MACOM
Real AI-Optical Content, Priced for Perfection

Snapshot: Executive Summary

MACOM Technology Solutions is a high-performance analog, RF and optical (photonic) semiconductor company serving three end markets, Industrial & Defense, Data Center, and Telecom. It has a real AI tailwind: MACOM supplies the analog and photonic front-end (laser drivers, transimpedance amplifiers, modulator drivers, and its own indium-phosphide photodetectors) into the 800G, 1.6T and 3.2T optical modules that wire AI clusters together. The business is executing well: FY2025 revenue $967.3M (+32.6%), record Q3 FY2026 revenue $342.2M (+35.8% YoY, +18% QoQ), adjusted EPS $1.40 (+55.6% YoY), adjusted gross margin 59.7%, record book-to-bill 1.6, and Data-Center growth raised to ~74% in FY2026¹.

The setup, and the catch:

• The AI-optical thesis is real, and it keeps delivering. The stock ran from a Dec-2025 close near $171 to a $418.90 high (May 27, 2026), based in the high-$200s through the summer, then popped on the Q3 beat to ~$311, +81% YTD and roughly doubled over the trailing year. The narrative (AI drives optical connectivity, MACOM owns differentiated content) is intact and durable.
• The valuation is still full, though earnings caught up. Q3's +55.6% adjusted-EPS jump compressed the forward P/E from ~45x to ~40x even as the stock rose. At ~$311 MTSI trades at ~40x forward EPS, ~18x trailing sales, ~14.6x forward sales and ~60x EV/EBITDA. GuruFocus' GF Value model still pegs fair value near $173, well below spot². These multiples still leave little room for a growth disappointment.
• We stay deliberately below the Street. The sell-side is overwhelmingly Buy with an average target near $396. We treat that ~$396 as the bull case, not the base. Our call rests on valuation discipline after a parabolic move, not on any doubt about the franchise.

This is a valuation call, not a thesis call. We rate MTSI Hold, 12-month PT $360 via a forward EV/Sales model on FY2027E revenue (Bull $420 / Bear $252) — raised from $300 as the Q3 beat and the ~74% Data-Center guide lift the FY27+ base. The probability-weighted blended fair value is ~$351, just below the base PT and ~13% above the current $310.82, so the risk/reward is balanced-to-modestly-favourable, upside we respect but do not chase at a full multiple. We would turn constructive on (a) a pullback that restores real downside protection, or (b) evidence that Data-Center optical is compounding fast enough to lift the FY27+ revenue base materially above $1.65B. Until then: own the franchise, respect the multiple.

Tactical: MTSI is trading at $310.82, ~16% below our $360 12-month base target. Q3 FY2026 was a clear beat (revenue $342.2M, +35.8% YoY; adjusted EPS $1.40, +55.6%) and we raise the target to $360, but we stay disciplined at ~40x forward EPS / ~18x sales after a +81% YTD run, deliberately below the Street's ~$396. The Data-Center optical story is real (FY26 growth raised to ~74%, record 1.6 book-to-bill, InP photonics, owned fabs) but well paid for. Rating Hold: the ~16% gap is upside we respect rather than chase.

Investment Thesis

Rating: Hold. Probability weights: 30% Bull / 45% Base / 25% Bear → blended fair value ~$351, just below the base PT $360 and ~13% above the current $310.82. The 45% base weight reflects confidence in the structural AI-optical content story; the 30% bull maps to the Street's ~$396 average target; the 25% bear respects the genuine valuation risk after a +81% YTD move to a full multiple. This is a valuation call, not a thesis call. It pairs analytically with the AI-optical-interconnect read on Marvell and Credo, where MACOM owns the analog/photonic front-end rather than the optical DSP.

Business Overview

MACOM reports three end markets: Industrial & Defense, Data Center, and Telecom. Unlike a pure-play AI name, MACOM is a diversified analog/RF and optical franchise, which is both its ballast (defense is steady and high-margin) and, to the bull, its lever (Data Center is compounding fast). In Q3 FY2026 Data Center became the largest segment: the mix was Data Center $137.6M (40%, +40% QoQ), Industrial & Defense $133.4M (39%), and Telecom $71.3M (21%). Over FY2025 the split was roughly I&D $419M (43%), Data Center $293M (30%, +48%), and Telecom $255M (26%, +41%).

What is inside each segment

• Industrial & Defense (~39%): RF, microwave and millimeter-wave components for radar, electronic warfare, satcom, aerospace and defense, plus test-and-measurement and industrial. Built on GaN, GaAs and silicon. Program-driven, high-margin, and relatively acyclical, the ballast of the business. IMS 2026 brought new GaN power amplifiers (a 125W X-band device, a 1.5kW S-band device) for radar.
• Data Center (~40% and rising): The AI growth engine, now the largest segment. High-speed analog and photonic content, laser drivers, transimpedance amplifiers (TIAs), modulator drivers, clock-and-data-recovery, and MACOM's own indium-phosphide (InP) photodetectors and lasers, sold into 400G, 800G, 1.6T and 3.2T optical modules. This is the whole AI thesis.
• Telecom (~21%): Optical and RF content for carrier, metro and long-haul networks, PON access, wireless infrastructure and satellite. Cyclical; recovering off a trough. Carries China and export exposure.

Why the segment mix matters

MACOM is not a one-story stock. Industrial & Defense gives it a stable, high-margin base that most AI-optical peers lack, which is a genuine quality attribute and part of why the multiple is defensible at all. But the growth, and therefore the multiple, is driven by Data Center. The flip side is margin and cyclicality: the Data-Center optical business is early-cyclical and, management has repeatedly said, "a very volatile market" with fast ramp-ups and ramp-downs. The bull case needs the AI-optical ramp to keep compounding; the bear case is an AI-capex digestion that hits the fastest-growing, highest-multiple leg.

AI Optical & Data Center

This is where the growth lives, and it is a genuinely differentiated position. Crucially, MACOM's AI exposure is not the optical DSP (that socket belongs to Broadcom and Marvell). MACOM supplies the high-performance analog and photonic front-end that sits around the DSP inside an optical module: laser drivers, transimpedance amplifiers (TIAs), modulator drivers, clock-and-data-recovery, and, distinctively, its own indium-phosphide (InP) lasers and photodetectors. As AI clusters push port speeds from 800G to 1.6T to 3.2T, MACOM's dollar content per module rises even though it never touches the DSP.

The InP and owned-fab advantage

MACOM is one of very few players that designs and manufactures its own indium-phosphide photonics, the material system behind high-speed lasers and photodetectors (for example 200G-per-lane InP photodetectors for 800G and 1.6T links). It runs its own fabs, including an expanding RF/photonics facility in Research Triangle Park (RTP), North Carolina. Vertical integration in InP and GaN is hard to replicate and gives MACOM a defensible content position, though owned fabs are a double-edged asset: a margin lever in an up-cycle and a fixed-cost drag if utilization falls.

The module-customer model

MACOM sells components into optical transceiver modules built by the module makers, Coherent, Lumentum, Innolight, Eoptolink and others, rather than owning the end socket. This diversifies customer risk (its content is spread across many module vendors) but also means MACOM rides the whole AI-optical cycle rather than a single contracted program. It has demonstrated TIA and laser-driver interoperability with Broadcom for links up to 1.6Tbps and launched a portfolio of 1.6T solutions, signals that its content is designed into the next optical generation.

The debate: content growth vs the cycle

The bull question is whether AI-optical content growth sustains Data-Center revenue above ~50% into FY2027. If it does, MACOM grows into a good deal of its multiple and the model moves toward the bull $420. The bear question is what happens when hyperscaler optical buying digests: the analog/photonic front-end is early-cyclical, and at a full multiple the de-rating is fast. New architectures (linear-pluggable optics, co-packaged optics) could either expand or compress MACOM's bill-of-materials, a genuine uncertainty. The franchise is excellent; the Hold is about what is already in the price.

Competitive Position

MACOM competes on two fronts. In Data Center optical it sits in the analog/photonic layer alongside Semtech and, adjacently, the retimer/DSP players (Broadcom, Marvell, Credo). In RF/analog (its Industrial & Defense and Telecom segments) it competes with Qorvo, Skyworks, Analog Devices, NXP and Microchip. The defining tension: MACOM is far smaller (~$1.3B revenue) than most of these peers, but it is differentiated in InP photonics, GaN and owned fabs, and it now trades at a valuation in the same premium tier as the much larger AI-optical names.

Where MACOM sits in an optical module

MACOM vs the peer set

The read

MACOM's differentiation is real: InP photonics plus GaN plus owned fabs is a moat that the mature RF peers and the fab-lite DSP players do not share, and a stable defense base makes the earnings less binary than a pure AI name. That is why it earns a premium multiple at all. But it trades at ~14.6x forward sales, roughly in line with Marvell despite a fraction of the revenue and none of the optical-DSP socket, and at a vast premium to the RF peers at ~2x. The bull says the content growth justifies it; our lean is that a smaller, more cyclical, component-into-module supplier should not carry the same premium as the socket owners without a larger margin of safety. As Data-Center growth becomes visible in the numbers, that premium is more likely to normalize than to expand.

Financial Health & Trends

Q3 FY26 print highlights (record quarter)

Record revenue with Data Center now the largest segment, adjusted operating margin to 31.5%, record book-to-bill 1.6, and a Q4 guide of $415 to $425M with adjusted EPS ~$2.00. Operationally this is a clean beat-and-raise. One earnings-quality caveat: FY2025 posted a GAAP net loss of ($0.73) per share despite positive GAAP operating income, driven by a one-time ~$193.1M loss on extinguishment of debt from the Q2 FY25 convertible refinancing, so the gap between adjusted and GAAP results, while narrowing (Q3 GAAP EPS $1.28 vs adjusted $1.40), is worth watching.²

Capital Allocation

Capital priorities are growth-first: fund analog/RF/photonic R&D and design-in engineering, invest in owned-fab capacity (notably the RTP facility), and keep a conservative, net-cash balance sheet. MACOM pays no dividend and buybacks are modest; the emphasis is on capacity and IP for the AI-optical and defense roadmaps. Liquidity is roughly $665M of cash and short-term investments against ~$0.41B of debt, so net cash is ~$0.25B, which is added (not subtracted) in the valuation.

The RTP fab and CHIPS Act

MACOM took operational control of the Research Triangle Park (RTP), North Carolina fab under a long-term (roughly 12-year) lease, adding on the order of 30% more RF/photonics capacity and supporting up to ~$70M of CHIPS Act funding. Owned fabs are strategically valuable for InP and GaN (they are hard to source externally and enable differentiated content), but they carry utilization risk: they are a margin tailwind while the AI-optical and defense cycles run hot, and a fixed-cost headwind if demand softens. The fab-absorption cycle is part of why gross margin dipped and is now recovering.

The FY2025 debt refinancing

In Q2 FY2025 MACOM refinanced its convertible notes, which produced a one-time ~$193.1M loss on extinguishment of debt and drove the full-year FY2025 GAAP net loss even though the underlying operating business was profitable. The charge is non-recurring and left the balance sheet in a cleaner, net-cash position, but it is a reminder that headline GAAP results and the adjusted numbers the market focuses on can diverge sharply.

R&D / NRE intensity

Sustained investment in RF/analog IP, GaN and InP process technology, and high-speed optical design is the core strategic spend. Underinvesting risks losing content position as optical architectures evolve toward 3.2T, co-packaged and linear-drive designs. Capital return is deliberately secondary while the AI-optical and defense roadmaps are in their build-out phase.

Valuation Overview

Why EV/Sales over P/E for MTSI

MACOM is in a revenue-ramp and margin-recovery phase, and its GAAP earnings are distorted by one-time items (the FY25 debt-extinguishment loss) and a low non-GAAP tax rate, so forward P/E is noisy. EV/Sales smooths those dynamics and lets us compare cleanly against the optical and RF peer sets. The cross-checks agree that the stock is expensive: ~40x forward EPS (roughly ~56x on FY26 adjusted EPS near $5.5) and ~60x EV/EBITDA, though Q3's earnings ramp compressed both from a year ago.

The crux of the Hold

At ~14.6x forward and ~18x trailing sales, ~40x forward earnings and ~60x EV/EBITDA, MTSI is priced for sustained, near-flawless execution. Three facts define the call: (1) the multiple has rerated to roughly double its own pre-AI range and sits well above the analog/RF-semi median; (2) GuruFocus' GF Value model implies fair value near $173, well below spot, a data point we do not take literally but do not ignore; and (3) the sell-side's ~$396 average target already embeds the bull case. Our base anchors on FY2027E revenue $1.65B at a still-premium 17.0x EV/Sales, which holds the re-rated multiple rather than normalizing it. The thesis is intact and durable, and the Q3 beat let earnings catch up, which is why we lift the target to $360 while staying below the Street.

EV/Sales Scenario Model

FY2027E revenue × EV/Sales multiple = enterprise value; plus net cash (~$0.25B) = equity value; divided by ~78.6M diluted shares = the 12-month price target. We use FY2027E (the fiscal year ending ~Oct 2027) as the base because the PT is a 12-month-forward target, so no extra discount is applied. Net cash is added because MACOM is net-cash, not net-debt.

Sensitivity grid: FY27E revenue × EV/Sales multiple (PT in $)

Quick PT calculator

Risk / Reward calculator

Upcoming Catalysts

Risk Factors

• Valuation, the #1 risk and the core of the Hold. After +81% YTD, MTSI trades ~40x forward earnings, ~18x trailing sales and ~60x EV/EBITDA, with GuruFocus' GF Value near $173. Even after Q3's earnings ramp compressed the multiples, the stock is priced for sustained, flawless execution; multiple compression is the dominant near-term risk, and most of the bear-case downside is de-rating rather than a business decline.
• Data-Center / AI-capex cyclicality. The fastest-growing, highest-multiple segment is also the most fragile. Management calls data center "a very volatile market" with fast ramp-ups and ramp-downs; a hyperscaler optical digestion in FY27 would hit the growth engine and the multiple together.
• Scale and competition. MACOM (~$1.3B revenue) competes against far larger, better-capitalized players, Broadcom, Marvell, Analog Devices, plus Credo and Semtech in AI interconnect. It sells components into modules rather than owning the DSP socket, so pricing pressure and design-win shifts at the module makers flow straight through.
• Distributor / channel concentration. Two distributors were each over 10% of FY25 revenue (12.4% and 11.2%), distributors were ~32% of revenue overall, and the top-10 customers were ~57%. Channel stocking can overstate true end demand, and a single relationship shift can move results.
• Gross-margin sustainability. Adjusted GM depends on a favorable, high-value mix (defense plus high-speed optical); GAAP GM runs several points lower (~55%). A mix shift toward lower-margin, higher-volume data-center parts, or under-utilization of the owned fabs, would pressure margins.
• Earnings quality (GAAP vs adjusted). FY2025 posted a GAAP net loss on a one-time ~$193.1M debt-extinguishment charge, and the headline adjusted numbers rely on a low non-GAAP tax rate. The gap is wide; the valuation rests on the adjusted figures.
• Owned-fab utilization. InP and GaN fabs are a differentiation asset but a fixed-cost drag if demand softens, the flip side of the RTP capacity lever.
• Geopolitical / supply. China end-demand exposure in telecom/optical and export-control dynamics for advanced components add tail risk.

Bull vs Bear Debate

Resolution = Hold: own the thesis, respect the price. The franchise is differentiated and the AI-optical content growth is durable, and the Q3 beat justified a higher target ($360), but at ~$311 the base case sees only ~16% upside, the blended fair value (~$351) sits ~13% above spot and below the Street, and the next leg up requires the Data-Center line to keep surprising. We turn constructive on a pullback or a clear acceleration above our base.

Technical Analysis

Trader's view

• Price $310.82; 52-week range $118.16 to $418.90; the $418.90 high printed May 27, 2026. +81% YTD 2026, roughly doubled over the trailing twelve months.
• Pulled back ~33% from the May high into the high-$200s after a parabolic Feb-to-May run, based through the summer, then popped on the Q3 beat to ~$311 (~26% off the high), with extremely high realized volatility.
• Key support: the ~$280 shelf, then the ~$250 level. Key resistance: the ~$350 May breakdown level, then the $419 high.
• R/R from ~$311 to the $420 bull target with a $252 stop is ~1.85:1, acceptable but not compelling, and consistent with the Hold: our $360 base PT is ~16% above spot while the bear is ~19% below. A pullback toward a real margin of safety would improve the setup further.
• Volatility note: realized vol is very high, typical of AI-thematic semis at a stretched multiple.