They raised optical guidance $50M on Keystone alone, prepaid foundries $40M, and generated $4.8M of operating cash on $168.8M of revenue.
Thesis: Keystone is the only 5nm PAM4 DSP shipping in volume at 100 Gb/lane, and the entire $50M raise comes from that one family with six months of order visibility plus $40M of wafer prepayments against the backlog. The bottleneck is 5nm allocation and TSMC/OSAT support, and management explicitly says they have it while competitors are capacity-constrained, which is why share gains (not just TAM) drove the raise. The balance sheet is thin but the prepayments are cash committed against non-cancellable backlog, so this is a working capital squeeze funding share capture, not a distress signal.
Verdict: LONG — Conviction: MEDIUM
Catalyst: Q3 prints the $215M midpoint with non-GAAP GM at 60%, then Rushmore completes quals through 2H26 with initial 1.6 Tb revenue in 2027, stepping ASPs above the 800G mix.
Key Risk: $93.7M cash against $85M of new obligations in one quarter, with $4.8M operating cash flow. A pullback in 800G optical demand meets a company that has prepaid foundry commitments it cannot flex.
The Tell: When Tore Svanberg pointed out they are approaching 30% operating margin, CFO Litchfield replied: 'you're absolutely right... I know we've been here before... We want to continue to show this kind of long-term sustainability.' He volunteered a prior-cycle comparison no analyst asked for, admitting this margin level has been reached and not held before.
Friction Level: MODERATE_FRICTION — Both sides accept the demand is real. Disagreement is over whether a 2.8% cash conversion rate and $93.7M cash balance invalidate a beat-and-raise cycle, or whether GAAP losses with 21.6% of revenue in SBC are just the cost of funding a share-gain ramp.
2026Q2 · Jul 23, 2026
5 of 5 readings agreed5 LONGRead the full call
One Product Carries the Whole Raise, Cash Funds the Wafers
Keystone is the only 5nm PAM4 DSP shipping in volume at 100 Gb/lane, and the entire $50M raise comes from that one family with six months of order visibility plus $40M of wafer prepayments against the backlog. The bottleneck is 5nm allocation and TSMC/OSAT support, and management explicitly says they have it while competitors are capacity-constrained, which is why share gains (not just TAM) drove the raise. The balance sheet is thin but the prepayments are cash committed against non-cancellable backlog, so this is a working capital squeeze funding share capture, not a distress signal.
When Tore Svanberg pointed out they are approaching 30% operating margin, CFO Litchfield replied: 'you're absolutely right... I know we've been here before... We want to continue to show this kind of long-term sustainability.' He volunteered a prior-cycle comparison no analyst asked for, admitting this margin level has been reached and not held before.
Revenue +55% YoY, +23% QoQ, Q3 guide of $210-220M implies another 24-30% sequential jump. Optical 2026 guide raised from ~$160M to $210-230M. Non-GAAP GM guided 60%, ahead of the 65% long-term path management reaffirmed.
Keystone is the only 5nm SoC shipping in volume at 100 Gb/lane. All incremental growth is 800 Gb PAM4 with ASPs above corporate average. GM expands to 60% even as CFO confirms wafer, packaging, and test costs are all rising.
Six months of order visibility, lead times long, and $40M of purchase obligations added in one quarter via wafer prepayments against backlog running a couple of quarters out. Management placed orders for Q1 already.
Prepaying foundries at premium prices while guides are raised is management committing cash against the demand forecast, not just narrating it.
CFO concedes wafer, packaging, and test cost increases across the board, says they are 'cautious' on pass-through, and is paying premiums to secure supply. TSMC/OSAT hold the relationship despite MXL prepaying.
Six months of visibility is two quarters. Management layers 'multiyear growth phase' language on top of an order book covering half a year.
Announced 'return to GAAP profitability' at $0.02 EPS while GAAP operations ran at -2%. The positive print came from below-the-line items; the operating business still loses money.
Every analyst opened with congratulations, one said 'Wow.' Management used 'exciting inflection,' 'exceptionally well positioned,' 'transformational' with zero hedging or macro qualifiers on the entire call.
+85
CONFIDENT"Direct numeric answers throughout: 60% GM guide, 6-month visibility, named product timelines (Rushmore 2027, Annapurna 2027 revenue, volume 2028). Zero hedging, zero macro qualifiers. CFO volunteered the $50M raise was all Keystone before being asked twice."
Segment-level revenue, margin guides, obligation deltas, inventory days, and visibility all disclosed with specifics. Gaps: no dollar quantification of wafer prepayments, no ASP figures, no customer names.
Q3 prints the $215M midpoint with non-GAAP GM at 60%, then Rushmore completes quals through 2H26 with initial 1.6 Tb revenue in 2027, stepping ASPs above the 800G mix.
$93.7M cash against $85M of new obligations in one quarter, with $4.8M operating cash flow. A pullback in 800G optical demand meets a company that has prepaid foundry commitments it cannot flex.
“We currently expect Q2 2026 revenue to be between $160 million and $170 million.”
“Total revenue for the second quarter was $168.8 million, up 23% from $137.2 million in the previous quarter, and up 55% from the $108.8 million in the second quarter of 2025.”
Companies tend to guide low: the median company we cover reports +1.7% against its guidance midpoint, so a small beat is the norm. A transcript can drop part of a guided range, so read both quotes.
“We currently expect revenue in the third quarter of 2026 to be between $210 million and $220 million.”
Changed less than 95% of calls
Wording compared with the Apr 23, 2026 call, ranked among 98 calls reporting on Q2 2026. A measure of how much the story moved, not a signal about the stock.
Figures are extracted from the transcripts and checked in code: every quote appears word for word. Language change is computed from word frequencies, with no model involved. Descriptive only: none of this feeds the verdict.
Revenue +55% YoY to $168.8M, infrastructure +145% YoY to $85M and now the largest segment. The $50M optical raise is 100% Keystone, meaning share gain inside a growing TAM, not one large design win. Keystone is the only 5nm PAM4 DSP in volume at 100 Gb/lane, with 40% lower power than competition, which is why foundries and OSATs are 'incredibly supportive' while supply is tight. Inventory days improved 128 to 123, DSO flat at 28, no channel stuffing. GM guided to 60% while absorbing input cost inflation. Wafer prepayments up $40M against backlog that runs into 2027. Rushmore inherits an installed footprint: management says they are designing across the board with all optical module players on 800G, and those same customers are 'eagerly working towards' 1.6 Tb. Rushmore, Washington, Annapurna all sampling with 2027 revenue, volume 2028. Street still models MXL as a broadband company. High conviction long.
Operating cash flow of $4.8M on $168.8M revenue is 2.8% conversion. GAAP operating margin is -2%; the $0.35 non-GAAP EPS strips out $36.5M of SBC and performance equity, 21.6% of revenue, roughly $146M annualized against a business producing under $20M of annual operating cash. Purchase obligations +$40M and other obligations +$45M in one quarter against $93.7M of total cash. The 'return to GAAP profitability' headline rests on below-the-line items. CFO admits wafer, packaging, and test costs are all rising and they are only 'cautiously' passing them through, so TSMC owns the cost curve. Backlog visibility is six months. No new revenue product until 2027, and the entire raise rides on one family. Peak-confidence language with zero caution anywhere on the call. Hold with caution.
Say/Do Gap
CEO headline claims 'return to GAAP profitability' at $0.02 EPS, but GAAP loss from operations was -2% of revenue. The profit is below the operating line; the operating business still loses money.
Zero Pricing Power Despite Constraints
CFO states wafer, packaging, and test costs are rising across the board, they are prepaying at premiums, and are only 'cautious' about pass-through. GM expansion to 60% depends entirely on infrastructure mix holding while input costs climb.
Backlog Quality
CFO puts visibility at 'on or about six months.' Two quarters of order book underwrites language about a 'multiyear growth phase' and a stock priced for transformation.
Unsustainable Trend Confidence
Zero hedging, zero macro qualifiers, unanimous analyst congratulations, and 'exciting inflection' / 'transformational' vocabulary across both prepared remarks. This is peak tone with no acknowledgment of risk.
LONG, medium conviction. The asymmetry is in the bottleneck: 5nm PAM4 DSP allocation is the constraint, MXL holds it, and management prepaid $40M of wafers to defend it. Share gain drove the raise, not just TAM, and GM expanded to 60% while every input cost rose. The balance sheet is the tax on the position: size it so a two-quarter delay in 800G demand does not force dilution at the wrong price. Do not add on the Q3 print alone. Add on Rushmore qual completion into 2H26, which converts the 2027 revenue story from narrative to booked orders.