They are tripling the only revenue stream that matters while the street worries about handset sensors.
Thesis: The market is pricing TSEM as a commoditized analog foundry exposed to weak mobile/imaging cycles. The edge is the mix shift. Infrastructure (SiPho/SiGe) is not just growing; it's exploding (3x YoY) and carries margins 'substantially higher' than the corporate average. The Fab 3 utilization of 65% is not a demand failure but a bottleneck opportunity—the $350M CapEx unlocks the remaining 20% capacity at incremental margins. You are buying a high-speed optical monopoly wrapped in a boring analog wrapper.
Verdict: LONG — Conviction: HIGH
Catalyst: Q4 earnings confirming the $150M annualized SiPho run rate and the first material revenue recognition from the Agrate facility.
Key Risk: Hyperscaler architecture shifts (e.g., NVIDIA moving away from current optical standards) could leave TSEM with specialized capacity and no sockets.
The Tell: CEO Ellwanger admitted, 'we have some wafers in line that indeed have been put on hold... for a mask change' regarding NVIDIA's platform. He dismissed the financial impact, but it confirms they are a downstream derivative of hyperscaler design churn, not a strategic partner.
Friction Level: HIGH_FRICTION — Valuation model mismatch. The Street models a cyclical legacy foundry; the data shows a structural shift to high-margin AI infrastructure (Silicon Photonics).
Report not found
The report data is no longer available. Please return to the archive.