CEO says grow faster in '26 than '25. CFO was asked the same question twice and retreated both times to 'outgrow the industry.'
Thesis: Infrastructure is genuinely scaling and will become the largest segment in 2026, with 28-week optical lead times and margin expansion from mix shift into PAM4 DSPs and storage accelerators. But $58M broadband (43% of revenue) is guided down for the year, the Q1 midpoint implies a sequential decline, and the CFO won't repeat the CEO's growth claim. The operating leverage is real, the narrative ahead of it is not fully backed by forward numbers.
Verdict: HOLD — Conviction: MEDIUM
Catalyst: Q1 2026 print: revenue landing at $130M vs $136.4M in Q4, plus OFC in March with Rushmore announcements and electrical retimers for scale-up Ethernet. Keystone ASP clarity emerging through H1.
Key Risk: Single-customer loss in Keystone hyperscaler ramps, PON Tier 1 carrier, or Panther AMD relationship erases the infrastructure growth rate. Management never disclosed customer concentration or backlog cancellation terms.
The Tell: Analyst Tim Savageaux asked directly: 'do you guys think you can grow faster than 30% overall in '26? Because that was the comment.' CFO Litchfield responded: 'we don't guide the whole year, and we're not going to do it here... I'm confident that we can outgrow the industry.' He deflected the same claim the CEO made in prepared remarks without repeating or quantifying it.
Friction Level: MODERATE_FRICTION — Both sides agree infrastructure is scaling and broadband ($58M, 43% of revenue) is declining. The disagreement is whether infrastructure growth outruns the base decline, and whether Keystone ASPs near $22 imply real pricing power or #3 entrant discounting.
2025Q4 · Jan 30, 2026
5 of 5 readings agreed5 HOLDRead the full call
Infrastructure +76%, Q1 Guide Implies Sequential Decline at Midpoint
Infrastructure is genuinely scaling and will become the largest segment in 2026, with 28-week optical lead times and margin expansion from mix shift into PAM4 DSPs and storage accelerators. But $58M broadband (43% of revenue) is guided down for the year, the Q1 midpoint implies a sequential decline, and the CFO won't repeat the CEO's growth claim. The operating leverage is real, the narrative ahead of it is not fully backed by forward numbers.
Analyst Tim Savageaux asked directly: 'do you guys think you can grow faster than 30% overall in '26? Because that was the comment.' CFO Litchfield responded: 'we don't guide the whole year, and we're not going to do it here... I'm confident that we can outgrow the industry.' He deflected the same claim the CEO made in prepared remarks without repeating or quantifying it.
CEO states design wins 'positioning us to grow faster in '26 than we did in '25.' CFO asked the same question twice by analyst Tim Savageaux and retreated both times to 'confident that we can outgrow the industry.' Q1 midpoint of $135M vs Q4 actual of $136.4M implies sequential decline while the growth narrative peaks.
28-week lead times in optical with backlog covering the first half of 2026. CFO states 'backlog that we have, it's in a much better position... particularly in the optical side... you've already got backlog, we're pushing to get some upsides.' Real demand pull, though cancellation terms never disclosed.
Non-GAAP gross margin 59.6% in Q4, guided 58-61% in Q1, CFO confident exiting year at 60s vs 50s. OpEx guided up only 4-5% against infrastructure revenue growing 76% YoY. Mix shift into optical DSPs and storage accelerators drives operating leverage in reported numbers, not projections.
Self-admitted #3 PAM4 entrant who 'got late' in the U.S. market. Keystone late, then leap to Rushmore 1.6T, then 'Sky' at 400G/lane. Each hop compounds miss risk while Broadcom and Marvell iterate on the node in front of them. CEO acknowledges incumbency advantage directly.
PON 'control plane conduit' design win, rack management serial transceivers, microcontroller telemetrics are real but small-dollar wins. CEO repeatedly frames against 'hundreds of millions of dollars' market sizing while declining to provide his own number: 'I don't want to provide market sizing at a level that we need to ascertain.'
Infrastructure became the growth engine in 2025 (30% full year, 76% Q4), scaling toward single largest revenue contributor in 2026. Keystone PAM4 ramping at major hyperscale data centers in US and Asia, Panther storage accelerators doubling, PON gaining share. AI datacenter optical buildout is a non-reverting demand driver.
+35
NEUTRAL"CEO delivers confident, specific answers with forward commitments (Keystone $100-130M, 4-6M units, Panther doubling). CFO gives clean operational numbers but evades the growth trajectory question twice, retreating from CEO's claim to 'confident that we can outgrow the industry.' Confidence and evasion coexist."
Specific unit volumes, revenue ranges, margin targets, and segment breakdowns throughout. Backlog quality, book-to-bill, and customer concentration metrics never disclosed. Growth claim CFO won't support with a number.
Q1 2026 print: revenue landing at $130M vs $136.4M in Q4, plus OFC in March with Rushmore announcements and electrical retimers for scale-up Ethernet. Keystone ASP clarity emerging through H1.
Single-customer loss in Keystone hyperscaler ramps, PON Tier 1 carrier, or Panther AMD relationship erases the infrastructure growth rate. Management never disclosed customer concentration or backlog cancellation terms.
“Total revenue for the fourth quarter was $136.4 million, up 8% from $126.5 million in the previous quarter and up 48% from $92.2 million in the fourth quarter of 2024.”
“We currently expect revenue in the first quarter of 2026 to be between $130 million and $140 million.”
Figures are extracted from the transcripts and checked in code: every quote appears word for word. Descriptive only: none of this feeds the verdict.
Q4 revenue $136.4M, up 48% YoY. Infrastructure $47M, up 76% YoY, becoming the largest segment in 2026. Non-GAAP gross margin 59.6% with CFO guiding exit year in 60s. OpEx up only 4-5% against revenue multiples of that. Inventory down $8M, DSO 31 days, operating cash flow positive. Optical has 28-week lead times with backlog covering H1 2026. Keystone guided $100-130M with explicit upside language and 4-6M unit volume. Panther accelerators doubling in 2026. $75M buyback authorized, $20M executed. Multiple dated catalysts: OFC in March (Rushmore, electrical retimers), Tier 1 carrier ramping Q2-Q3, new customer ramps through the year. The Street still models this as a cable/broadband story while infrastructure scales past it.
Q1 midpoint of $135M vs Q4 of $136.4M is a sequential decline in the quarter management calls its strongest entry point. Broadband at $58M (43% of revenue) guided down for the full year with DOCSIS 4 recovery not starting until H2 2026. CEO says 'grow faster in '26 than '25'; CFO asked twice and won't repeat it. GAAP operating loss of -$15M (11% of revenue) vs non-GAAP +16% (16% of revenue), a $34M gap from $28.1M SBC plus $6M acquisition costs. Annualized dilution ~$112M vs $20M quarterly buyback. Operating cash flow of $10.4M doesn't cover the $20M buyback. Keystone ASP math: $100-130M across 4-6M units implies $17-32, midpoint ~$22. Self-admitted #3 PAM4 entrant. No book-to-bill, no backlog cancellation terms, no customer concentration disclosure. Unexplained arbitration attached to buyback commentary with no size given.
Say/Do Gap
CEO claims 'grow faster in '26 than '25' in prepared remarks. CFO asked twice by Savageaux, retreats both times to 'confident that we can outgrow the industry.' Q1 midpoint implies sequential decline. The growth claim the CEO leads with is not one the CFO will defend with a number.
Backlog Quality
28-week lead times and 'robust bookings' asserted without book-to-bill ratio, cancellation terms, or backlog-to-revenue figure. CFO volunteered lead times instead of backlog metrics when asked about optical demand durability.
AI Label Without Substance
PON control plane, rack management transceivers, and microcontroller telemetrics are real design wins with small dollars. CEO frames against 'hundreds of millions' market sizing then explicitly refuses to provide his own: 'I don't want to provide market sizing.'
Generation-Skipping Execution Risk
CEO admits 'within the U.S., just the timing of our product offerings, we got late as the #3... fighting to get to the #1 or #2 takes a little bit of a taller order and incumbency has incredible value.' Keystone late, Rushmore next, Sky beyond. Each leap compounds miss risk.
Capacity Ceiling
28-week optical lead times and acknowledged supply chain tightness dismissed as 'not concerned' with 80% of business 'not exposed.' Optical is where the growth is. If supply constrains the ramp, the Keystone upside case compresses.
Unsustainable Trend Confidence
Peak positive language ('robust,' 'exciting,' 'excited,' 'confident' dozens of times) while the largest segment ($58M, 43% of revenue) is guided down for the year and Q1 implies a sequential decline. Zero hedging on the call.
HOLD. Operational execution is real: inventory down, DSO tight, positive operating cash flow, margin expanding on mix shift, 28-week optical lead times pulling product out the door. The infrastructure rotation is structural. But the near-term setup has binary risk: Q1 midpoint implies sequential decline, broadband troughs through H1, and the CFO won't back the CEO's growth claim with a number. Cannot size the risk until Q1 print confirms or denies the ramp. Size entry after Q1 and OFC in March, when Keystone ASPs and Rushmore traction become visible. If Q1 lands at low end and Keystone ASPs confirm ~$22 discounting, the growth narrative re-rates down.