Open Lending reported Q4 2025 earnings on March 12. The stock ripped 20% to $1.45 the next session. Even after the move, this trades at 2.9x guided EBITDA with $176M in cash. The question is whether you're buying a turnaround or a value trap.

After reading the full transcript and cross-checking every material claim against ALLY, COF, KMX, CACC, and CPSS earnings calls, here's what's real and what's borrowed.

The Setup

LPRO is an asset-light auto lending enablement platform. They don't lend — they provide pricing, analytics, and insurance facilitation to credit unions making near-prime auto loans. Revenue comes from program fees (per-loan technology fee) and profit share (72% of insurance underwriting profit on the loans they enable).

The company nearly imploded in 2024 when a $96.1M profit share write-down on 2021-2023 loan vintages crushed reported revenue to $24M. That write-down is over (+$0.4M change-in-estimate in 2025 vs -$96.1M in 2024). FY2025 revenue recovered to $93.2M, gross margin to 77%, and adjusted EBITDA turned positive at $15.6M.

The balance sheet is the floor: $176.6M cash, $85.1M debt (term loan due September 2027), $150M undrawn revolver. Net cash $91.5M, or $0.78/share. Gross cash per share is $1.50. Stock at $1.45 is above net cash but below gross cash — the operating business is valued at roughly $0.67/share, or $78.5M enterprise value.

Six insiders bought open-market throughout 2025 at $1.50-$2.50, including the CFO putting in $620K+.

What the Call Revealed

Three data points matter. Everything else is noise.

1. CERT pace fully recovered. Management gave specific daily metrics: 353 certified loans per business day since February 1, matching the 60-day average before the Q4 disruption. During the impacted period (mid-November through January 16), pace was 293/day. Applications are running +20% YoY through February. The Q4 miss — certified loans dropped from 26,065 to 19,308 — was a self-inflicted pricing mistake (rate changes implemented without elasticity modeling), not structural decline. It's fixed.

2. 2025 vintage outperforming model. 60-day delinquency is running 200bps better than 2023-2024 vintages at the same seasoning point. Management books 2025 loans at 72.5% implied loss ratio but expects actual performance in the mid-60s. That ≈750bp gap is CIE optionality: if mid-60s materializes, positive change-in-estimate adjustments flow back as revenue in 2027-2028. Nobody is modeling this.

3. First guidance since 2022. Management provided full-year 2026 guidance for the first time in four years: 100,000-110,000 certified loans, $25-29M adjusted EBITDA, Q1 at 21,000-22,000 loans. The CFO said free cash flow would be "relatively in line with EBITDA guidance." At $25-29M FCF on a $170M market cap, that's a 15-17% FCF yield.

The call also revealed: zero customer losses in Q4 (four for the full year, 46 new logos added), credit builder borrowers (≈30% of incoming applications, currently shut off) being re-enabled in Q2 via a new pricing solution called Project Red Rocks, and early-stage traction on Apex One (prime auto decisioning platform — 2 paying customers, not material yet).

Only three analysts were on the call. No one asked about profit share per loan trajectory, insurance partner health, or competitive dynamics. Thin coverage is the information asymmetry opportunity — and the risk.

What's Real vs. What's Borrowed

This is where the cross-ticker work pays off.

The 2025 vintage improvement is mostly industry beta. ALLY beat the low end of its NCO guidance (1.97% vs 2-2.25% range). Capital One's auto charge-offs are "stable at pre-pandemic levels," delinquency down 72bps YoY. KMX explicitly separates "2024 and 2025 post-contraction vintages continue right in line with expectations" from 2022-2023 vintages still generating $71M in additional losses. Even deep subprime CPSS shows 2025 vintage recovery rates at 43.4% vs 20.5% for 2022.

Every lender that tightened underwriting in 2022-2023 is seeing newer vintages perform better. The structural drivers — normalized used car values (Manheim stable at 205.5), tighter post-COVID underwriting, improved recovery rates — benefit all participants. LPRO's 200bps improvement is consistent with the industry pattern, not exceptional.

So where's the idiosyncratic alpha? Three places:

Conservative booking creates CIE optionality. LPRO books at 72.5% loss ratio when they expect mid-60s. That conservatism — adopted after the 2024 catastrophe — means future positive CIE adjustments are likely as 2025 vintages season. Competitors don't have this same buffer because they didn't have the same write-down forcing them into extreme conservatism.

Customer acquisition against a declining market. NCUA data shows credit union auto lending declined $6.5B (-1.3%) in H1 2025 despite CU profitability up 31.5% and loan-to-share ratio declining (capacity exists). CUs have the capital but are choosing mortgages and commercial over auto. LPRO is gaining share — 46 new logos, zero Q4 losses — in a flat-to-declining pool. That's execution, not cycle.

Project Red Rocks has no disclosed competitor equivalent. A real-time simulation engine that models rate and credit box changes before implementation addresses the exact operational failure that caused Q4. On time, on budget, preliminary benefits rolling quarterly.

Factor Decomposition

Statistical regression: 93% idiosyncratic variance, 7% market beta, R² = 7%. For context, ALLY is 25% idio, COF 26%, CACC 35%, CPSS 47%. LPRO is dramatically more company-specific than every auto lending peer. The auto credit cycle moves ALLY and COF. It doesn't move LPRO. LPRO moves on LPRO.

The thesis decomposes into seven independent factors:

Aggregate idiosyncratic exposure: ≈80%, above the 75% threshold. This is a legitimate idiosyncratic bet, not a disguised factor play.

Aggregate edge exposure: ≈60%. The edge concentrates in granular volume data the market doesn't track, CIE optionality math nobody is modeling, and information asymmetry from thin institutional coverage.

What the Market Implies

At $1.45, the market prices the operating business at $78.5M EV — 2.9x guided EBITDA. Specialty finance peers trade at 5-12x. An asset-light 77% gross margin business with a $150M undrawn revolver doesn't typically get a sub-3x multiple unless the market expects the earnings to evaporate.

Extracting implied probabilities from a three-scenario framework:

The market assigns roughly 4x the probability to the bull case that we do in reverse — we see 2.5x the bull probability the market implies, and the market sees 1.6x the bear probability we assign. The gap is a 15pp divergence on both tails.

Probability-weighted EV: $1.80. From $1.45, that's 24% raw return, roughly 10% annualized idio alpha after edge adjustment.

The Bear Case

This isn't a table-pounding screaming buy. The bear case is real:

Program fees have declined three consecutive years ($64.1M to $57.3M to $54.3M). Management said nothing about fee rate stabilization on the call. No analyst asked. If the per-loan fee rate keeps falling, volume recovery may not save program fee revenue.

Management proved they can self-inflict damage. The Q4 pricing mistake — implementing rate changes without elasticity modeling — cost roughly 7,000 certified loans and cratered the quarter. Project Red Rocks is the guard rail being built, but it's not complete. What's the next unforced error?

Full valuation allowance on $84.4M in deferred tax assets. Management's own auditors assessed it "more likely than not" the DTAs won't be realized. This is the company's tax team saying they don't expect sustained profitability. If the turnaround works, the VA release is a potential $50-84M one-time benefit. But its existence is a bearish signal on internal expectations.

Subprime auto ABS delinquency is at an all-time record — 6.18% 60+ DPD in December 2025. LPRO operates in near-prime, not deep subprime, and 2025 vintages are performing well. But if macro deteriorates, near-prime follows subprime.

Insurance partner concentration is an existential risk that got zero airtime on the call. Three active partners, one at 24% of revenue. No analyst asked about partner health, satisfaction, or renewal. Earned premium declined to $318.4M from $336.9M. We don't know why.

Entry and Timing

Don't chase the 20% rip. The stock moved from $1.21 to $1.45 in a single session on 2.7x average volume. That's discovery, and discovery can retrace.

The next information event is Q1 2026 earnings on May 7 — 53 days away. No reason the stock can't drift to $1.25-1.35 in the interim as momentum traders take profits.

Preferred entry: $1.25-1.40 (pullback from the rip). Starter size: 1.5-2%. Add on Q1 confirmation if certified loans land within the 21-22K guide and profit share per loan holds above $320. Full position (3-4%) after Q2 confirms credit builder re-entry impact and Red Rocks effectiveness.

Hard cut: Q1 certified loans below 20,000. Management just committed to specific guidance for the first time in four years. Missing Q1 after that commitment kills the credibility thesis.

Conclusion

LPRO is a moderate-alpha turnaround with genuine downside protection (cash) and genuine execution risk (management). The market prices 10% odds on the bull case; we see 25%. That 15pp gap, combined with 93% idiosyncratic variance and thin institutional coverage, makes this a textbook micro-cap information asymmetry play.

The vintage improvement is mostly beta — every auto lender is seeing it. The real alpha is in LPRO's conservative booking discipline creating CIE optionality, share gains in a declining CU auto market, and a simulation engine that addresses their biggest operational weakness.

At $1.45 post-rip, the asymmetry is compressed but still positive EV. Wait for the pullback. Q1 earnings on May 7 is the proof point. If they hit their own guidance, this re-rates. If they miss, you'll be glad you waited.

Evidence