They have 6 of the top 10 optical integrators locked in for AI interconnects, yet the street prices them like a cyclical commodity play waiting for auto inventory to clear.
Thesis: TSEM is an arbitrage on AI interconnects mislabeled as an analog foundry. While the street obsesses over the cyclical bottom in Power Management (Fab 9), the real story is the structural mix shift to 300mm RF SOI and Silicon Photonics (SiPho). SiPho is already 5% of revenue and margin accretive, critical for 800G/1.6T AI clusters. With Fab 7 effectively sold out, the Agrate ramp isn't just capacity—it's a margin expansion engine. You're buying the AI picks-and-shovels at a value multiple.
Verdict: LONG — Conviction: MEDIUM
Catalyst: Q2/Q3 revenue ramp confirming 'Q1 was the low' in Power, combined with SiPho breaking out as a named reporting segment or crossing 10% revenue share.
Key Risk: Execution risk on the Agrate qualification. If the 300mm transition delays, they remain capped at Fab 7 and bleeding cash at Fab 9.
The Tell: CFO Shirazi admits Fab 9 utilization is 60% due to 'worldwide decreased demand' and that automotive inventory is 'still burning off,' yet relies on 'forecasts' rather than firm orders to call Q1 the bottom. The recovery in Power is hope; the growth in RF is math.
Friction Level: MODERATE_FRICTION — The $2.66B revenue target. Bears see a pipe dream requiring 2x growth from current run rates; Bulls see the 300mm capacity unlock at Agrate making the math inevitable.
Report not found
The report data is no longer available. Please return to the archive.