#GoldmanSachsBullishOnCXMT

Goldman Sachs has put ChangXin Memory Technologies, or CXMT, firmly into the global semiconductor conversation.

In an August 23 research report, Goldman initiated coverage with a Buy rating and a CNY 129 12-month price target, building its thesis around three major forces: AI-driven memory demand, aggressive capacity expansion and China’s push toward greater domestic DRAM supply.

The capacity numbers are what make this story particularly significant.

Goldman estimates CXMT’s monthly wafer capacity could increase from approximately 270,000 wafers in 2026 to 447,000 in 2028 and 665,000 by 2030. That would represent more than a doubling of monthly capacity over four years if the expansion is executed as projected.

The report also estimates that CXMT could eventually have the capability to supply around 50% of China’s DRAM demand by 2028. That figure should be understood as potential future supply capability, not current market share. The company still has to scale production, improve yields, qualify products and secure customers.

Then there is HBM.

Goldman expects HBM to become a much larger part of CXMT’s business, with its revenue contribution potentially rising from around 2% in 2026 to 27% by 2030. That is important because AI accelerators rely heavily on high-bandwidth memory, making HBM one of the most strategically valuable areas of the memory market.

But this is also where the execution challenge becomes much bigger.

HBM is not simply conventional DRAM with a different label. It requires advanced manufacturing, packaging, high yields and demanding customer qualification. Goldman’s long-term valuation therefore depends on CXMT successfully moving further up the memory technology chain.

CXMT’s IPO already showed how aggressively the market is pricing the China-memory narrative. The company priced its Shanghai STAR Market offering at CNY 8.66 and opened at CNY 49.50 on July 27, a gain of more than 470% from the IPO price. The offering raised roughly CNY 57.9 billion, making it one of the most significant semiconductor listings on China’s domestic market.

That extraordinary debut creates an important second side to the story.

The bullish case is built around AI demand, memory shortages, Chinese semiconductor self-sufficiency and CXMT’s potential production expansion.

The risk case is valuation.

A company can have a strong technological and strategic outlook while its stock price already reflects a large portion of that future success. Goldman’s model itself assumes several things go right simultaneously, including continued favorable DRAM pricing, higher production yields, successful capacity expansion, stronger product mix and substantial margin improvement. Goldman reportedly models gross margin rising from roughly 41% in 2025 to 82% by 2030.

That makes the next few years more important than the headline target.

Investors and industry observers will be watching whether CXMT can turn capital expenditure into actual wafer output, whether yields improve as capacity scales, whether customers adopt its higher-end products and whether the global memory cycle remains strong enough to support attractive pricing.

There is also a geopolitical layer.

China is pursuing greater semiconductor self-sufficiency, while restrictions surrounding advanced semiconductor equipment and technology continue to influence the industry. CXMT’s progress therefore matters beyond one company: successful expansion could strengthen China’s domestic memory ecosystem and increase competitive pressure on established DRAM suppliers.

The global memory market could consequently face an interesting two-stage dynamic.

Short term: AI infrastructure is supporting extraordinary demand for memory.

Long term: Expanding Chinese capacity could increase supply and intensify competition, particularly in conventional DRAM.

That distinction is crucial when thinking about HBM versus mainstream memory. Conventional DRAM could experience greater competitive pressure as capacity expands, while advanced HBM remains a much higher technical hurdle.

So Goldman’s CNY 129 target is best understood as a long-term execution thesis, not a guarantee of future performance.

The numbers tell the story:

CNY 129 — Goldman’s 12-month target
270K → 665K — projected monthly wafer capacity from 2026 to 2030
~50% — potential share of China’s DRAM demand capability by 2028
2% → 27% — projected HBM revenue contribution from 2026 to 2030

CXMT is becoming more than a domestic memory supplier. It is increasingly a test case for whether China can build a large-scale, competitive DRAM ecosystem while AI simultaneously pushes global memory demand to new levels.

Goldman has outlined the potential destination.

Now the semiconductor industry gets to watch whether CXMT can execute the journey.