China's Chip Industry Profits Surge 18.5x in First 7 Months; Goldman Sachs Forecasts 66% Advanced Node Self-Sufficiency by 2035
Nashnova编辑部
China's IC industry posted an 18.5× year-on-year profit surge in the first seven months of 2026, driven by AI-demand tailwinds; Goldman Sachs raised its 2030 capex forecast to $82 billion and projects advanced-node self-sufficiency leaping from 8% to 66% by 2035 — with yield catch-up as the make-or-break variable.
Profits up 18.5× — where is the money coming from?
China's National Bureau of Statistics reports IC-industry profits rose 18.5× year-on-year in January–July 2026, accounting for over 80% of the electronics sector's profit growth.
This means → the profit engine of China's electronics industry has shifted from assembly and finished devices to the chips themselves.
The driver is AI compute demand: both compute and memory chips saw volume and price gains simultaneously. The broader electronics sector's profit doubled (+1.1×), contributing 9.3% of profit growth across all large industrial enterprises nationwide.
Who else along the supply chain is cashing in?
Downstream surged hardest: computer manufacturing profits rose 3.3×, peripherals 2.5×, and industrial-control computers 1.6×.
Upstream kept pace: electronic specialty materials grew 226.8%, discrete semiconductors — the simplest individual components inside a chip — rose 45.8%, and electronic-circuit manufacturing climbed 37.1%.
In plain terms = the chip profit boom is spreading in both directions along the supply chain — materials makers and system builders are all eating.
Why is Goldman sharply raising its bet on China semis?
Goldman lifted its China 2030 semiconductor capex forecast to $82 billion, 79% above its estimate from a year ago.
Two reasons: generative-AI demand keeps expanding + semiconductor localization is accelerating.
This means → Goldman sees China's chip investment not as a short-term pulse but as a long runway stretching at least to 2030.
Can advanced-node self-sufficiency really hit 66% by 2035?
Goldman projects China's domestic supply of ≤7 nm advanced-node wafers — the silicon slices used to make the most cutting-edge chips — will grow at a 46% CAGR from 2025 to 2035, far outpacing 17% demand growth over the same period.
By 2035, domestic monthly capacity is forecast to reach roughly 410,000 wafers, covering 66% of estimated monthly demand of ~619,000 wafers. In 2025, that self-sufficiency rate was just 8%.
In plain terms = from "almost entirely imported" to "two-thirds homemade" in a decade — that is Goldman's most optimistic trajectory.
Yield and lithography — where are the two hardest bottlenecks?
Goldman forecasts China's advanced-node yield — the share of usable chips per production batch — rising from 23% in 2026 to 50% in 2030 and 75% in 2035.
TSMC's 7 nm yield already exceeds 90%, per ICsmart data. This reflects that even on Goldman's most optimistic path, the yield gap between China and TSMC will not fully close by 2035.
Lithography remains the biggest chokepoint: China relies heavily on ASML's deep-ultraviolet (DUV) tools, while U.S. export controls block access to extreme-ultraviolet (EUV) equipment — the higher-precision machines essential for the most advanced nodes — capping the speed of yield catch-up.
70% self-sufficient by volume — why is the value share so much lower?
On equipment: Goldman expects domestic wafer-fab equipment suppliers' market share to rise from 31% in 2026 to 38% in 2028.
Memory is another pillar: CXMT (DRAM) and YMTC (NAND flash) are expanding rapidly through the current memory upcycle.
By shipment volume, China's domestic chip output as a share of demand has climbed from 38% in January 2010 to roughly 70% by June 2026 — but by value, the ratio remains significantly lower. This means → the bulk of what China produces is mid-to-low-end; high-value advanced-node chips still depend heavily on overseas suppliers — whether yield improves along Goldman's forecast path is the core variable for judging if the 2035 target is achievable.
市场有风险，内容仅供研究参考，不构成投资建议。