China-made Loongson 12-core chip is approximately three times slower than six-core Ryzen 5 9600X — 3B6000 hampered by low clock speeds in Linux benchmarks

Benchmarking by Phoronix shows Loongson's 12-core 3B6000 (LA664) desktop CPU, clocked at ~2.5GHz, performs roughly three times slower than AMD's six-core Ryzen 5 9600X across most Linux workloads, though it matches or approaches high-end chips in a few niche tests (C-Ray 2.0, OpenSSL). The vendor's claimed IPC is comparable to Zen 3, but low clock speeds materially limit real-world performance; Loongson is developing an LA864 architecture targeting ~3.0–3.5GHz to narrow the gap with Intel Raptor Lake-class CPUs. For investors, these results underscore a substantial competitiveness deficit for China-made consumer x86 alternatives today, implying limited near-term commercial threat to Intel/AMD dominance absent meaningful architecture and clock-speed improvements.

Analysis

Market structure: This result confirms incumbents (AMD, INTC) retain near-term product and pricing advantages — expect Western CPU vendors to keep >80% share of high-performance PC/server markets outside China for the next 12–24 months. Loongson’s 3B6000 shows core-count ≠ competitive performance when clocks and ecosystem are weak, so Chinese domestic OEMs will likely reserve Loongson for security-sensitive, low-to-mid performance segments, preserving ASPs for AMD/Intel in premium tiers.

Risk assessment: Main tail risks are geopolitical policy shifts — a Chinese mandate to substitute domestically could re-route 1–5% of AMD/Intel revenue within 1–2 years, and a successful LA864 launch (3–3.5GHz) could halve that timeline. Hidden dependencies include software/compiler optimization and board/memory configs (bench variability), meaning one-off benchmarks aren’t reliable predictors; watch official LA864 silicon/bench releases over 3–18 months as catalysts.

Trade implications: Short-term (days–months) this favors selective long exposure to AMD (AMD) and modest INTC exposure while avoiding China-specific hardware plays; longer-term (6–24 months) consider small thematic longs in semiconductor equipment (ASML/LRCX) if China scales fabrication investment despite export controls. Use options to skew risk: favor defined-risk bullish structures to capture upside without one-way exposure to geopolitical risk.

Contrarian angle: Consensus overweights the narrative of an inevitable Chinese CPU catch-up; the missing point is systems/ecosystem inertia (OS, compilers, OEM qualification) which takes years and favors incumbents. If LA864 fails to deliver demonstrable parity within 12 months, re-rate China’s domestic CPU TAM to sub-10% of global high-performance demand for the next 3 years.

Key Decisions for Investors

• Establish a 2–4% portfolio overweight in AMD (AMD) via equity or 6–12 month exposure; size to 2% if portfolio concentrated, add if shares decline >8% on headline noise. Rationale: preserve exposure to secular demand and near-term performance moat.
• Initiate a defined-risk bullish options position on AMD: buy 3‑month call spread at ~10%/30% OTM (sell higher strike) sized to 0.5–1% of portfolio to capture upside into product cycles while limiting downside. Close or roll at expiry unless LA864 shows credible performance vs Zen 4/5.
• Add a tactical 1–2% long in Intel (INTC) funded by selling 1–3 month covered calls (collect premium) to improve yield; trim if company lowers China revenue guidance by >10% or if AMD materially outperforms consensus guidance.
• Establish a 0.5–1% long in semiconductor equipment exposure (ASML or LRCX) for 6–18 months to profit from potential Chinese onshore capex; increase allocation only if China announces >$5B/year domestic fab incentive or LA864-related procurement programs within 6 months.