Will Tower Semiconductor's SiPho quarterly wafer revenue remain above $90M in Q2 2026?

SiPho quarterly revenue has been on an exponential trajectory: ~$40M Q1, ~$55M Q2, ~$70M Q3, $95M Q4. Even stripping NRE from Q4 (estimated $10-15M), production wafer revenue was $80-85M. The quarterly trajectory shows consistent 25-35% sequential growth. For Q2 2026, the question is whether the trend sustains to $90M+ on production wafers alone. With 5x capacity expansion underway, Fab 2 and Fab 9 ramping (60-65% utilization), and 70%+ capacity reserved with prepayments, the demand-supply dynamics strongly favor continued growth. The $90M threshold is actually below the implied Q2 trajectory if Q1 continues from Q4 base.

The critical nuance is the NRE component. Q4 SiPho revenue was $95M but CEO acknowledged some non-wafer NRE. If NRE was $10-15M, production wafer revenue was $80-85M -- which is already close to the $90M threshold. For Q2 2026 to clear $90M in production wafers, there needs to be genuine capacity ramp. Management 5x target by end 2026 implies tools are being qualified in H1, with meaningful production wafer output increase. However, tool qualification typically takes 3-6 months, and some tools ordered in Q4 may not be producing revenue by Q2. The base case is that Q1 qualifications drive incremental Q2 output, pushing above $90M, but there is a nonzero chance qualification delays keep Q2 production revenue at $80-90M range.

The structural demand argument is compelling: SiPho vs EML advantage is permanent at 1.6T, data center bandwidth requirements compound, and Tower is the dominant manufacturer. Q2 2026 NRE may actually be higher than Q4 if Tower is doing Gen+1 and Gen+2 development work for 3.2T, which would boost total SiPho revenue even if production wafer volume is flat. Taking the question at face value using management reported SiPho revenue figure, $90M+ in Q2 2026 is highly likely given the $95M Q4 base and continued capacity ramp.

The quarterly trajectory strongly supports exceeding $90M. However, I weight two risks: (1) Q1 2026 guidance shows a seasonal step-down to $412M total revenue from Q4 $440M. If SiPho experiences a proportional step-down, Q1 SiPho could be $85-90M, and Q2 may only recover to $90-95M. (2) AI infrastructure spending normalization is a genuine risk -- if any major hyperscaler reduces or delays orders in Q1-Q2 2026, the effect cascades to module makers and then to Tower within 1-2 quarters. The 70% reservation probability is high, but not overwhelming given these specific timing risks.

The question specifically asks about wafer revenue which I interpret as including NRE since management reports them together. On this basis, Q4 $95M base is strong. But Q2 is only two quarters out from the data vintage (Q4 2025), and the capacity ramp is back-loaded to H2 2026. Q1-Q2 2026 SiPho revenue may be closer to Q4 levels rather than materially above. The risk scenario: SiPho production revenue was really $80-85M in Q4, stays flat in Q1-Q2 while tools are being qualified, and NRE drops as Gen+1 development completes a phase. In that scenario, total SiPho revenue could dip to $85-90M range.

Weight of evidence favors YES. The demand trajectory is strong, reservations provide visibility, and the $90M threshold is below the current run rate. Even in a downside scenario where production wafer revenue is flat at $80-85M, NRE or initial 3.2T development revenue would likely push total SiPho revenue above $90M. The main risk is a genuine demand shock (major hyperscaler capex cut) which is unlikely but not impossible in this 6-month window.

Q4 2025 SiPho revenue was $95M. The exponential growth trajectory and 5x capacity expansion plan make $90M+ in Q2 2026 very likely. Customer prepayments and reservations provide demand certainty. The threshold is below the current run rate.

NRE component makes Q4 baseline less certain. If true production wafer revenue was $80-85M, the $90M threshold becomes more of a stretch for Q2 specifically, before major capacity additions qualify. Still more likely than not to exceed $90M given the growth trajectory.

Strong demand signals and capacity expansion support exceeding $90M. The main downside is timing -- H1 2026 is tool qualification, H2 is full ramp. Q2 may be in the transition window. But management track record of sequential growth and customer lock-in tilts probability toward YES.