SIVERS SEMICONDUCTORS - CROSS-VERTICAL OPTION VALUE ON A DUAL RF-PHOTONICS STACK

Apr 29, 2026

∙ Paid

SEQH CAPITAL RESEARCH - TEAR SHEET
SIVERS SEMICONDUCTORS - CROSS-VERTICAL OPTION VALUE ON A DUAL RF-PHOTONICS STACK

WHAT THIS REPORT ANSWERS

The report extends SEQH’s April 23 Sivers deep dive by treating Sivers Semiconductors (SIVE / SIVEF) as a platform of real options on three end markets, rather than three separate revenue lines, and quantifies the cross-vertical convexity this creates.
It shows that once imperfect correlation, shared fab learning and capacity rebalancing are modeled explicitly, the platform is worth about SEK 4–7 per share more than a simple sum-of-the-parts view, lifting the 12 month target from SEK 38.50 to a convexity adjusted range of SEK 41–45 while reaffirming OVERWEIGHT.

Platform and thesis upgrade

Sivers is reframed as a dual stack company: a wholly owned Glasgow InP cleanroom plus a GlobalFoundries 22FDX / 45RFSOI RF stack feeding three main vectors, CPO external light source, Ka band SATCOM and FMCW LiDAR, with 5G / 6G RF as a fourth.
These verticals share process, IP and engineering, but have different ramp speeds and different obsolescence horizons, so the chance that all three fail at once is far lower than the per vector risk would suggest.
SEQH’s central upgrade is that the joint option value of this structure is strictly greaterthan the sum of single vertical NPVs or option values, and that this gap is large enough to matter for price targets.

Each major vector is modeled as a call option on steady state revenue: CPO ELS (Ayar, POET, O Net / Enablence), LiDAR (Aeva on NVIDIA DRIVE Hyperion), SATCOM(ALL.SPACE Ka band), Tier 1 5G / 6G, Daybreak FR3 and defense / IRIS².
Sum of single vector option values is about SEK 2,680M, with LiDAR the largest single piece, SATCOM long dated but high quality, and the CPO triad meaningful but more volatile and shorter lived.
LiDAR has the shortest time to expiry and lowest volatility after the Hyperion design in, while SATCOM has the longest life and strongest per terminal dollar content, and CPO carries the most sigma but is divided across three architectures.

A pairwise correlation matrix shows the three CPO paths are highly correlated with each other, but only weakly correlated with LiDAR and SATCOM; average off diagonal correlation is about 0.18, which keeps the blended platform volatility near 58 percentinstead of 70 plus percent.
Using a multivariate Black Scholes Monte Carlo with compound options for the CPO triad, plus explicit terms for shared Glasgow yield learning, GF 22FDX node sharing and Glasgow / WIN capacity switching, SEQH derives a platform real option value of about SEK 3,240M, or SEK 10.66 per share.
Adding pro forma net cash, wireless in place earnings and legacy photonics yields total enterprise value around SEK 4,155M before additional unmodeled upside, corresponding to about SEK 13.67 per share, and lifting the convexity adjusted 12 month EV range to SEK 4.96–5.56B or SEK 16.30–18.27 per share at the platform level.
Reconciling this back to the original deep dive’s SEK 14.50 / 35 / 59 bear base bull range and SEK 38.50 probability weighted target gives SEK 41–45 once the 10–18 percentconvexity uplift is applied, with the SEK 28.36 April 23 spot used as the valuation anchor.

Four concrete mechanics drive the convexity: etched facet plus on wafer test at Glasgow that let LiDAR yield work benefit CPO wafers, shared GF 22FDX node between Daybreak FR3 and the Tier 1 5G Gen 3 beamformer, Glasgow plus WIN capacity rebalancing between vectors, and compound de risking across Ayar, POET and O Net.
SEQH argues this makes Sivers’ risk adjusted profile superior to single architecture CPO peers like POET, AAOI and Credo, because per vector volatility is high but the portfolio sigma is lower and the information ratio higher.
The note also tightens the bear case: with joint failure of CPO, LiDAR and SATCOM modeled at only about 4–5 percent, the realistic downside floor is closer to SEK 18–20 than the prior SEK 14.50 once existing contracts and the Bootstrap Europe convertible are considered.
Key convexity specific risks are correlation drift if the CPO architectures share a hidden common mode failure, delayed WIN Semi qualification that reinstates single fab dependence, LiDAR single program exposure to Aeva on Hyperion, and the possibility that linear drive optics and OCS displace CPO before ELS reaches scale.

SEQH maps each upcoming event to its option Greeks: Q4 2026 Aeva LiDAR ramp is the biggest delta event, Ayar Labs H2 2027 laser volume qualification is the biggest vegaevent, and a Nasdaq New York listing submission around Q3 2026 mainly affects the discount rate and institutional access.
Re rating triggers include the Nasdaq submission, on time Aeva ramp, and the ordering of CPO commercialization across POET, O Net and Ayar, where divergent timing is positive for convexity and convergent stalling is negative.
SEQH reaffirms OVERWEIGHT, positions Sivers as its flagship RF photonics sleeve name, and sets a convexity adjusted 12 month target of SEK 41–45, noting that the main thing to watch for a thesis break is any sign that cross vertical correlations are rising due to shared yield or thermal failures.

FULL 15-PAGE REPORT ATTACHED BELOW IN PDF FORMAT: