Kanzhun (BZ) Q4 2024 Earnings Call Transcript

Kanzhun reported FY2024 GAAP revenue of RMB 7.36 billion, up 24% y/y, and Q4 revenue of RMB 1.82 billion, up 15% y/y; adjusted net income was RMB 2.71 billion for the year (+26% y/y) and RMB 720 million in Q4 (+15% y/y). Operational metrics strengthened: MAU 53.0 million (+25.3% y/y), paid enterprise customers 6.1 million (+17.3% y/y), full-year adjusted operating income RMB 2.32 billion (+41% y/y) and adjusted operating margin 31.5% (Q4 margin 36.1%); the company repurchased $229 million of stock in 2024 (reducing shares by 3.7%). Management set Q1 2025 revenue guidance of RMB 1.90–1.92 billion (+11.5–12.7% y/y) and a 2025 non-GAAP operating profit target of RMB 3.0 billion (+30% y/y), while highlighting substantial AI deployments (Nanbeige, DeepSeek-R1) and strong AI-driven engagement (>200M AI-generated resumes, AI-related job postings +60% y/y).

Analysis

Kanzhun’s AI push is less about a one-time product lift and more about permanently lowering marginal cost of matching and outreach; that implies operating leverage can continue even if top-line growth reverts to trend. The real optionality lies in recruiter-side automation: if AI agents sustain conversion uplifts, pricing power shifts from pure ad/visibility to outcome-based and workflow subscriptions, which compresses CAC and lengthens LTV.

A distinct second‑order risk is “signal pollution” from mass AI‑generated resumes and JDs — this degrades match quality and could force the market toward stronger provenance/verification primitives (proof-of-personhood, authenticated interview logs) or new paid validation services. That outcome would create a new adjacent TAM (third‑party verification, test/assessment providers) while penalizing platforms that can’t defend data quality.

Competitively, open‑source large models lower the barrier to entry for mid‑sized rivals and reduce differentiation from pure model capability; Kanzhun’s durable edge becomes verticalized behavior data, AI‑specific workflow integrations for blue‑collar recruiters, and execution (buybacks + tight cost control) that supports valuation multiple resilience. However, regulation (AI content rules, data privacy enforcement) or a meaningful drop in match accuracy are binary downside catalysts that could compress monetization within months.

Net: the setup favors a tactical overweight in the equity to capture margin expansion and buyback support, but only with active downside protection — the same AI tailwind that powers efficiency also creates a plausible self‑inflicted product quality shock if left unmanaged.

Key Decisions for Investors

• Long BZ (ADS), size 1–2% NAV, 6–12 month horizon — rationale: capture continued margin re‑rating and buyback support as AI features monetize; plan to scale into strength and take partial profits on a 30–40% move higher. Risk: buy protective puts or size to withstand a 20–30% drawdown if regulation or AI‑noise hits matching quality.
• Staged call spread on NVDA, 9–12 month expiry (buy near‑the‑money calls / sell higher strike) — rationale: asymmetric way to express continued enterprise demand for GPUs/inference hardware driven by customers like Kanzhun; target 2–3x payoff if AI adoption forces accelerated capex. Tail risk: macro slowdown or inventory cycles that compress GPU pricing.
• Protective hedge: buy BZ 3–6 month puts (approx. 15–25% downside protection) sized to offset 50% of long position — rationale: insures against rapid deterioration in match quality or China‑specific regulatory action which are high‑impact, short‑dated catalysts.