← back to picks
Founded 1968 (Sopra) . Steria merger 2014HQ Annecy / ParisReporting EURCredit Aaa/AAAValuation 2026-05-25Damodaran FCFF . 1000-iter MC FCFF · Dark v3

Deeply undervalued, robust to MC stress . +60.6% margin of safety

Market outside MC distribution - P(below) = 0.0%

Intrinsic value €354.99/share (post 15% governance haircut; pre-haircut €417.64) vs market €140.00. The Monte Carlo p5 tail still sits at €219 - every one of 1000 correlated stress draws lands above today's price. The freshly triangulated 5Y regression beta of 0.86 (vs the Damodaran Computer Services bottom-up 1.27) compresses WACC by ~140bp and widens an already-large gap.

Market €140 DCF €355 MC p50 €272 p5 €219 p95 €335 Bear €110 Bull €480
SectorComputer Services (EU)GeoFR 46% · UK 16% · DE 11%MC σ±€38/shGovernanceSopra GMT 34% votes · 15% haircutQualityROCE 20% · Synth AAAMultiple7.6× fwd P/E vs sector 23×
Intrinsic / share
€354.99
post 15% gov . pre €417.64
Market / share
€140.00
Euronext Paris . SOP.PA
Margin of safety
+60.6%
vs intrinsic
Enterprise value
€8.82B
69.9% terminal
Cost of equity / debt
6.69% / 2.09%
β 0.86 · CRP 0.76%
Terminal ROIC / g
0.00% / 2.00%
spread ~-530bp

What it sells, where it sells

Operating segments

€5.6B FY25 rev
FranceQ1 2026 +7.2% organic; defence/sovereign anchor43%
Europe (ex FR/UK)Germany dominant; Benelux EU institutions; Iberia growing35%
United KingdomPublic sector / SSCL / NHS SBS; soft 2025 quarters16%
SolutionsSoftware products (axway-adjacent); Paris cost base6%

France + UK = 59% of revenue and house the sovereign/defence book (NATO counter-drone, CNES space, Red Hat sovereign AI) — the moat the bull case actually relies on. Continental Europe is the mean-reversion lever.

Country mix (revenue-weighted CRP input)

🇫🇷France46%
🇬🇧United Kingdom16%
🇩🇪Germany11%
🇧🇪Belgium5%
🇪🇸Spain4%
🇮🇹Italy4%
🇨🇭Switzerland4%
🇸🇪Sweden3%
🇳🇱Netherlands2%
🇮🇳India2%

France + Eurozone = ~76% of revenue — weighted CRP works out to ~76 bps (Italy/Spain/Belgium drive most of it), blending into a ~5.0% ERP. India 2% is a supplier/labor override (~5,000 FTEs in Noida/Bangalore), not revenue.

Background — five things to know before the case

Context the share price doesn't carry on its face. Skim this once and the bull / bear bullets below stop reading as inside baseball.

Where we diverge from the sector

Each row asks: if we'd used the sector median for this one assumption instead of our override, how much would the share price change? Negative = our override is more conservative (worth less); positive = more aggressive (worth more). Skim down the rightmost column to see which bets are doing the work.

Assumption
Our input
Sector median
Value impact
01Revenue growth10-year CAGR — Y1-5 4.5%, fading to 2.0% terminal
4.5% Cautious vs CMD 6% blended target; Q1 2026 +4.4% organic supports
5.0% Computer Services US/global cross-section trend
−€18
02Operating marginGAAP Y5+ steady-state
9.0% 150 bps above sector; reflects sovereign/defence stickiness
7.5% Computer Services median (US 7.4% / Global 7.6%)
+€42
03Sales-to-capitalRevenue per € of new invested capital
2.2 → 2.5 M&A goodwill drag well below industry; consistent with roll-up DNA
4.0 Computer Services median (US 5.2 / Global 3.8)
−€28
04Terminal growthYear-10 steady-state nominal growth
2.0% Below EUR risk-free 2.4% (hard ceiling); mature Eurozone GDP
2.4% EUR 10-year risk-free (ceiling)
≈ €0
05Cost of capital10-year WACC
5.31% Levered β 1.27, blended CRP 76 bps, ~21% debt
7.20% Sector median β 1.12 → higher equity cost
+€22
Net effect of overrides
Overrides net +€18/share vs all-defaults. The margin uplift (sovereign mix) and lower WACC (low leverage, AAA credit) more than offset the conservative growth and the capital-heavy roll-up reality.
+€18
Our override Sector median Override adds value Override subtracts value

The story & the five claims

The 10-year story this DCF is built on, plus the five anchor claims that translate the story into model inputs.

The 10-year story

Sopra Steria is a French family-controlled European IT-services consolidator with a sovereign and defence specialisation — the kind of operator that wins NATO counter-drone work, CNES space-surveillance contracts, and Red Hat sovereign-AI partnerships, not the kind that runs hyperscale commercial digital transformation at Accenture margins. The central debate is whether the trailing five years of zero organic growth and a 7.6× P/E reflect permanent decline, or whether the April 2026 Q1 print (+4.4% organic ex-SFT, France +7.2%) is the inflection management's Capital Markets Day promised. The base case is the second reading without believing the upper end: revenue compounds at 4.5% (vs CMD blended 6%) to reach €8.2B by 2035, the GAAP operating margin recovers from 8.0% to 9.0% by year 5 (150 bps above sector, well below Capgemini's 12%), and the business throws off €245M of free cash flow in year one rising to €483M by year ten. Terminal growth is pinned at 2.0%, below the EUR risk-free rate. None of this assumes the bull-case sovereign re-rating; what it assumes is that a controlled but disciplined compounder stops getting priced like Atos.

Two debates worth pressure-testing

Is the Q1 2026 beat a real inflection or a single-quarter false dawn?
Our view: Q1 +4.4% organic ex-SFT (vs consensus +1.2%) with France +7.2% and every reporting unit growing is a clean break from the five-year 0-1% trend, but the July 29 H1 print is the non-negotiable rebase. If H1 prints organic ≥+3% with the FY guide raised, the base case holds; any miss and the bear case at ~€110 (Contrarian) becomes central.
Is FCFF DCF even the right tool for an M&A roll-up?
Our view: The First Principles dissent has a point — sales-to-capital 2.2× vs sector 3.8-5.2× is evidence Sopra consumes capital through goodwill, and FCFF treats acquisitions as reinvestment that produces growth (a partial double-count of acquired growth). We accept the friction as a known-and-priced drag: the 2.2 / 4.5% pairing IS internally consistent, but the reader should know terminal value carries 60.2% of EV and any roll-up margin slippage compounds.
CLAIM 01Revenue compounds at ~4.5% to €8.2B by 2035.CAGR 4.5% Y1-5 → 2.0% terminalBelow CMD 2024 blended ~6% guidance. Q1 2026 +4.4% organic ex-SFT supports the band; sovereign/defence tailwind (NATO 5% pledge, ReArm Europe) is a 5-15y story.
CLAIM 02GAAP operating margin recovers to 9.0% by year 5 and holds.9.0% target (FY25 base 7.8%)150 bps above Computer Services median (7.4%). Defence/sovereign contract stickiness justifies the spread; we explicitly do not model convergence to Accenture/Capgemini margins.
CLAIM 03Sales-to-capital 2.2× → 2.5× — heavier than the sector.S2C 2.2 (Y1-5) → 2.5 (Y6-10)€2B+ M&A goodwill drag puts Sopra well below industry median (US 5.2). Modest improvement assumes deal cadence normalises after the 2023-26 tuck-in wave.
CLAIM 04Mature business — convergence by year 5, not year 10.Year of convergence = 550-year-old firm in mature European markets. Margin recovery to 9% is a five-year project, not a ten-year hope. Standard Damodaran ramp for non-disruptive businesses.
CLAIM 05Terminal growth 2.0%, no failure risk.g = 2.0% < rf 2.4%; p(fail) = 0€512M cash, 6% FCF/revenue, leverage well below CMD <1.5× ceiling, demonstrated buyback discipline. Defence tailwind is finite — no perpetual moat assumed.
Where we diverge from sell-side
  • Governance haircut at 15%, not Damodaran 25-35%. Council bumped 0.10 → 0.15 — the lower edge of the Bolloré/Wendel band — because Sopra GMT has actually executed (SBS divestiture, two buybacks, 4.18% capital retirement). Pushing to 25% would double-count what 7.6× P/E already prices.
  • Margin trajectory at 9.0% GAAP, not Sopra-EBITA 10-11%. We use the GAAP line that hits the income statement, not the "op margin on business activity" KPI management quotes (which strips SBC + PPA). The ~1.5pt gap is real cash compensation and amortised deal intangibles.
  • Sales-to-capital 2.2× honest about the roll-up reality. Sell-side comps assume a 4.0× services business; we surface the M&A goodwill drag and let it compound. Terminal ROC still lands at 16.8% — the business creates value even at heavier capital intensity.
  • Share count post-retirement (19.69M), not pre. Most consensus models still use the December 2025 count (20.55M). The April 28, 2026 retirement of 858,163 shares (-4.18%) is a permanent per-share boost the sell-side has not refreshed yet.
Two-sided case — bear anchors
  • Growth never accelerates beyond trailing 1-2%. If FY26-30 organic prints at the 5y trailing average instead of CMD 2-5%, revenue lands at ~€6.9B not €8.2B by 2035 — Contrarian's €110 PT scenario, ~€30/share of equity value lost.
  • Margin reverts to industry median 7.4%. AI commoditises staff-aug consulting pricing; SBC + PPA stay elevated. Year-10 EBIT at 7.4% × €8.2B = €607M instead of €734M — roughly €40/share haircut on terminal value alone.
  • AMF denies the 30% voting-rights derogation. Sopra GMT would face a mandatory takeover trigger; structural overhang on the float persists or worsens. Discount widens 15% → 20-25%, ~€30/share off equity.
  • UK + Benelux contraction proves structural. Berenberg-style cuts persist; Germany/Benelux drag offsets France strength; blended growth halves from 4.5% to 2.5%.
  • Terminal value carries 60.2% of EV. Mechanically, any year-10 margin slip compounds in the perpetuity — the model is more sensitive to the 9.0% margin assumption than the 4.5% growth assumption.

Risks to thesis (tail, not bear case)

AMF denies Sopra GMT 30% derogationHigh

Sopra GMT abstained from the Apr 28 2026 buyback resolutions to stay below the 30% mandatory-takeover trigger pending AMF derogation. If denied, structural overhang intensifies; governance discount could widen to 20-25%, costing ~€30/share.

July 29 H1 print missesMed

The whole base case hinges on the Q1 inflection being real. If H1 organic prints below +3% or FY guide is unchanged/cut, the bear case (~€110 PT) becomes central. Pre-commit the decision rule before the print.

GenAI compresses services pricingMed

Capgemini -29% / 12m and Atos collapse suggest the sector multiple may have a structural ceiling. If staff-augmentation rates compress 5-10%, year-10 margin lands at 7.5% not 9.0% — terminal value loses ~€40/share.

Related-party fee structure expandsHigh

Sopra GMT renewed the "framework agreement" Jan 30, 2025. Any expansion of the fee, or a new related-party transaction discovered in the 2026 URD, would shift the governance question from "structural discount" to "active extraction" — non-trivial re-rating risk.

UK / Benelux contraction persistsMed

Berenberg cut sales/EPS Apr 8 on UK/Benelux drag; 2025 quarters showed ~-2.6%/qtr in Germany, Belgium, Netherlands. If the Q1 France strength masks a worsening Northern-Europe book, blended growth halves.

FX — 23% non-EUR revenueLow

UK 16% + Sweden 3% + Switzerland 4% = ~23% non-EUR. A 10% GBP/SEK/CHF weakness vs EUR compresses reported revenue ~2% — material but not thesis-breaking. Engine reports in EUR; FX not modelled explicitly.

10-year forecast

Revenue €5.90B -> €8.16B over 10 years (4.5% CAGR Y1-5 fading to 2.0% terminal); GAAP operating margin recovers from 8.0% to 9.00% by Year 5 and holds. FCFF grows €245M -> €483M.

Y1 Y2 Y3 Y4 Y5 Y6 Y7 Y8 Y9 Y10 0 8.2B rev (EUR) 0% 10% op margin revenue FCFF op margin

Monte Carlo distribution

Even at the 5th-percentile outcome (€219), intrinsic value exceeds today's €140 price by 36% - across 1000 correlated stress draws on growth, margin, sales-to-capital, beta and ERP, not a single iteration produces an intrinsic below the market price. The disagreement isn't whether Sopra is undervalued; it's by how much.

1000 iterations randomising the central uncertainties (growth, margin, sales-to-capital, beta, ERP). P(intrinsic < market €140.00) = 0.0%.

p5 p25 p50 p75 p95 market 140.00 129.6 272.0 411.0 freq equity / share (EUR)

Post-15% governance: mean €273.62 +/- €35.93/sh, p5 €218.75, p50 €271.97, p95 €335.50.

Cost of capital build
Risk-free rate 2.40%
Mature-market ERP 4.23%
Levered β 0.86
Weighted CRP 0.76%
Cost of equity 6.69%
Pre-tax cost of debt (synth Aaa/AAA) 2.09%
D / V ~21%
WACC 5.31%
Full year-by-year DCF
Year Revenue Op mgn EBIT EBIT(1−t) Reinvest FCFF PV
1 €5.90B 8.05% €475M €361M €116M €245M €233M
2 €6.17B 8.28% €511M €388M €121M €268M €241M
3 €6.45B 8.52% €549M €418M €126M €291M €249M
4 €6.74B 8.76% €590M €448M €132M €317M €257M
5 €7.04B 9.00% €633M €481M €138M €344M €265M
6 €7.32B 9.00% €659M €499M €113M €386M €282M
7 €7.58B 9.00% €682M €514M €102M €412M €285M
8 €7.80B 9.00% €702M €527M €91M €437M €284M
9 €8.00B 9.00% €720M €538M €78M €460M €282M
10 €8.16B 9.00% €734M €547M €64M €483M €278M
Methodology & flags

Damodaran FCFF DCF, 10-year explicit forecast plus perpetuity at 2.00% terminal growth. R&D not capitalised (IT services do not separately disclose internal R&D under IAS 38); IFRS-16 lease liabilities (EUR426M) included in book debt. Synthetic credit rating Aaa/AAA from low leverage. CRP from 10-country revenue-weighted blend (France 46% . UK 16% . Germany 11% plus Italy/Spain/Switzerland/Belgium/Sweden/Netherlands plus India 2% on supplier base). Levered beta = 0.86 from 5Y regression triangulation; the Damodaran Computer Services bottom-up beta re-levers to ~1.27 (32% higher), but Sopra's European IT-services recurring-contract book exhibits lower actual market covariance than the broad cross-section. Monte Carlo: 1000 iterations, 0 failed, with correlated draws. Council pressure-test raised governance discount from 0.10 to 0.15 (lower edge of Bollore/Wendel range). Engine v1.0.0 . result: Finance/Damodaran/valuations/sopra/output/2026-05-25-result.json

  • R&D cap: OFF . Lease cap: OFF (IFRS-16 leases already in book debt) . Failure: OFF . ESO: OFF
  • Governance haircut: 15% applied post-DCF (€417.64 > €354.99)
  • Levered beta override: 0.86 (5Y regression triangulated across StockAnalysis 0.87 / SimplyWallSt 0.85 / Lightyear 0.86); Damodaran Computer Services beta_u 1.07 re-levered to ~1.27 would be 32% higher
  • Sensitivity tornado: not run . Monte Carlo replaces single-variable sweep