AlphaLens
Research
首覆Equal-weightTP $95.00005月3日 · Morgan Stanley

Empyrean Technology: Growth Reacceleration Hinges on Closing Digital IC EDA Gaps

301269.SZ
中文EN⚠ quality lint: see notes

Empyrean Technology: Growth Reacceleration Hinges on Closing Digital IC EDA Gaps

Core Thesis

Empyrean Technology's 2025 revenue growth of 8.4% materially lagged the domestic EDA market narrative. The fundamental driver was a 38% sales decline from Huawei, its largest customer, to Rmb248mn. Non-Huawei sales grew 31%, but the Huawei drag capped total growth. China EDA market share crept only 0.4ppt to 9.7% in 2025. The company achieved ~80% digital IC EDA tool coverage, but two critical missing tools – design-for-test (DFT) and placement-routing (P&R) – block full-flow capability. Management targets acquiring university teams to commercialize these by year-end 2026. Until full-flow is proven, revenue reacceleration remains contingent, not given.

Evidence Chain

  • Revenue mix reveals fragility. 2025 total revenue: Rmb1,325mn. Huawei accounted for Rmb248mn (18.7% of sales), down from ~Rmb400mn in 2024. Non-Huawei revenue grew 31% to ~Rmb1,077mn, but aggregate growth was only 8.4%. By contrast, Cadence's China revenue grew 19%, while Synopsys' China revenue fell 18% (largely due to export controls). Empyrean underperformed the local market leader in a year of accelerating localization.

  • Tool coverage gap is the binding constraint. Empyrean has launched digital IC tools for simulation and sign-off in early 2026, reaching 80% coverage. Two tools remain: DFT (critical for scan chain insertion and test pattern generation) and P&R (the core placement-and-routing engine that determines chip physical layout). Without these, Chinese digital IC designers cannot run a full tape-out flow on Empyrean tools. Management's plan to acquire university teams is realistic but execution-dependent; a 12-month delay would push full-flow to 2027.

  • Localization trend is policy-driven, not derailed by AI. Cadence is aggressively embedding agentic AI into EDA platforms. Empyrean lags in AI features. However, China's push for domestic EDA is politically mandated. Local GPU design houses shifting production back to China are co-developing PDKs with Empyrean and domestic foundries. This provides a non-discretionary revenue base that is less sensitive to feature parity.

Key Risks and Debates

  • Huawei concentration risk persists. In 2025, Huawei still contributed ~19% of sales. Any further decline (due to Huawei's own revenue volatility or shift to other tools) could suppress growth beyond 2026. Non-Huawei growth of 31% is strong but cannot fully offset a 38% cut from a single customer.

  • Full-flow timing slippage. The company's 80% coverage already includes complex tools; completing DFT and P&R is nontrivial. If commercialization extends into 2027, the expected 2028 revenue reacceleration (forecast at 27% growth) would be pushed out.

  • Valuation premium is extreme. At 28x NTM P/S, Empyrean trades at 2-3x the multiples of Cadence (15x) and Synopsys (9x). This premium assumes full-flow success and sustained market share gains. If the tool gap persists, multiple compression could be severe.

Valuation or Trading Implications

The analyst cut 2026/27 EPS estimates by 63%/39%, reflecting lower sales and elevated R&D spending. Price target reduced to Rmb95 (from Rmb115), implying ~7% upside from current Rmb88.82. The Equal-weight rating signals fair risk-reward. The bull case (Rmb130) assumes 30% revenue CAGR and >20% China EDA market share by 2030; the bear case (Rmb55) assumes stagnant ~10% share. For now, the market is pricing in full-flow success. Any negative news on tool development timeline would trigger de-rating. Entry should wait for evidence of DFT and P&R commercialization.

Related (同 ticker)