In less than three months, ChangXin has made two decisions that surprised the market in succession.
Recently, reports emerged that ChangXin Memory Technologies refused to quote a price to Apple for DRAM (Dynamic Random Access Memory, a type of memory chip) supply demand; and just in August, Apple pressured ChangXin for price cuts on LPDDR5X mobile DRAM supply, which likewise failed to reach an agreement.
In the consumer electronics supply chain, let alone refusing Apple twice in a row, even refusing just once is a rare thing.
Over the past two decades and more, accepting Apple's harsh and forceful bargaining power has always been the tacit admission ticket for suppliers to squeeze into the "Apple chain."
In the new round of price pressure in 2026, Apple demanded that Samsung Display and LG Display lower the supply price of OLED panels for the iPhone 18 Pro series to $70. By comparison, the previous procurement price for iPhone 17 Pro Max panels was about $110 to $120.
Although Apple is quite forceful, many suppliers still yearn to enter the "Apple chain." After all, reaching cooperation with this company means the prospect of securing large orders in the coming years, while also obtaining the qualification endorsement of a high-end supply chain.
But the foundation of Apple's power is being rewritten.
American economic historian Chris Miller asserted in "Chip War" that in the computing power era, whoever controls chip production capacity holds a new type of power.
The DRAM track is precisely a footnote to this logic; it is no longer a cold line of cost on a bill of materials, but a solid bargaining chip on the table. That is why ChangXin today has the confidence to say no to Apple.
However, this game has only just begun.
Why Apple Was Refused Twice by ChangXin Over the Same Matter
The two rounds of maneuvering between Apple and ChangXin in the second half of this year, although different in the process of contact, share the same underlying nature.
In both contacts, Apple hoped to obtain lower quotes from ChangXin than from other companies. According to Wccftech and multiple industry chain media reports, Apple's expected price level was lower than that of the three major original manufacturers: Samsung, SK Hynix, and Micron.
Judging from ChangXin's two replies, its attitude also gradually hardened.
In August, ChangXin replied to Apple that the quote could not be lower than that of the Korean companies, and that some models using self-developed processes would even be priced higher. Although it was a refusal, ChangXin still left room for negotiation.
This time, ChangXin simply shut down the quotation channel and would not play along. At almost the same time, a DDR5 memory order from a Taiwan-funded server manufacturer was also turned away by ChangXin.
Some may wonder, since the August round did not reach an agreement, why Apple would initiate another inquiry.
First, Apple's supplier evaluation is inherently a long-term action; one failed negotiation will not make it give up continuously probing ChangXin.
Moreover, as AI servers continue to siphon off global DRAM capacity, global memory prices keep rising, and Samsung and Hynix have successively raised contract prices. According to statistics from The Wall Street Journal, since Google, Microsoft, Meta, and Amazon announced significant increases in capital expenditure budgets last year, the prices of memory and storage chips have all risen threefold.
Former Apple CEO Tim Cook once described the current tight supply-demand relationship for memory and storage chips as "a once-in-a-century flood."
Under such circumstances, Apple will naturally include all eligible memory suppliers in its evaluation list, and it is entirely reasonable for it to approach ChangXin multiple times.
In addition to market supply and demand factors, Apple also has another clever calculation.
The Korean semiconductor media outlet "Digital Daily" believes that Apple hopes ChangXin will offer LPDDR5X quotes lower than those of Samsung and SK Hynix, both to control iPhone costs and to use the lower quotes as a bargaining chip to pressure Korean memory manufacturers.
The Korean media's speculation is not groundless. At the end of July this year, at Apple's earnings call attended by Tim Cook in his capacity as CEO for the last time, he said that the dynamic random access memory market is currently mainly controlled by three suppliers, and that having more suppliers would help improve supply and have a positive impact on product pricing.
Business is like war; making suppliers bid against each other and then reaping the lowest price oneself is itself a manifestation of the strength of leading companies like Apple.
Only this time, the chosen ChangXin did not play along with the performance.
However, ChangXin's refusal to quote does not equal a complete refusal to enter the "Apple chain," nor does it mean it has gained the upper hand in the game with Apple. In the future, given the pace of ChangXin's capacity ramp-up and technological progress, it is entirely possible that the two sides will return to the negotiating table.
Each Version Has Its Own Way of Playing
In the previous version of Earth Online, for a Chinese electronics manufacturing enterprise, entering the "Apple chain" meant changing its fate.
In 2005, Cook became Apple's COO; he abandoned building his own factories and instead built the "Apple chain" globally.
Apple was still quite generous: it would spend its own money to procure expensive specialized equipment and place it in supplier factories to help suppliers cross the threshold of heavy asset investment; of course, this was not for charity—the equipment's property rights belonged to Apple, and the production lines had to prioritize Apple's orders.
Therefore, this model is essentially still a method for Apple to forcefully control the supply chain, which the industry calls the "Apple Bank model."
In 2007, the first-generation iPhone was released, the great cycle of smartphones began, and the "Apple chain" became a land of milk and honey for electronics companies.
In 2011, Luxshare Precision entered Apple's supply chain through the acquisition of Kunshan Liantao. That year, the company's revenue was 2.556 billion yuan and net profit was 257 million yuan; ten years after joining the Apple chain, Luxshare Precision's annual revenue reached 153.946 billion yuan and net profit was 7.071 billion yuan.
It is worth mentioning that in the same year Luxshare Precision "joined the chain," Cook officially became Apple's CEO.
In addition to Luxshare Precision, Lens Technology and Goertek also achieved substantial revenue and profit growth after entering the Apple chain. Therefore, these three companies are not only regarded as the "three giants of the Apple chain," but also as a microcosm of the rapid development of China's high-value-added core component manufacturing.
However, Cook has now stepped down as Apple CEO, and the version of Earth Online has also been updated.
In 2026, smartphones are no longer the incremental market they once were. According to data from the internationally renowned consulting firm IDC, in the second quarter of this year, global smartphone shipments were 276.3 million units, down 7.4% year-on-year; the firm also predicts that this year consumers' replacement cycle may lengthen to more than 42 months, setting a record high. In 2021, this cycle was only about 24 months.
The new main storyline of the current version is AI, a story far grander than smartphones.
According to forecasts from IDC and other institutions, this year the global smartphone market size will reach $613 billion, while total AI spending will reach $2.67 trillion, up about 49.5% year-on-year, and will further increase to $3.63 trillion in 2027.
At the same time, in the previous version, although DRAM was an essential supporting component of smartphones, it could never become a core selling point; but now, DRAM has become the child of the version, a strategic resource, because no matter how powerful the GPU for large models is, it cannot run without DRAM.
The roc rises with the wind one day, soaring ninety thousand li.
ChangXin has reaped the dividends of this version. According to the company's recently released earnings forecast, in the first three quarters of 2026, the company expects operating revenue of approximately 265 billion to 270 billion yuan; net profit of approximately 131 billion to 132.5 billion yuan.
A comparison will show just how great the "potential energy" of AI is in the current version.
In the first ten years after Luxshare Precision entered the "Apple chain," the company's net profit grew about 27.5 times, which is already a rather enviable figure.
But this increase is not even in the same dimension as ChangXin's net profit growth in the past two years. In the first three quarters of this year, ChangXin's net profit reached at least 130 billion yuan, but in the same period last year, it was still losing 5.28 billion yuan.
More importantly, unlike previous "Apple chain" suppliers, ChangXin's performance growth does not rely on Apple and other foreign companies.
At present, about 90% of ChangXin's capacity is used to supply domestic customers such as Huawei, Xiaomi, and Lenovo; Xiaomi's self-developed Xuanjie 03 chip is equipped with ChangXin's globally first commercially launched LPDDR6.
In other words, for ChangXin today, whether or not it enters the "Apple chain" does not affect its performance. Therefore, ChangXin's two refusals of Apple are essentially a refusal of the old version's rules of the game.
The Next Round Full of Variables
When will the next round of the game between ChangXin and Apple arrive?
Judging from estimates by market research firm Yole, the two sides may have to wait a long time before meeting again. The firm believes that merely meeting China's domestic DRAM demand is a ten-year project.
But variables are accumulating.
In September this year, ChangXin completed the largest IPO on the STAR Market of the year, and subsequently put forward an expansion plan. Its capital ammunition is being converted into catch-up speed.
According to analysis and exclusive data from Citrini Research, if existing projects proceed as scheduled, by the end of 2026, ChangXin's monthly DRAM wafer capacity will reach about 350,000 wafers, closing in on Micron, one of the world's three major memory manufacturers.
Market reports also show that ChangXin's monthly capacity will expand from the current approximately 320,000 wafers to 420,000 wafers in 2027, while its longer-term expansion target has already set its sights on more than 500,000 wafers.
The faster capacity ramps up, the sooner ChangXin's capacity can feed domestic customers, and the closer the day will be when it returns to the negotiating table with Apple.
There is another variable, on the other side of the ocean.
In June this year, according to media reports, Apple was actively lobbying the U.S. Department of Commerce and senior White House officials over the procurement of ChangXin DRAM.
In fact, ChangXin is not on the U.S. Department of Commerce's "Entity List," so U.S. private companies are currently not restricted from purchasing ChangXin's products.
The reason Apple is doing this is to seek written confirmation or assurance from the U.S. government that ChangXin will not be added to the U.S. Department of Commerce's "Entity List" in the future.
From this it is not hard to see that Apple indeed needs ChangXin. After all, well-known Apple analyst Ming-Chi Kuo has publicly stated that the pressure Apple currently faces has escalated from simply "soaring memory prices" to "an expanding supply gap."
In fact, ChangXin will likewise need Apple in the future. Citibank pointed out in a research report that a vote of confidence from Apple could redefine ChangXin's market image from "a Chinese domestic substitute" to "a credible fourth-largest global DRAM manufacturer."
Since both sides have something the other needs, the arrival of the second round is most likely just a matter of time. It is just that, by then, the posture in which the two sides sit down at the negotiating table will not be entirely up to themselves.
ChangXin's rise cannot be separated from the broader context of AI memory being in a super cycle. On SK Hynix's recent earnings call, company executives said that "customer demand is twice the supply."
But if the current version makes a balance adjustment in this regard and heavily nerfs the strong attributes of AI memory, ChangXin's way of winning the first round may not carry over into the second round.
But at least for now, these two "small victories" have already given the overall game a good start.
This article is from the WeChat public account "Xinghai Intelligence Bureau" (ID: junwu2333), author: Xinghai Lao Ju
