Taiwan's DRAM maker Nanya Technology is sharply ramping up capital expenditure. President Lee Pei-ing, speaking at the earnings call on July 10, 2026, presented a preliminary plan to invest over NT$200 billion in 2027—roughly four times the NT$52 billion capex ceiling set for 2026. However, according to Reuters, this figure has not yet been approved by the board of directors. Will the company redirect profits from soaring prices into production equipment for a new fab whose construction has dragged on for years? Nanya Technology is now moving from being a beneficiary of the DRAM shortage to becoming a contributor to supply.
What the Over NT$200 Billion Will Build at the New Fab
The scale of the planned 2027 investment cannot be explained as a simple extension of past years for Nanya Technology. Capital expenditure was NT$13.2 billion in 2023, NT$16.1 billion in 2024, and NT$13.4 billion in 2025. For 2026, the ceiling was raised to NT$52 billion, but actual spending in the first half was only NT$6.9 billion. The preliminary plan for 2027 exceeds twice the NT$94.7 billion total spent over the four years from 2023 through 2026 combined.
The funds are destined for the 12-inch DRAM fab under construction at the Nanlin Technology Park in Taishan District, New Taipei City. As a first phase, Nanya Technology aims to establish a monthly wafer input capacity of 30,000 sheets by 2028. The company ultimately plans to expand this to 45,000 wafers per month, with total investment—including the building itself—expected to reach NT$480 billion. The over NT$200 billion planned for 2027 represents a single year's execution of this overall total.
There has also been a shift in timeline. When construction broke ground in June 2022, Nanya Technology set the total investment at NT$300 billion and planned to complete the fab by 2025. Now, the schedule calls for equipment move-in in the first quarter of 2027, with the first phase reaching 30,000 wafers per month by 2028. The total investment estimate has grown by NT$180 billion, and the capacity ramp-up has been pushed back. The company has not disclosed the reasons for the increase.
The new fab will use Nanya's own 10nm-class process technology. Currently under development are the 1C, 1D, and 1E generations, along with support for EUV lithography. The massive investment figure does not simply represent the completion of a building—it includes the costs of installing advanced manufacturing equipment, qualifying processes, and improving yields. In the 2026 capex plan, wafer fabrication equipment accounts for roughly 30% of the total.
Price, Not Volume, Drove the 79.5% Gross Margin
What made the expanded investment possible was a sharp reversal in DRAM market conditions. Nanya Technology's revenue for the second quarter of 2026 came to NT$82.549 billion, up 68.2% from the previous quarter. Gross profit reached NT$65.619 billion, with a gross margin of 79.5%. In the same quarter a year earlier, the company posted a gross loss of NT$2.165 billion, for a gross margin of negative 20.6%. In just one year, profitability swung completely in the opposite direction.
The main driver of higher revenue was not shipment volume. In the second quarter, the average selling price rose more than 60% quarter-on-quarter, while bit shipments remained flat. In the first quarter as well, the average selling price rose more than 70%, even as shipments fell by a mid-single-digit percentage. In other words, the surge in profit did not come from selling significantly more product, but from buyers paying higher prices for limited supply.
Net profit for the second quarter reached NT$50.192 billion, bringing the total for the first half of 2026 to NT$76.25 billion. That is about 11.5 times the full-year 2025 net profit of NT$6.603 billion. Market swings of this magnitude are significant. Even so, there is a rationale behind the capex plan: converting cash earned during a period of high prices into future production capacity.
Nanya Technology plans to grow full-year 2026 bit shipments by a percentage in the high teens year-on-year. Adjusting product mix and production efficiency at existing fabs can boost near-term shipments. But expanding wafer input capacity itself on a large scale requires the new fab. Between today's high margins and the supply increase expected in 2028, there is a gap of at least a year and a half.
HBM Demand Is Pushing Up Even Commodity DRAM Prices
Nanya Technology's rapid earnings recovery also reflects how AI-driven memory demand has spilled over into the broader DRAM market. The company explains that AI servers and general-purpose servers require massive amounts of HBM and RDIMM, which is squeezing memory supply not only for PCs and smartphones but also for automotive and consumer electronics applications. When major manufacturers shift production capacity toward higher-margin products, supply of mature products such as DDR4 also tightens. This is why even devices that have nothing to do with AI data centers are being affected by rising memory prices.
Nanya Technology supplies DDR5 and LPDDR5/5X, as well as DDR4, LPDDR4/4X, DDR3, and LPDDR3. In the first half of 2026, AI infrastructure and server-related products accounted for more than 20% of revenue. DRAM content is increasing not only in the core memory of AI computing systems but also in peripheral equipment such as enterprise SSDs, networking equipment, and management controllers. With its broad product lineup, the company has found opportunities to redirect products toward markets where supply has opened up.
That said, the supply shortage is far too large to be resolved by Nanya Technology's expansion alone. The company expects the tightness to persist for several more quarters, and plans to expand its new capacity from 2028 onward in line with multi-year long-term supply contracts and demand forecasts. This is also a move to avoid the DRAM industry's cyclical pattern in which overproduction causes prices to collapse—locking in customer contracts first, and moving forward with equipment investment only within a range where future buyers are visible.
NT$198.4 Billion in Cash, but Approval and Mass Production Are Separate Matters
Nanya Technology has accumulated the cash needed to begin the investment. Net cash at the end of June 2026 stood at NT$198.4 billion, close to the preliminary plan for 2027 capex. In the second quarter alone, the company generated NT$55.013 billion from operating activities, and after subtracting NT$4.046 billion in capital expenditure, free cash flow reached NT$50.967 billion.
In addition, the company raised NT$78.72 billion in April through a private placement subscribed by four major customers. This placement represents 10.19% of shares outstanding after issuance. The company has stated that the proceeds will be used for factory facilities and advanced memory manufacturing equipment. Having secured both funding and long-term customer relationships simultaneously makes it easier to proceed with investment in the new fab.
Even so, cash reserves and mass-production capacity are not directly linked. First, the 2027 budget must be approved by the board of directors. Following that will come equipment move-in, process qualification for the 1C generation and beyond, customer product evaluation, and yield improvement. The 30,000-wafer-per-month capacity target is itself a planned figure to be reached by 2028.
Whether the 2027 investment amount is formally budgeted, whether equipment move-in begins as scheduled in the first quarter of 2027, and whether the fab reaches 30,000 wafers per month by 2028—these three factors will determine whether the current price surge can be converted into an actual increase in supply. If Nanya Technology gets there, the supply channel for commodity DRAM, narrowed by AI demand, will finally begin to widen.