Semiconductor equipment demand is closely tied to manufacturers’ decisions about capacity, process migration and technology development. In its 2019 outlook for the second half of that year, Mirae Asset Daewoo argued that equipment suppliers could be better positioned than memory makers amid weakening prices and uneven near-term investment. The reasoning rested on two potential sources of future demand: renewed NAND and DRAM capacity spending from 2020, and more complex processes in foundry manufacturing.
This was a forward-looking assessment made in 2019, not a statement about current market conditions. Its central lesson for automation and manufacturing decision-makers is that equipment demand depends not only on near-term chip sales, but also on how producers manage inventory, convert production lines and prepare for new process nodes.
Why semiconductor equipment stood out in the 2019 outlook
The report described a memory market facing continued price declines and uncertainty about the timing of recovery. NAND demand was supported by growing storage content in devices and demand for enterprise solid-state storage, while DRAM demand reflected smartphones and servers. But these demand drivers did not automatically translate into immediate investment: producers were also adjusting output and working through inventory.
Against that backdrop, the outlook favored equipment suppliers because manufacturers were expected to resume or increase investment after reducing spending in 2019. The anticipated recovery was not uniform across producers or memory types. It depended on company-specific plans, facility expansions and migration to newer technologies. For equipment suppliers, the relevant signal was therefore the direction and composition of capital expenditure, not simply the headline market outlook.
NAND capex: restrained spending before planned additions
The 2019 NAND outlook combined modest expected capital-spending cuts with plans for later capacity growth. The report anticipated that industry-wide NAND capex would contract year over year, while wafer capacity would remain broadly flat. Some producers were reducing legacy capacity or converting production toward 3D NAND, illustrating how investment can shift from adding wafer starts to changing what existing facilities can manufacture.
Company plans differed. Samsung Electronics was expected to reduce NAND capex in 2019, with partial migration to 128-layer production at its P1 facility and limited additions at Xi’an II. The report then projected higher 2020 spending, based on further Xi’an II capacity additions. SK Hynix was also expected to cut 2019 NAND capex, while migrating production and planning additions at M15; its 2020 spending outlook assumed another capacity expansion there.
These projections point to several distinct equipment requirements: tools for process conversion, equipment supporting 3D NAND production, and tools for new capacity. The timing matters. A period of lower aggregate capex can coexist with ongoing investment in specific process transitions, while larger expansion programs may create a later demand step-up.
DRAM investment was expected to recover later
The report similarly expected DRAM spending to weaken in 2019 before strengthening in 2020. Samsung’s 2019 plan included process migration, with no wafer-capacity capex slated for the year in the report’s assessment. Its 2020 outlook assumed additions at the Pyeongtaek P1 and P2 facilities. SK Hynix was expected to reduce 2019 DRAM capex while undertaking technology migration, with later spending growth linked to capacity expansion at its C2 facility.
For semiconductor manufacturing, this distinction between process migration and capacity expansion is important. Both can require equipment, but they reflect different production needs. Migration upgrades a line’s capability; capacity additions increase the amount of production the company can support. The 2019 outlook suggested that a pause in wafer capacity growth did not mean investment activity had stopped altogether.
Demand recovery also appeared gradual in the report’s DRAM analysis. Although a pickup in demand was expected to help reduce inventory, it cautioned that reaching optimal inventory levels could take time. That uncertainty could affect the pace of pricing and profitability improvement—and, in turn, the confidence and timing of manufacturers’ investment decisions.
Advanced foundry processes create another equipment pathway
Memory was not the only basis for the equipment preference. The report highlighted competition in foundry manufacturing, where Samsung sought to gain ground against TSMC through advanced process development. It described Samsung’s move to EUV at 7 nm and its plan to introduce a gate-all-around, or GAA, transistor structure at the 3 nm node.
The shift from FinFET to GAA was presented as a potential driver of greater use of atomic layer deposition (ALD). The report connected this need to limitations in existing deposition methods and the requirements of the newer device structure. For equipment suppliers, this illustrates how a change in transistor architecture can create demand for process-specific tools, even when the overall chip market is subdued.
What industrial decision-makers can take from the analysis
The outlook’s investment case depended on a sequence of events rather than a single market rebound: inventory adjustment, continued process conversion, planned facility additions and the development of advanced foundry technologies. For teams tracking semiconductor manufacturing, useful indicators include:
- Capital expenditure by memory type: Separate NAND and DRAM plans, and distinguish wafer additions from migration-related investment.
- Production-line conversion: Track changes such as the transition to 3D NAND or newer process nodes, which can create equipment needs without immediate expansion in wafer capacity.
- Facility-level schedules: Company-specific projects and their timing shape when planned investment may translate into equipment demand.
- Process architecture: Developments such as GAA can influence demand for specialized deposition equipment, including ALD tools.
Conclusion
The 2019 assessment favored semiconductor equipment suppliers because it saw a potential investment recovery after near-term spending restraint, alongside ongoing process migration and advanced foundry development. Its practical implication is that equipment demand should be evaluated through the details of manufacturing plans—not inferred from chip demand or memory prices alone. Capacity additions, technology transitions, inventory conditions and process roadmaps all influence when and where manufacturers may invest.