SoFi and NU Holdings run loan portfolios many times larger than their revenue, while MercadoLibre and Sea Limited keep lending as a sidecar to diversified businesses. This distinction is the single most important structural difference among these four companies. SoFi's loan book stands at nearly 10× its revenue; NU's sits at 1.8×. MercadoLibre and Sea Limited both hover around 0.3×. These ratios reveal fundamentally different business models, risk profiles, and growth strategies — all operating under the loose umbrella of "fintech."
The data below draws from each company's FY2024 results (year ended December 31, 2024), sourced from 10-K and 20-F filings, earnings press releases, and supplemental financial disclosures.
All figures are in USD and reflect the fiscal year or balance sheet date of December 31, 2024.
| Metric | NU Holdings | SoFi Technologies | MercadoLibre | Sea Limited |
|---|---|---|---|---|
| Loan portfolio (gross) | $20.7B | $26.3B | $6.6B | $5.1B |
| Total revenue | $11.5B | $2.68B | $20.8B | $16.8B |
| Gross profit | $5.27B | ~$1.33B* | $9.58B | $7.21B |
| Shareholders' equity | $7.65B | $6.53B | $4.35B | $8.37B |
*SoFi, structured as a bank holding company, does not report a traditional gross profit line. The $1.33 billion figure represents total segment contribution profit (lending + financial services + technology platform), which deducts provision for credit losses and directly attributable expenses from segment revenue. This is the closest available proxy but is not perfectly comparable to IFRS/GAAP gross profit reported by the other three companies.
Definitions of loan portfolio by company:
| Ratio | NU Holdings | SoFi | MercadoLibre | Sea Limited |
|---|---|---|---|---|
| Loans / Revenue | 1.80× | 9.83× | 0.32× | 0.30× |
| Loans / Gross Profit | 3.93× | 19.83×* | 0.69× | 0.71× |
| Loans / Book Value | 2.71× | 4.03× | 1.52× | 0.61× |
*Uses segment contribution profit as proxy for gross profit.
These ratios split the four companies into two clear groups. SoFi and NU are credit-first businesses where the loan book dwarfs revenue, meaning their economic engine is fundamentally balance-sheet-driven. Interest income comprises 84% of NU's revenue and 64% of SoFi's, confirming that credit intermediation is their primary value-creation mechanism. MercadoLibre and Sea Limited are platform businesses with lending bolt-ons — their loan portfolios represent roughly one-third of a single year's revenue, and fintech segments contribute just 41% and 14% of total revenue, respectively.
The loans-to-book-value ratio is especially telling for credit risk. SoFi's 4.03× ratio means its loan portfolio is four times its equity cushion, typical of a leveraged bank model funded primarily by deposits ($26B in deposits supports the $26.3B loan book). NU's 2.71× reflects meaningful leverage but with more equity headroom and capital adequacy ratios well above regulatory minimums (18.1% in Brazil). Sea Limited's 0.61× means its equity base alone could absorb the entire lending portfolio — an extremely conservative posture. MercadoLibre at 1.52× sits in between, with its credit book growing fast (74% YoY) but still manageable relative to equity.
SoFi's 9.83× loans-to-revenue ratio is by far the highest in this group and reveals a structural reality: SoFi generates relatively modest revenue per dollar of credit deployed. This is characteristic of a bank holding diversified, lower-yielding assets (student loans carry lower rates than NU's Brazilian consumer credit or MELI's merchant cash advances). SoFi's $26.3B portfolio generates $2.68B in total net revenue, implying an effective portfolio yield well below NU's.
NU Holdings, by contrast, achieves $11.5B in revenue on a $20.7B portfolio — a loans-to-revenue ratio of just 1.80×. This reflects the extraordinarily high yields on Brazilian consumer credit (where annualized rates on revolving credit cards routinely exceed 100%). NU's gross profit margin of 45.6% is remarkably strong for a bank, driven by these high yields and the company's low-cost digital distribution model. NU posted $1.97B in net income for FY2024 with a 29% ROE, dramatically outperforming SoFi's $499M net income (which included a $265M one-time tax benefit).
SoFi's model depends on loan origination volume and gain-on-sale economics alongside net interest income. The company originated $23.2B in new loans during FY2024 and actively sold or securitized portions of its book, generating $256M in origination and sales revenue. This originate-to-distribute model means the on-balance-sheet figure is somewhat fluid — the portfolio would be even larger without active sales.
MercadoLibre and Sea Limited share nearly identical loans-to-revenue ratios (0.32× and 0.30×) and loans-to-gross-profit ratios (0.69× and 0.71×), but their lending trajectories differ meaningfully. MELI's credit portfolio grew 74% year-over-year to $6.6B, with credit cards alone surging 118%. The company is aggressively scaling Mercado Crédito and now holds a $1.7B allowance for doubtful accounts — a sign of both portfolio maturity and meaningful credit risk in Latin American markets.
Sea Limited's $5.1B lending book grew 64% YoY with an impressively low 1.2% NPL rate (90+ days past due), suggesting conservative underwriting standards across Southeast Asian markets. SeaMoney generated $2.37B in revenue and $657M in segment operating income in FY2024, making it a highly profitable but still secondary business line behind Shopee's $12.4B e-commerce revenue. With $10.4B in cash and investments and the lowest loans-to-equity ratio in the group (0.61×), Sea Limited has enormous balance sheet capacity to scale lending if it chooses.
MercadoLibre faces a more constrained equity position. Its $4.35B in shareholders' equity is the smallest in the group, and its loans-to-book-value ratio of 1.52× is climbing as the credit portfolio scales rapidly. The company's accumulated other comprehensive loss of $920M (largely from Latin American currency depreciation) has compressed book value. While $20.8B in annual revenue provides ample earnings power to absorb credit losses, the equity cushion bears monitoring as Mercado Crédito grows.
The comparison reveals a spectrum from pure-play bank to platform-with-lending, and the ratios quantify where each company sits.
The core insight from this analysis is that loan-portfolio-to-revenue ratios differ by more than 30× between SoFi (9.83×) and Sea Limited (0.30×), making "fintech" a nearly meaningless label for grouping these companies. SoFi and NU are banks in all but branding — their economics are driven by net interest margins on large loan books carried against depositor and equity capital. MercadoLibre and Sea Limited are technology platforms that happen to lend, and their lending units, while growing rapidly, remain modest relative to their overall economic footprint. For investors, the practical takeaway is that credit cycle sensitivity varies enormously within this group: a deterioration in consumer credit quality would be existential for SoFi's model, painful for NU, manageable for MercadoLibre, and barely noticeable for Sea Limited.