网红分析师长推复盘 Sivers Semiconductors(SIVE)激光器产能/客户与美股上市预期，无新增事实仅为渠道推断

Last major $SIVE disclosure was 100M CW DFB laser capacity Q4 2027. And a new foundry partner (that makes 2 including Win Semi).

Expected $JBL revenue coming in H1 2027 (with lasers for Jabil's 1.6T LRO scaling up H2 2027).

There's channel checks along the way like CIOE where Sivers was engaged with Chinese pluggable makers + laser ASPs being raised.

70mW/100mW ranges were the most tight to Chinese pluggable makers, which was surprising.

Biggest industry read through was $LITE selling out into 2029, making agreements for 2030, and having hyperscalers pay for capacity expansion for 2031.

I don't think anyone has enough information to do any sort of accurate modeling for $SIVE like you can for $AAOI given they don't release product breakdown + capacity targets by quarter.

The 200mW NPO lasers Sivers is sampling is likely different ASP than your 70mW lasers shipped to pluggable makers. And your custom CPO lasers are different pricing too.

Like all sampling/qualification players from $AEHR to $AAOI back in 2025, usually the risk/reward comes in whether qualifications convert and what new customers they have, which is always a risk.

In a bull case, if $SIVE comes online with 300m laser supply (2:1 implied ratio, already has "tremendous" supply online from likely US fab) during a structural shortage.

Pluggable players like $JBL, 6 other vendors.

$GFS as the reference design laser for their CPO/NPO platform.

Your minor players vendors like $POET, $AEVA and other players end up scaling.

And all the other custom products they hinted they were working on in the past for CPO years back.

I think the probability that they cook is higher since there's so many ongoing qualifications during a bottleneck.

We've already seen ownership changes from Swedish investors to US/International retail from $SIVE shareholder disclosures.

My guess is that trends institutional from retail before US listing next year amid the market correction.

Think people forgot institutions have already financed their expansion, which was a 700M SEK round 100%+ ago (multiple times oversubscribed).

And that $SIVE uses primarily external fabs, so it's not the same as a lot of the major financing concerns affecting Neoclouds.

Anyway, up to management execution, I'm not running Sivers so no clue what other companies are working with them now, whether it's Cambridge or some other US player like Semtec.

TLDR:

- Just need to wait for earnings for more disclosures around qualifications + partners, only so much public can guess from channel checks and supplier mapping.

Not quite the same as $LITE CEO where you can see the guy on Fox news or Bloomberg interviews every week (which is a good thing)