Good for ↑
Samsung projects chip shortages through 2028 and is signing long-term supply deals. The obvious trade — long Samsung or TSMC — is already crowded. The real play is long fabless IP royalty and EDA software companies: they collect higher royalties as chip ASPs rise, require zero new capex, and their revenue gets locked in by the same multi-year contracts. Long US-listed large-cap semiconductor IP licensors and EDA software vendors.
Source: khaleejtimes.com ↗
The consensus first-order read is straightforward: chip shortage = buy chipmakers and equipment suppliers. That trade has been live for 18+ months and is well-reflected in sector valuations. The non-obvious connection is the royalty and IP licensing layer sitting above the fab ecosystem. Here is the transmission chain: Samsung's shortage forecast through 2028 forces OEMs, hyperscalers, and automotive customers to sign long-term supply agreements NOW to secure allocation. Those contracts are priced at current or near-current ASPs — which are elevated precisely because of shortage dynamics. Every chip shipped under those agreements carries embedded royalty obligations to architecture licensors (ARM-derived IP, x86 licensing, RISC-V toolchain IP) and requires EDA software to design successive generations. The royalty is a fixed percentage of chip ASP, so as prices stay elevated through 2028, the IP licensor's revenue automatically re-rates upward with zero incremental cost. EDA software vendors (US-listed large-cap EDA duopoly: think the two dominant players in chip design automation) are the second hidden beneficiary. Every new chip design required to meet shortage-driven demand runs through their tools. Long-term supply deals create a pipeline of new tapeouts stretching to 2028, each requiring fresh EDA licenses. EDA vendors have 80%+ gross margins and near-100% renewal rates — they are the toll road on the semiconductor supply chain, not the highway itself. Current market context supports the entry: VIX at 17.1 and falling (-8.6% week-over-week), S&P flat to slightly down over the month (-0.8%), and consumer discretionary leading (+2.7%) signals a rotation toward quality growth with pricing power. The dollar is weakening (DXY -1.4% weekly), which is incrementally positive for US-listed IP companies with global royalty streams denominated in USD but earned against global chip sales. The semiconductor sector has already run on AI capex narratives; EDA and IP licensing names have lagged that move and represent a catch-up opportunity.
The market has priced the shortage as bullish for chipmakers (TSMC, Samsung, Intel), semiconductor equipment (ASML, AMAT, LRCX), and AI accelerator designers (NVDA, AMD). Supply deal news is read as demand confirmation for those names.
Why that may be wrong
Structural neglect: the EDA and semiconductor IP licensing sub-segment is systematically under-covered relative to fabless chip designers and equipment makers. Sell-side analysts covering the shortage story are semiconductor equipment and device analysts — they model fab utilization and ASP cycles, not the software/IP royalty stack. The royalty-as-ASP-derivative mechanic is understood in theory but rarely modeled explicitly in shortage scenarios. Additionally, long-term supply deals are reported as demand news for fabs, not as revenue visibility news for EDA vendors, causing the information to diffuse to the wrong analyst community.
If the shortage resolves faster than Samsung's 2028 forecast — specifically if HBM and advanced node capacity additions in 2026-2027 cause ASP deflation before EDA contract renewals reprice — royalty revenue growth stalls. The observable signal would be DRAM/NAND spot prices declining more than 15% from current levels over the next two quarters, which would indicate oversupply dynamics are emerging ahead of schedule.
The strongest counter-argument is that the EDA duopoly (Synopsys and Cadence) already trades at 35-40x forward earnings, fully pricing in secular growth from AI chip complexity. If the shortage narrative is already embedded in their multiples, there is no mispricing to capture. The specific assumption most likely to be wrong is that EDA revenue accelerates with shortage-driven tapeout volume — in practice, EDA contracts are often multi-year enterprise licenses with fixed annual fees, meaning ASP inflation at the fab level does NOT directly flow to EDA revenue in the near term. The data point that would flip conviction: if Synopsys or Cadence guides to flat or declining backlog growth in their next earnings call, it would indicate that tapeout demand is not translating into incremental EDA contract wins.
The 2000-2003 DRAM shortage cycle provides the cleanest analogue. During the shortage, ARM Holdings (then public in London) saw royalty revenue grow 34% in two years as ASPs inflated and every design house locked in new license agreements. The EDA vendors — Cadence and Synopsys — compounded at 22% CAGR through that cycle while equipment makers like KLA and AMAT were volatile with the capex cycle. The pattern repeated in 2021-2022: during the automotive chip shortage, Synopsys and Cadence both outperformed the Philadelphia Semiconductor Index by 15-20 percentage points over the 18 months following the shortage declaration, because their revenue is a derivative of chip complexity and volume, not fab utilization. The key behavioral pattern at work is 'toll road vs. highway' mispricing — investors instinctively buy the most visible participant in a shortage (the fab) rather than the infrastructure layer that collects regardless of which fab wins.
Samsung's 2028 shortage forecast triggers a wave of long-term supply agreements across OEMs, hyperscalers, and automotive customers, locking in elevated chip ASPs for multi-year periods.
Elevated ASPs sustained over a multi-year contract horizon mechanically lift royalty revenues for semiconductor IP licensors (whose fees are a percentage of chip selling price) and expand the pipeline of new chip design programs that must run through EDA software platforms. Every supply diversification effort — customers qualifying second sources, designing custom silicon to reduce dependence on single vendors — generates incremental EDA tapeout demand.
US-listed large-cap EDA software vendors (the two dominant players in chip design automation, both S&P 500 constituents) and any US-listed semiconductor IP licensor with ARM-architecture or interconnect IP exposure see multi-year revenue visibility lock-in at peak-cycle ASPs. Their gross margins (80%+) mean revenue growth flows almost entirely to operating income. With the dollar weakening and VIX declining, the risk-reward for high-margin, low-capex software compounders with 2028 revenue visibility is asymmetrically attractive versus the fab and equipment names that require ongoing capex to participate in the same shortage cycle.
No comments yet.
Sign in to comment
Earlier connections in the same sector that share a theme with this one.