Si-Cap TAM expands, legacy DRAM capacity may see new opportunities
AI summary card
Si-Cap TAM expands, legacy DRAM capacity may see new opportunities
Morgan Stanley believes SEMco's large silicon capacitor contract validates Si-Cap demand expansion, and applications such as next-generation TPUs, AI servers, and high-performance computing could drive more long-term supply opportunities.
- SEMco announced a silicon capacitor supply contract worth about KRW 1.5 trillion, or about US$1 billion, with a term from 2027-01-01 to 2028-12-31.
- The report believes this contract will likely mainly supply EMIB-T needed for next-generation TPUs, but the customer base may later expand to more applications such as mobile devices.
- Because Si-Cap supply is oligopolistic, customers are more likely to sign long-term contracts with suppliers, improving revenue visibility across the value chain.
- Legacy DRAM processes are suitable for Si-Cap manufacturing; the report believes Taiwan DRAM suppliers have an opportunity, and with PSMC at full capacity, Winbond may become a potential SEMCO partner.
Report interpretation
Overview
This report focuses on Si-Cap opportunities in the Greater China and Asia Pacific semiconductor industry. The key catalyst is SEMco's announcement of a two-year silicon capacitor supply contract worth about KRW 1.5 trillion, or roughly US$1 billion, with a global large company. Morgan Stanley believes the contract may primarily address demand for EMIB-T used in next-generation TPUs and shows that the addressable market for Si-Cap is expanding.
Core views
The report's core view is: first, silicon capacitor demand is no longer limited to a single AI server use case and may expand into high-performance computing, autonomous driving systems, and mobile devices; second, the number of Si-Cap suppliers is limited, so customers are more likely to lock in supply through long-term contracts; third, legacy DRAM processes are very suitable for Si-Cap manufacturing, so Taiwan DRAM suppliers may benefit; fourth, with PSMC at full capacity, Winbond may become a potential partner for SEMCO, though its 2027 revenue contribution is expected to remain in the low single digits.
Analysis framework
The report combines event-driven analysis with value-chain mapping: it uses SEMco's newly signed supply contract as a demand-validation signal, and then infers the potential impact of Si-Cap TAM expansion on regional semiconductor companies by considering downstream application expansion, the supply structure, the likelihood of long-term contracts, and the suitability of legacy DRAM processes.
Methodology notes
Judge Si-Cap market expansion based on new contracts and application growth
The report treats SEMco's large long-term contract as evidence of expanding silicon capacitor demand and further focuses on potential new demand sources such as AI servers, high-performance computing, autonomous driving, and mobile devices.
Higher supply concentration increases the likelihood that customers will sign long-term contracts
Because Si-Cap supply is relatively oligopolistic, the report believes customers are more likely to secure supply in advance, improving supplier order visibility and revenue visibility.
Legacy DRAM processes are suitable for Si-Cap manufacturing
The report believes older DRAM processes can be used to produce silicon capacitors, so some Taiwan DRAM suppliers may gain new business opportunities due to capacity and process compatibility.
Asset mapping & comparison
Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).
- Winbond Electronics CorpPotential beneficiary in the Si-Cap supply chain
- Strengths
- Its legacy DRAM process is compatible with Si-Cap manufacturing, and the report believes it may become a potential SEMCO partner when PSMC is at full capacity.
- Weaknesses
- The report expects related 2027e revenue to account for only a low-single-digit share, so near-term financial upside may be limited.
- Comparison
- Compared with PSMC, which is already at full capacity, Winbond is identified by the report as the more likely potential partner to take on the collaboration.
- Risks
- There remains uncertainty around deal execution, customer qualification, pricing, capacity allocation, and the actual ramp-up pace of mass production.
- Powerchip Semiconductor Manufacturing CoReference for related legacy DRAM/foundry capacity
- Strengths
- It has a capacity base related to legacy DRAM and other mature processes.
- Weaknesses
- The report says PSMC is at full capacity, limiting its ability to take on additional Si-Cap collaboration.
- Comparison
- Under capacity constraints, Winbond is relatively more likely to become a potential SEMCO partner.
- Risks
- Tight capacity may prevent it from fully participating in new Si-Cap demand.
- AP Memory Technology CorpSource of industry demand signals
- Strengths
- The report cites remarks from its conference call about the possible appearance of more customers, supporting the view that Si-Cap customer adoption is expanding.
- Weaknesses
- The report does not provide its specific order amount or direct financial impact.
- Comparison
- Compared with the SEMco contract, AP Memory provides more qualitative signs of demand diffusion.
- Risks
- The pace of customer expansion and the scale of commercialization still need to be verified later.
- SEMCOThe contract signer and the core demand-validation event for silicon capacitors
- Strengths
- It signed a large two-year Si-Cap supply contract worth about KRW 1.5 trillion and plans to expand beyond AI servers into high-performance computing, autonomous driving, and mobile devices.
- Weaknesses
- The report does not disclose the customer name, and details on the exact product, gross margin, and capacity arrangement are limited.
- Comparison
- The SEMCO contract is the main event evidence the report uses to judge Si-Cap TAM expansion.
- Risks
- Contract execution, customer concentration, technology shifts, and supply-chain delivery could all affect the final outcome.
Key data
- SEMco silicon capacitor contract amountAbout KRW 1.5 trillion, about US$1bnThe report says the contract was signed by SEMco and a global large company.
- Contract term2027-01-01至2028-12-31The contract covers a two-year supply period.
- Potential primary applicationEMIB-T for next-generation TPUsMorgan Stanley believes the contract will likely mainly supply this type of demand.
- Potential expansion applicationsAI servers, high-performance computing, autonomous driving systems, mobile devicesSEMco mentioned plans to diversify supply destinations, and the report also cited comments from AP Memory's conference call about more customers.
- Winbond potential revenue contributionLow-single-digit share of 2027e total revenueThe report believes that even if it becomes a potential SEMco partner, the contribution may still remain in the low single digits.
- Morgan Stanley global rating coverage distributionOverweight/Buy 1546 firms at 42%; Equal-weight/Hold 1568 firms at 43%; Underweight/Sell 555 firms at 15%This comes from the disclosure section's rating distribution table and does not represent a new rating on any single company in this report.
Impact & implications
If Si-Cap demand expands as the report expects, legacy DRAM capacity may find a new structural application outlet beyond the mature memory cycle. For suppliers, long-term contracts help improve order visibility; for investors, it is important to watch legacy DRAM capacity, Si-Cap manufacturing capability, customer qualification progress, and the pace of actual revenue contribution, rather than inferring broad earnings upside from a single contract.
Risks
- The report does not disclose the SEMco contract customer name, creating incomplete visibility into the demand source and final application.
- Si-Cap expansion from AI servers into mobile, autonomous driving, and other scenarios still needs customer adoption and mass-production validation.
- Whether legacy DRAM suppliers can convert capacity into Si-Cap revenue depends on customer qualification, capacity allocation, yield, and pricing.
- Winbond's potential contribution is expected to be only a low-single-digit share of 2027e revenue, so the near-term earnings impact may be limited.
- Morgan Stanley discloses that it has investment banking, shareholding, or other service relationships with multiple covered companies, and investors should pay attention to potential conflicts of interest.
What to watch
- Further disclosure of SEMco's customers, application scenarios, and shipment pace.
- Order changes in the supply chain related to next-generation TPUs and EMIB-T.
- Updates from AP Memory, Winbond, PSMC, and others on Si-Cap demand, capacity, and customer qualification.
- Whether long-term Si-Cap contracts are replicated among more customers and suppliers.
- Changes in legacy DRAM capacity utilization and the feasibility of migrating it to silicon capacitor manufacturing.