Tactically Positive

Research evidence excerpt

Tactically Positive

15–18. At the MarQ earnings print, however, TER outlined a SiPh P/E 47.4 41.5 29.4 23.4

EPS ($)§ - 7.34 9.89 13.04

TAM of approx. $100 million this year, growing to $300–700 million over the Div yld (%) 0.2 0.2 0.2 0.2

medium term, below market expectations. In hindsight, CPO alone was never going Unless otherwise noted, all metrics are based on Morgan Stanley ModelWare

to take TER to $15+ of FY27 EPS. Achieving that level would also have required framework** = Based on consensus methodology

contributions from NVIDIA GPUs, memory, Trainium, and most importantly §e == MorganConsensusStanleydataResearchis providedestimatesby Refinitiv Estimates

networking. Has the operating environment changed for those other four Quarterly EPS ($)

drivers? Not really. Taiwan import data suggest TER’s JunQ networking revenue is 2026e 2026e 2027e 2027e

Quarter 2025 Prior Current Prior Current

likely to be extremely strong once again, the memory environment has continued to Q1 0.75 - 2.56a - 2.15

improve, and while we remain dismissive of meaningful NVIDIA merchant-GPU share Q2 0.57 - 2.05 - 2.82

Q3 0.85 - 1.69 - 3.12

gains, TER is well positioned to benefit from Groq LPU and Vera CPU. Q4 1.80 - 1.71 - 3.29

e = Morgan Stanley Research estimates, a = Actual Company reported data

Tactically positive into the quarter. This leaves us tactically positive into the

quarter for 3 reasons: 1) JunQ upside: Similar to MarQ, we believe TER’s reported

JunQ results will track ahead of guidance, supported by pull-ins primarily related to

networking. 2) TPU qualification: We expect TER to comment on its TPU

qualification efforts and the potential to qualify during this CY. We are not making

an explicit call on qualification, but we do believe management will speak to the

opportunity.