Episode Details

MaxLinear, Inc. (MXL) Q2 2026 — MaxLinear (MXL), the fabless communications-chip maker that has reinvented itself as an AI data-center optical play, reported an explosive fiscal Q2 2026 (quarter ended June 30, 2026): revenue of $168.8M (+55% YoY, +23% QoQ) beat estimates, and non-GAAP EPS of $0.35 topped the ~$0.33 consensus (vs $0.02 a year ago). The engine was Infrastructure — the optical AI data-center business — up 145% YoY to $85.0M, now 50% of revenue (Keystone PAM4 DSP ramping for 800G, with a 1.6T generation on the roadmap). Management guided Q3 revenue to $210–220M (~+27% QoQ), well above the Street. But GAAP told a soberer story: a -2.5% operating margin (a small operating loss) and just $0.02 GAAP EPS, with $27.5M of stock-based comp (~16% of revenue) doing much of the work in the 'adjusted' numbers, and H1 free cash flow roughly breakeven-to-slightly-negative. After a ~7x run off a $12.77 low to ~$91, the stock actually FELL ~4% after the print. At ~11x forward sales and ~50x forward non-GAAP EPS, even our probability-weighted DCF lands fair value near $60 — and even the Street's ~$70 average target sits ~23% below the price. Our call: HOLD, 2/5. MaxLinear is one of the market's wildest turnaround stories — a sleepy, fabless broadband chipmaker out of Carlsbad, California that has reinvented itself almost overnight as an artificial-intelligence data-center play, and whose stock has gone up roughly seven-fold in a year (from a $12.77 low to ~$91). Fiscal Q2 2026 (quarter ended June 30, 2026) was genuinely explosive: revenue of $168.8M jumped 55% YoY and 23% QoQ, beating estimates, and non-GAAP EPS of $0.35 topped the ~$0.33 consensus (against just $0.02 a year ago). The engine is Infrastructure — the optical AI data-center business — up 145% YoY to $85.0M and now fully half of revenue (it was 32% a year ago), driven by the Keystone PAM4 DSP ramping at multiple hyperscalers for 800G, with a 1.6T generation on the roadmap and a long-term $3B infrastructure ambition. Management guided Q3 revenue to $210–220M (~+27% QoQ), well above the Street, and non-GAAP operating margin has climbed from 7% a year ago to 22%. So the debate isn't whether the turnaround is real — it clearly is. It's the price. On a GAAP basis MaxLinear still ran a small operating loss (-2.5% margin) and just $0.02 of EPS, with $27.5M of stock-based comp (~16% of revenue) doing much of the work in the 'adjusted' numbers; H1 free cash flow was roughly breakeven-to-slightly-negative on a working-capital build; the balance sheet carries ~$124M of debt against ~$65M of cash (net debt) and ~$319M of goodwill; and Silicon Motion's ~$160M+ arbitration claim over the collapsed 2023 merger still lingers off the balance sheet. After a ~7x run, the stock trades near 11x forward sales and ~50x forward non-GAAP EPS — and it actually FELL ~4% after this beat, the classic 'priced for perfection' tell. Even leaning our probability-weighted DCF toward the optimistic AI-optical compounder path (owner earnings ramping toward ~$390M by FY31), fair value lands near $60 — our base case is closer to $37, and even our full bull case barely reaches today's ~$91. Our call: HOLD, 2/5 — a real turnaround, already priced in. Notably, even Wall Street's own average price target (~$70, on a nominal 'Buy') sits ~23% below the price, so we're broadly aligned with the Street on value; the 'Buy' rating is a momentum label. We're not shorting a company growing 55%, but there's no margin of safety here — we'd want a pullback toward the $50s–low $60s. Not financial advice. THE CALL: HOLD (2/5, A REAL AI-OPTICAL TURNAROUND, ALREADY PRICED IN — NO MARGIN OF SAFETY AFTER A 7x RUN) — base-case value ~$60 vs ~$91 today. What to watch: hard evidence MaxLinear is keeping real optical-DSP share as the market moves to 1.6T, non-GAAP operating margins pushing toward the high-20s, and — critically — GAAP profits and actual free cash flow finally showing up, which would justify the premium multiple and