MaxLinear Lifts 2026 Optical Outlook to $250 Million-$270 Million as 1.6T Revenue Pulls Forward; Management Targets Teens DSP Share Within Two Years
Q3 2026 earnings call, October 22, 2026: Infrastructure now larger than all other segments combined
MaxLinear reported third quarter revenue of $218.1 million, up 29% sequentially and 72% from $126.5 million a year earlier. Fourth quarter guidance of $240 million to $260 million points to another large step-up. The more important change is in the business mix. Infrastructure revenue grew 210% year over year to about $125 million and now exceeds the rest of the company combined. The company raised its 2026 optical data center revenue outlook to $250 million to $270 million. CEO Kishore Seendripu said bookings visibility extends beyond 2027 and that "we expect another significant step-up in revenue next year."
1.6T Timing Moves Forward, and Content per Port More Than Doubles
The most material update was the timing of the 1.6 terabit cycle. Last quarter management pointed to a mid-2027 start. Seendripu now expects an initial design win to begin shipping in the first half of 2027, with a stronger ramp in the second half. He said the initial volume is unclear but "meaningful in the sense that it sets up the second half very nicely." Commitments cover both the Rushmore DSP and the companion TIA.
Pricing is the second driver. Stifel's Tore Svanberg noted the market expects ASPs to double in the move from 800 gig to 1.6T. Seendripu confirmed this and said that the doubling, which "historically never stuck when you change speeds," is holding this time. With a TIA added to the bill of materials, MaxLinear expects content growth of more than 2x per port. He guided the TIA attach rate at roughly 80% near term and 100% or more in about 18 months. He said only one other TIA vendor meets the performance and power requirements at 1.6T, and that MaxLinear intends to be the second.
CFO Steve Litchfield said that 2026 optical growth is almost entirely 800 gig. 2027 is the start of 1.6T, with a much larger 2028 in which the mix is "substantially 1.6." The 800 gig business could be sustained until about 2029. Management declined to endorse Svanberg's estimate of a $400 million-plus DSP run rate for 2027. Litchfield did say he expects MaxLinear to outgrow the market, and a Northland analyst's suggestion of near-doubling did not draw pushback.
Share Targets: Teens Within Two Years, Parity With a Competitor Is the Goal
Asked about the 5% to 7% DSP share range management has cited at investor conferences, Seendripu said the figure is backward-looking and that "we are well above that number." He said share should reach "the teens for sure" in two years. He added that within 12 to 24 months MaxLinear "could be neck and neck with 1 of our 2 competitors. That's the hope." These are management aspirations, not guidance, and the company would not endorse a sell-side extrapolation to a $1 billion optical business. Litchfield did say the market-size assumptions behind that math, including about 200 million transceivers in 2027, "make a lot of sense."
The customer base is broadening. No customer exceeded 10% of revenue in the quarter. Seendripu said revenue spans all the major module makers, and that US hyperscalers are moving to 1.6T ahead of their Chinese counterparts. He said US data centers are only beginning to ramp 1.6T "outside of an NVIDIA and a Google, in that order." He declined to break out the US share of current 800 gig revenue.
Supply Is the Constraint, but Management Rejects the Double-Ordering Thesis
Susquehanna's Christopher Rolland asked whether longer lead times of 26 to 28 weeks could encourage customers to over-order DSPs. Seendripu dismissed the idea: "I don't think so at all." He said nobody in the chain is signing long-term agreements, and that hyperscalers qualify DSP and optics combinations as dedicated supply chains. He characterized this as discipline in the supply chain rather than hoarding. He said optics supply is easing while the real bottleneck is foundry capacity for DSPs, with "a fab capacity shortage" because new capacity takes years to build. He said the optical pluggable market of roughly 120 million units could reach about 200 million in 2027 or 2028.
One internal constraint is worth noting. Seendripu said MaxLinear is "bottlenecked by our own ability to support multiple quals simultaneously," because 10 module makers may want qualification at once, often for the same data centers. This is unrelated to wafer supply. It suggests engineering resources, not demand, may limit how fast 1.6T design wins convert to revenue.
CPO Delayed, NPO and LRO Favored
Seendripu said co-packaged optics "has gotten delayed because the challenge of the optics in a hybridized co-package configuration is a very huge problem." He expects near-packaged optics to supplant it, citing serviceability, modularity and avoiding lock-in to switch or GPU ASIC vendors. He called linear pluggable optics a "very small fraction point technology" without a future, but said LRO, the half-DSP approach, is a real option in the US at 200 gig per lane and probably at 400 gig. His personal view, not a market forecast, is that by 2030 pluggables will be 60% to 70% of volume, with NPO at about 20% and other forms at about 10%. MaxLinear's TIA, driver and DSP lineup is meant to cover each scenario.
Annapurna Targets a Modest Slice of the AEC Market
Management gave a candid view of the active electrical cable opportunity. Credo holds an overwhelming share of AEC, which Seendripu put at "93% or maybe 99%." MaxLinear will sell silicon only, to cable makers or paddle card suppliers, at pricing similar to its optical PAM4 DSPs. It plans to focus on one or two US data centers and aims for 20% to 30% share in the 1.6T cycle over the next couple of years. Initial Annapurna and Washington (TIA) revenue is expected in 2027, with a stronger ramp in 2028. XGS-PON control plane deployments at a hyperscaler, which have completed qualification, are also expected to start in 2027. Panther storage acceleration revenue should roughly double this year.
Margins Improve, but Cash Flow and Working Capital Are Under Strain
Non-GAAP gross margin was 60.0% and Q4 guidance is 59% to 62%. Litchfield said margin is rising despite ecosystem cost increases, some of which are being passed to customers, and that further mix shift toward infrastructure should lift it. Non-GAAP operating margin reached 29.5%, with non-GAAP EPS of $0.62 against GAAP EPS of $0.30. Q4 non-GAAP operating expenses are guided to $65 million to $70 million.
The balance sheet is the weak spot. Operating cash flow was negative $13.6 million and cash fell to $109.9 million, reflecting a large wafer prepayment. Needham's Quinn Bolton noted prepaid assets rose about $50 million. Litchfield said he expects prepayments to continue rising as revenue projections climb. The company expanded its revolver by $70 million to $200 million, and interest expense is guided up to $3.5 million to $4.3 million in Q4. Inventory days fell to 111 from 123, and DSO rose to 31 days from 28. Growth is therefore being financed with debt capacity and tied to foundry supply commitments.
Legacy Markets: Seasonal Softness, Growth Expected in 2027
Broadband revenue was about $53 million and connectivity about $25 million, both up sequentially on fiber PON and Wi-Fi 7 gateway deployments at Tier 1 carriers. Both are guided down seasonally in Q4. Litchfield expects broadband growth in 2027 from telco spending and PON share gains, without quantifying it. Industrial and multimarket revenue was about $15 million and is guided up in Q4.