deliberately consumed by buybacks on an asset-light business; Graham is not informative here and is not weighted.
Weight goes to FCF-based frames, and they disagree — which is the honest finding.
| Frame | Metric | Multiple |
|---|---|---|
| Reported FCF | EV $5.38B ÷ $492.5M | 10.9x |
| SBC-adjusted FCF (owner earnings) | EV $5.38B ÷ $333.4M | 16.1x |
| Operating income | EV ÷ TTM EBIT $196.5M | 27.4x |
| Revenue | EV ÷ $1,736M | 3.10x |
| Normalized GAAP earnings | — | ~40x |
(EV = 171.77M diluted shares × $33.01 − $289.9M net cash = $5.38B.)
DCF, on SBC-adjusted FCF, 10% discount:
| Scenario | Assumptions | Value/share |
|---|---|---|
| Bear | ACV growth → 0%, margin compresses; SBC-adj FCF $300M, no growth | $19 |
| Base | ACV stabilizes ~7–8%, FCF $355M growing 5% for 5yr then 3% | $35 |
| Bull | GTM fixed, ACV reaccelerates; FCF $380M growing 8% then 3% | $42 |
🔴 The FY2026 guidance is arithmetically almost unreachable
PEGA issues guidance once a year and does not update it quarterly. The February guide — $2.0B revenue (+15%), 15% ACV growth, $575M FCF — is therefore still the posted number, and it is stale in a way a screen cannot detect.
- H1'26 actual revenue $850.7M (−1%). Hitting $2.0B needs H2 of $1,149M = +29.8% YoY.
- H1'26 FCF $288.3M. Hitting $575M needs H2 of $286.7M = +40% YoY vs H2'25's $204.2M.
Stillwell said recovering the H1 shortfall "will be very difficult" and that the $575M target "remains in view, but any ACV shortfall could pressure that goal." There is no formal cut on the record. Treat the posted guidance as void; my base case assumes ~$500–520M FY26 FCF.
Fair value range: $19–$42, central ~$35. At $33.01 the stock sits just below the central case — roughly fair, with a wide bear tail, not a bargain.
5. Sentiment & positioning
The de-rating is recent and specific: four downgrades in 48 hours on 2026-07-22/23 after the Q2 print (−19.2% premarket to ~$25). William Blair Outperform→Market Perform; Loop Capital Buy→Hold ($55→$25); KeyBanc Overweight→Sector Weight; Citizens Outperform→Market Perform. PT cuts: RBC $60→$35, DA Davidson $55→$32, Barclays $48→$39, Rosenblatt $58→$47. In April, Citigroup still carried $71. Yahoo mean target $43.00; a widely-quoted $59.82 consensus is stale and un-marked.
🚩 Zero open-market insider purchases in the entire 13-month window — including after a −52% drawdown. Trefler sold ~659,672 shares for ~$36.9M in metronomic 45k/90k blocks (clearly a 10b5-1 program, not signal-bearing), and those sales stopped after 2026-02-18 — plausibly plan expiry or blackout, not conviction. Stillwell sold 44,000 @ $44.81 in March; other officers sold on option exercises. Not one insider bought at $25–$33. That is the cheapest available vote of confidence and it was not cast.
🚩 Governance: Trefler holds ~46%, so PEGA is effectively activist-proof. The usual value-unlock pressure valve does not exist. Institutional ownership 61% shows no mass exit; no 13D or activist campaign.
Short interest 17.8% of float / ~9.6% of shares out, 6.05 days to cover — crowded but not extreme once the founder-float distortion is removed.
6. ⚖️ Appian litigation — a live, unreserved tail risk
- 2022: Fairfax jury awarded Appian $2.036B for trade-secret misappropriation + a Virginia Computer Crimes Act violation.
- 2024: Court of Appeals reversed, remanded for new trial.
- Jan 2026: the Supreme Court of Virginia AFFIRMED the new-trial order and REJECTED Pega's cross-appeal seeking dismissal. Reversal was on four damages-related evidentiary/instructional errors — not on whether misappropriation occurred. The appellate courts did not exonerate Pega.
- Pega did NOT appeal the Virginia Computer Crimes Act liability finding. That liability stands.
- Status conference was set for 2026-05-07; no retrial date found. No reserve or accrual disclosed.
A corrected retrial should produce a far smaller number than $2.036B, but the floor is not zero. Against a $5.43B market cap and $361.9M of liquid assets, even a heavily reduced award is material. Pull the Q2'26 10-Q contingencies footnote before sizing any position.
7. Synthesis — weighted verdict
Tensions between the agents, resolved:
| Tension | Resolution |
|---|---|
| Fundamentals: "FCF is excellent, 10.9x EV/FCF" vs Moat: "ACV +7% and decelerating" | Both true and not in conflict. FCF is a lagging collection of ACV signed 1–3 years ago. Flat FCF on decelerating ACV is exactly what you would expect before the deterioration reaches cash. Weight the leading indicator: ACV. |
| Moat: "passes the pricing test" vs Fundamentals: "growth halved anyway" | The diagnostic was built to detect seat/price compression. PEGA's failure mode is volume and execution. Applying the 2-for-2 rule here without that distinction produces a false positive. The pricing pass is real but non-dispositive. |
| Valuation: reported-FCF frame ($43/sh) vs SBC-adj frame ($30/sh) | Weight the SBC-adjusted frame — SBC dollars are rising, so the gap will not self-close. Central $35 sits between, closer to the honest frame. |
Verdict: WATCH — conviction 5.5 / 10
The bull case in one paragraph. A debt-free, net-cash, 76%-gross-margin business at 10.9x EV/FCF and 16.1x owner earnings, retiring ~7% of its shares a year, whose per-case pricing model is structurally immune to the exact risk that de-rated the entire per-seat software cohort — and which raised prices in June 2026 while competitors deferred theirs. If GTM execution is fixed and ACV stabilizes at 7–10%, this is worth $35–42 and you are paid to wait.
The bear case in one paragraph. Total ACV went sequentially negative; growth halved in two quarters; backlog growth halved; the posted FY26 guide needs +29.8% H2 revenue and will be missed; the reported P/E is about to double on a tax roll-off with no business change; SBC dollars are rising while headcount is flat; management admitted it broke its own go-to-market; an unreserved trade-secrets retrial is pending on a liability finding it declined to appeal; and not one insider bought a share at $25.
Why 5.5 and not higher: the pricing-power insight is genuinely differentiating and the cash multiple is genuinely low — but [[principle-down-a-lot-is-not-cheap]] applies directly, and the leading indicator is still deteriorating. You do not need to catch this knife. There is no scenario where waiting one quarter for ACV to stabilize costs you the thesis, and one where it saves you 30%.
Key risks, named: (1) ACV growth prints below 7% in Q3 — the thesis breaks; (2) Appian retrial award; (3) FY26 FCF guide missed materially, which is the one metric holding the valuation up; (4) Gartner Leader-quadrant loss confirmed; (5) founder control removes any activist backstop.
A portfolio-specific passage was removed from the public build.
8. What would change this verdict
🟢 Upgrade to 6.5+: ACV growth stabilizes ≥8% for two consecutive quarters · the seat-vs-case ACV split is disclosed and confirms majority case-metered · an insider buys in the open market · Appian settles for a disclosed, contained sum · FY26 FCF lands ≥$550M.
🔴 Downgrade / avoid: ACV growth <5% or a second sequential decline · FY26 FCF below $450M · an adverse Appian retrial verdict · Pega Cloud ACV growth below 15% · SBC/revenue rising above 11%.
Next catalyst: Q3 FY26 print, ~2026-10-21. Watch ACV growth and the sequential ACV dollar level first, backlog second, FCF third. Quarterly GAAP revenue is the least informative line in the release and will mislead in both directions.