Semtech二季度业绩和三季度指引显著超预期，AI数据中心业务继续加速

Well, I had high expectations going in to the print/call.

But man, they exceeded on basically every front.

Going into this print they guided:

-

$328m rev

-

$0.61 adjusted EPS

-

54% adjusted gross margins

-

35% sequential data center growth ($97m rev)

They delivered:

-

$341.9m rev

-

$0.71 adjusted EPS

-

54.5% adjusted gross margin

-

24.4% adjust opearting margin (vs 21.9% guided midpoint)

-

Data center rev $100m (up 39% sequentially and 91% YoY.

And they guided data center to grow another 45% sequentially and roughly 160% YoY in Q3. So ~$145m next quarter.

I mean come on. Wow.

So the DC sequence:

Q1 - $71.6m

Q2: $100m

Q3: ~$145m

And the crazy thing is that we aren’t even at the inflection period that I am so excited about.

I think it’s possible for them to hit $500m in data center rev in FY27 (current FY). That would require a ~$164m Q4. I think they can do that.

More on their Q3 guide, $410m in total rev, 58.3% adjusted gross margin, 31% adjusted operating margin and $1.05 adjusted EPS.

Now some call notes.

-

800G industry expectations going from ~50m to 80-90m. Existing customers are still increasing orders and their fiberedge backlog for 800G remain healthy

-

800G fiberedge share was around 18% two years ago and now above 50%. Now designed into every major module provider in their target market, with several beign sole source. Expects fiberedge share at 1.6T to exceed 50% by the end of the fiscal year.

-

1.6T FiberEdge qualified earlier than mgmt had assumed. They were being conservating because qual time is uncertain but now customers '“want parts tomorrow”. So Q3 should be the first Q with a meaningful 1.6T contribution.

-

The AAOI read through supports the component demand picture we were talking about. AAOI said DSP and TIA availability is constraining how many 800G and 1.6T transceivers it can ship. Semtech supplies TIAs and they set prepped their capacity early enough to support both planned ramps and some drop in demand.

-

Pricing looks good. Availability currently carries more weigh than pricign with no near term erosion expected on orders already sitting in backlog. Passing through most cost increases and new products also carry higher margins

-

I really wanted strong NPO commentary and I got it! Engaged in roughly 10-15 programs across module manufacturers. Positioning their TIA arrays, linearized lasar arrays and future photodiode arrays as a more complete solution.

-

They are expanding their photodiode offering. Hired an industry leading. New team is working with SMTCs existing TIA engineers and is already engaged with key customers. Goal is a co optimized photodiode array and TIA solution for later optical geneartions.

-

CopperEdge products through 1.6T are ready for volume deployment. Lead hypercaler volume deployment begins in Q4.

-

Multiple hyperscaler and ODM engagements around onboard and active backplane applications. Some area approaching the finish line in the near term. Design in activity through 3.2T.

-

HieFO gain chip biz shipping in volume. Original three anchor customers are increasing demand, they are adding other customers as well. Custom confidence in HieFo improved after Semtech bought it,

-

High power CW lasers sent to 5-6 module manufacturers with positive feedback around tech specs. Cw laser rev for transceivers expected to begin durign H1 FY 2028.

-

They started initial SOA and customer qualification expected over the next couple of months. A SOA boosts optical power inside a photonic path and becomes useful in architectures with splitters, couplers, modulators and other sources of optical loss. No specifi rev or production timing, but a nice addition.

-

Capacity expansion on track. Acquired a nearby already facilitized fab and expects the additional space to increase photonics fab capacity by roughly 3-4x by year end.

-

Large content step up with 3.2T. $80-$90 per transceiver. Deisgn window beginning in around 12 months and meaningful deployment in ~2 years.

-

I’m really stoked on margins. Q3 should be 58.3% consolidated adjust gross. And excluding cellular modules it will be 63.9%. 31.0$ operating margins. Product mix should continue to drive margins higher as well.

-

Fy27 data center target fully booked. Fy28 “over 70% there”.

-

Near tear capacity sufficient for planed ramps. But might fall short in 2nd half FY28 which is why they are pushing for more front end wafers, testers, OSAT capacity, manufacturing partners etc. Includng prepayments and capex arrangements.

-

LoRa is a legit 2nd growth engine. Rev $58m, up 31% sequentiall and 58% YoY. Q3 guide 15% sequential.

Now for financial modeling and valuation, here is where I stand:

I am modeling using FY28 scenarios here.

My base is: $1.7B total rev, 64.5% gross margin, 38% operating margin, $5.44 adjusted EPS. So today were trading at ~24x P/E.

Bull: $2.00B rev, 65.5% gross margin, 41.0% operating margin, $6.51 adjusted EPS. ~20x P/E.

Based on that, I think $190 is very reasonable for a base case, $220+ for bull.

While I am still being quite conservative on any AI / Semis, I did end up buying more shares in the low 130s and plan to add more on dips.

I am surprised how muted this earnings reaction was.

**Financial Disclaimer:** This content is for informational and educational purposes only and should not be construed as financial, investment, legal, tax, or other professional advice. Nothing contained herein constitutes a recommendation, solicitation, or offer to buy or sell any security or financial instrument. I am not a registered investment adviser, broker-dealer, or financial planner, and the information presented is based on my personal research, opinions, interpretations, and assumptions. I may hold positions in securities discussed, and those positions may be increased, reduced, or closed at any time without notice. Any financial models, valuation ranges, revenue estimates, earnings projections, market-share assumptions, price targets, or other forward-looking statements are estimates only and may differ materially from actual results. Company guidance, management commentary, industry forecasts, third-party research, and other sources are believed to be reliable, but accuracy and completeness cannot be guaranteed, and information may become outdated. Investing involves substantial risk, including the possible loss of principal. Individual circumstances, objectives, time horizons, and risk tolerances vary, so readers should conduct their own due diligence and consult qualified financial, legal, or tax professionals before making investment decisions. Past performance is no guarantee of future results.
What a quarter