COUR — Coursera, Inc. — Research note
Executive summary
Coursera operates a global online-learning platform with two segments — Consumer (subscription/courses) and Enterprise (B2B skills training), serving ~205m registered learners and ~1,729 paid enterprise customers as of Q1 2026. The trajectory over the filings shows growth decelerating from the low-20%s YoY in 2022 to single-digits (9% in 2025, 9% in Q1 2026), while management has aggressively expanded margins, taking Adjusted EBITDA from negative in 2022 to $63.5m (8.4%) in 2025, and pivoted to a transformative all-stock merger with Udemy (approved by shareholders April 2026, HSR cleared Feb 2026). The single most important valuation point today: the combined entity trades at $1.57bn market cap against ~$1.1bn pro-forma cash, giving an EV of ~$0.5bn on ~$1.5bn combined revenue — striking-looking on multiples, but the market is clearly pricing significant skepticism about AI substitution risk, growth deceleration, and integration execution.
Fair value estimate
- Fair value range: $7.50 – $10.50 per share (current pro-forma combined share count basis)
- Implied market cap range: $2,150m – $3,010m (mid: ~$2,580m)
- Methodology: blend of EV/Sales (1.0–1.5x combined revenue ~$1.5bn) and EV/EBITDA (8–12x on combined ~$140m EBITDA), grossed up by ~$1.1bn pro-forma cash, less an integration-risk discount.
- Latest disclosed market cap: $1,569m
- Upside to midpoint: ~+64%
Key assumptions: combined entity revenue ~$1.5bn (COUR $757m FY2025 + estimated Udemy ~$790m); combined EBITDA ~$140m (COUR $63.5m + estimated Udemy ~$80m); pro-forma cash ~$1.1bn (COUR $792.6m at FY2025 + estimated Udemy ~$300m); 286m diluted shares outstanding. The fair value would be ~$9–12/share if you fully credit cash; we narrow it to reflect AI-substitution and integration risk 2025-12-31 10-K; 2026-03-31 10-Q.
Sector context
Confirmed: Software & IT Services / vertical SaaS for online learning. Within the sector, Coursera is below typical SaaS peers on growth (9% vs. mid-teens median for scaled SaaS) but above typical peers on balance-sheet quality (net cash equivalent to ~50% of market cap). Margins (8% EBITDA) trail diversified SaaS but are above pre-profitability scale ed-tech peers. Listed comps: Udemy (UDMY — pending merger partner), Duolingo (DUOL — much higher growth/multiple), Pluralsight (private). The market structure pre-merger is fragmented and AI-pressured.
Investment thesis
- Cash-adjusted valuation provides a substantial cushion of safety. Standalone COUR carries $789.8m cash, no debt, and generated $78.5m of Free Cash Flow in FY2025 2025-12-31 10-K. The combined entity post-merger will have ~$1.1bn cash against a $1.57bn market cap — effectively the market is paying ~$0.5bn EV for ~$1.5bn of combined revenue and a positive-EBITDA business.
- Demonstrated operating leverage as platform matures. Coursera moved Adjusted EBITDA from negative $9.9m in 2023 to $41.5m in 2024 to $63.5m in 2025 — a 53% YoY improvement on 9% revenue growth — while gross margins expanded from 51.9% in 2023 to 54.6% in 2025 on a more favorable content-cost mix 2025-12-31 10-K, MD&A. The 2026 guide of $70–76m EBITDA continues the trajectory.
- Industry micro-credential franchise with embedded AI catalog gives optionality on enterprise reskilling demand. Catalog includes 100+ entry-level Professional Certificates, 1,100+ generative AI courses (15 enrollments/min in 2025 vs. 8/min in 2024), and partnerships with Google, IBM, Microsoft, Anthropic, OpenAI (ChatGPT app) 2025-12-31 10-K and 2025-09-30 shareholder letter. If AI drives net-positive skills demand rather than substitution, COUR has a credible vehicle for capture.
Key risks
- AI substitution risk to core product. ChatGPT/Claude already provide on-demand tutoring, custom course creation, and Q&A — directly competing with structured online courses. The 10-K acknowledges: "providers that leverage AI capabilities... are increasingly offering education-focused capabilities, such as self-paced learning, custom course creation and on-demand tutoring" 2025-12-31 10-K, Competition.
- Growth deceleration and weak enterprise retention. Revenue growth has fallen from 21% (2023) to 9% (2025). Enterprise Net Retention Rate is 90% in Q1 2026 vs. 91% prior — meaning existing enterprise customers are contracting on average; growth is dependent entirely on new logo acquisition 2026-03-31 10-Q, Key Business Metrics.
- Udemy merger integration and dilution risk. All-stock deal closing 2026 will add ~120m shares (from 167.9m to ~287m); $40.5m termination fee risk; the 10-K flags concentration of integration uncertainty, possible loss of key personnel, and culture risk 2025-12-31 10-K, Risk Factors. Three CFO transitions in 18 months (Hahn → Foley interim → Foley permanent) and a CEO transition (Maggioncalda → Hart, Feb 2025) compound execution risk.
Operating leverage
The cost base is predominantly fixed: R&D ($121.6m in 2025, ~16% of revenue), sales & marketing ($255.7m, 34% — partially variable through performance marketing), G&A ($114.4m, 15%). Content costs are the main variable cost (rev-share with creators) and ran at ~36% of revenue, with management actively negotiating lower rev-share arrangements that drove a 110 bp gross-margin improvement YoY 2025-12-31 10-K. Gross profit margin of 55.7% non-GAAP, with content costs as a percentage of revenue declining in both Consumer (38.6% from 40.3%) and Enterprise (30.3% from 31.4%) segments. At this margin structure, a 10–20% upside revenue surprise should drop a high share to incremental EBITDA: rough math at a ~70% incremental contribution margin (gross minus marketing) would convert $80–160m of incremental revenue into $55–110m of incremental EBITDA — i.e. roughly doubling from the current $63.5m base. The flywheel is real but capped because growth is decelerating, not accelerating. Observable inflection points: (i) the platform fee introduced in Jan 2026 2026-03-31 10-Q, (ii) the still-pending Udemy synergies, and (iii) AI-driven content production capability (Course Builder) that lowers content acquisition cost.
Value-trap signals
- Decelerating growth across all periods covered (21% → 9% → guidance 6–8%) — classic value-trap signature.
- Enterprise NRR below 100% (89–93%) indicating existing customer contraction.
- Heavy SBC dilution: $95.1m stock-based comp in 2025 (~12.5% of revenue); share count up 4.9% YoY (160.1m → 167.9m before Udemy dilution).
- Repeated leadership turnover: CEO change Feb 2025, CFO change Oct 2025 (Hahn resignation), interim CFO appointment, then permanent CFO appointment March 2026.
- Related-party transactions: ongoing content-sourcing agreement with DeepLearning.AI Corp, owned by Chairman Andrew Ng ($8.7m in 2025) 2025-12-31 10-K, Note 12.
- Repurchased $95m of stock in 2023–2024 at an average price (~$12) more than 2x the current $5.48 — meaningful capital-allocation misstep.
- Structural AI substitution threat acknowledged in own Risk Factors.
Earnings vs. expectations
Across the eight quarters covered in detail, Coursera has more recently beaten and raised with notable consistency: Q1 2025 ($179.3m vs. ~$170m implied), Q2 2025 ($187.1m beat, raised FY guide by $17m), Q3 2025 ($194.2m beat, raised FY guide $10m to $750–754m), Q4 2025 ($196.9m at the high end of $193–197m guide). Earlier in 2023–2024 the pattern was more mixed, with the 2024 full-year revenue ($694.7m) coming in below original investor expectations. Net pattern: more beats than misses in the last 4–6 quarters, with positive guidance revisions — though against expectations that had already been reset lower.
Conviction
Conviction: 3 (moderate).
- Anchors: clean disclosure with full 10-K and quarterly 10-Q filings; large net cash position is unambiguous; FY2025 financials are recently audited (Deloitte, unqualified opinion); guidance methodology is consistent.
- Limits: (1) the Udemy merger creates a step-change in the entity being valued and the filings provide no Udemy financials, forcing me to estimate the combined business; (2) AI substitution risk is fundamental uncertainty in the business model itself — courses delivered by leading universities/companies may or may not survive the rise of agentic AI tutoring. A wide range is unavoidable.