Kanzhun (BZ) Q4 2025 Earnings Call Transcript

Kanzhun reported Q4 revenue RMB 2.08 billion (+14% YoY) and FY2025 revenue RMB 8.27 billion (+12.4% YoY) with adjusted operating profit of RMB 900 million in Q4 (+37% YoY) and a record FY adjusted operating margin of 40.8% (Q4 43.3%). Net income was RMB 2.7 billion, adjusted net income RMB 3.6 billion (+33% YoY), cash of RMB 19.9 billion, and the board approved a three-year return policy allocating at least 50% of adjusted net income to dividends/buybacks while increasing the repurchase authorization to USD 400 million. Management flagged Q1 revenue guidance of RMB 2.05–2.085 billion (+6.6–8.4% YoY) affected by later Chinese New Year, highlighted accelerating closed-loop AI revenue (hundreds of millions RMB) and strong SME/blue-collar user growth, while noting higher tax and a one-off intangible-asset impairment as risks.

Analysis

Kanzhun's AI push creates an asymmetric margin lever: small, cheap models that outperform larger peers on targeted tasks materially lower inference cost and enable margin-dense, outcome-based pricing for placement services. If management can scale closed-loop placements from “hundreds of millions” to low billions within 12–24 months, incremental revenue will flow almost entirely to the bottom line because the delivery is software/AI-driven and labor-light — meaning current margins can expand further even with incremental SG&A for go-to-market.

The structural move into SMEs, blue-collar, and lower-tier cities mutes churn and ARPPU cyclicality; customers are numerous, ticket sizes are small, and renewals are sticky — a profile that favors unit-economics predictability and makes pay-per-hire or subscription-plus-success-fee products viable at scale. The board’s explicit payout linkage and larger buyback authorization change the capital structure calculus: if management uses repurchases opportunistically, EPS sensitivity to modest revenue growth becomes magnified over 12–18 months.

Key tail risks are tax and regulatory regime changes (OECD top-up and potential AI/data restrictions) and product risk if rivals replicate small-model gains or if marketplaces push back on AI agents in hiring. Near-term seasonality noise (calendar shifts) can create buying windows; the real re-rating catalyst will be observed multi-quarter acceleration of closed-loop placements and evidence of repeatable unit economics for pay-for-hire contracts.

Key Decisions for Investors

• Core long: Buy BZ stock on any >10% post-earnings pullback. Horizon 12–18 months. Thesis: closed-loop revenue scaling + buybacks → 30–60% upside; tail risk: regulatory/top-up tax could produce 30% drawdown. Stop-loss 30% below entry.
• Options overlay: Sell 3–6 month cash-secured put ~10% OTM to collect premium and set a lower entry price; if assigned, cost basis reduces expected downside and you own shares for the 12–18 month thesis.
• Volatility capture: If implied vols spike on China/seasonality headlines, buy a 12-month BZ call spread (buy ATM, sell +30% strike) to limit cost while keeping upside participation; target 2:1 reward/risk if closed-loop adoption accelerates.
• AI hedge: Size a small long position in NVDA calls (6–12 month) as a convex hedge — if AI demand re-accelerates broadly, NVDA upside offsets valuation compression risk for BZ driven by macro/AI sentiment shocks.