revoAIution

Stocks | AI | Semiconductors
Tracking the next great revolution
Not Financial Advice

[revoAIution @revoAIution]

Do we really have $NVDA and $AMZN at the end of a potential $SIVE supply chain through Ayar Labs?

Mapping out the NVLink Fusion ecosystem and stumbled onto something pretty crazy if it comes to fruition.

NVLink Fusion is Nvidia's open interconnect standard. The spec every serious AI infrastructure company is now building around. Not a product. A standard. Like USB but for AI chips at hyperscale.

Ayar Labs got officially certified inside NVLink Fusion in June 2026 which of course is huge for $SIVE.

And back in December 2025, Amazon confirmed Trainium4 is being built on that exact same NVLink Fusion standard. Seven months ago. Quietly sitting there.

$SIVE has been the confirmed laser supplier inside Ayar Labs since 2022.

So the chain could look like this.

$SIVE laser ---> Ayar Labs optical engine --->NVLink Fusion = $NVDA AI infrastructure and potentially $AMZN Trainium4 both building on top of it.

Two of the most powerful companies on earth building on the same standard. And $SIVE sitting at the laser layer underneath all of it through Ayar Labs.

This is not yet confirmed whether Ayar Labs optical technology will be used inside Trainium4 specifically.

Bit of a moonshot, but never say never.

[revoAIution @revoAIution]

I don’t think it’s of any relevance. I think that would have been more of a coincidence. Stockholm opened at quite low and ran down sharply within 20 minutes, most likely by sell pressure, stops, and extreme fear of the headline that was released two hours before the market opened… meanwhile news was actually dropped a few days ago that they converted.

[revoAIution @revoAIution]

$SIVE down ~18% today on a convertible loan conversion that was already disclosed in February. This is exactly why the Nasdaq dual listing cannot come soon enough.

Actual dilution is 6 to 7 percent. The loan disappears. The balance sheet gets stronger. Nothing about the business changed today.

The ~18% drop is Stockholm being Stockholm.

Thin market.
Thin liquidity.
Fear of selling pressure from a lender sitting on profit. Stop losses cascade.
Small cap overreaction.

US institutions do not panic sell 6 to 7 percent dilution on something that was already disclosed in February.

Irrational reaction. Thesis completely unchanged.

[revoAIution @revoAIution]

@IWANTTOBEHAPPYI It’s been ulgy. However, any catalyst materializes and $SIVE can have some massive days

[revoAIution @revoAIution]

@ProfitTen10 $SIVE’s main exchange is stockholm.

[revoAIution @revoAIution]

@APl2ck We will see what happens in the next 2 weeks with the unlocks

[revoAIution @revoAIution]

@Mylo14870344 I’m not. I think it’s irrational.

[revoAIution @revoAIution]

@HerschelRothman I think you’re wrong here. Time will tell

[revoAIution @revoAIution]

@Foddmann Fraction to what the Americans trade

[revoAIution @revoAIution]

@_MHequities I’ll be watching if the CEO sells shares, only because he just recently bought. The rest is noise.

[revoAIution @revoAIution]

@GSEconomics Compare that to US markets and it is still night and day. The daily dollar volume on $SIVE in Stockholm is a fraction of what a comparable US listed stock would see. That is not an opinion that is just the size difference between the two exchanges.

[revoAIution @revoAIution]

@darkstalwart Bootstrap Europe is a lender not a long term investor. They converted debt to equity. They are almost certainly going to sell every single share to lock their profit. That is just what lenders do.

[revoAIution @revoAIution]

@IWANTTOBEHAPPYI lol whatever you say

[revoAIution @revoAIution]

$AMD's next gen chip the MI500 dropping in 2027 uses GFS SCALE platform to move data using light inside the chip. $SIVE lasers are confirmed inside that platform. This connection is just sitting there quietly.

[revoAIution @revoAIution]

Scale up vs scale out...

Things just got more interesting after looking further into the $NVDA Kyber delay and possible cancellation. And why $SIVE, Ayar Labs and $GFS SCALE may now front run scale up and become the most desirable architecture right now.

Scale up = GPUs talking to each other inside a rack. Think of it like a team working in the same room passing notes across a table. Fast but limited by how many people you can fit in the room.

Scale out = entire racks talking to each other across a network using optics. Think of it like connecting multiple rooms together so the whole building can work as one. That is where the real growth in AI infrastructure is heading.

And that scale out side is exactly where Ayar Labs, $GFS and $SIVE sit.

1. Ayar Labs joined $NVDA NVLink Fusion in June 2026 with $SIVE as the confirmed laser supplier inside their optical engine since 2022.

2. $GFS SCALE platform is built specifically for this. Co packaged optics for scale out AI networking with $SIVE lasers baked into the reference design. Ayar Labs optical engines doing the same thing through NVLink Fusion.

Markets are choppy but the $SIVE thesis keeps getting stronger and stronger

[revoAIution @revoAIution]

@avshushka Ayar Labs confirmed $SIVE as their laser supplier publicly. Alchip is the confirmed backend design partner for Amazon Trainium. That is all documented.

The last link connecting Ayar Labs optics specifically into Trainium is still speculative

[revoAIution @revoAIution]

Another possible $SIVE chain to another hyperscaler? $AMZN?

Sivers laser ---> Ayar Labs light source ---> Alchip chip design ---> Amazon Trainium?

Insane at sub 2b

[revoAIution @revoAIution]

@babyfolio Scale up is delayed but if anything that just expedites the need for scale out which is exactly what Ayar Labs and $SIVE are doing

$AAOI is probably not bothered by this at all. They will just print with pluggables.

[revoAIution @revoAIution]

SemiAnalysis just posted that $NVDA Kyber has been delayed over 12 months to 2028. Sounds scary for CPO timelines but how does this actually affect $SIVE?

Maybe this is actually good news for the Ayar Labs connection?

The scale up vs scale out narrative.

Nvidia was building technology to connect their own GPUs to each other inside a rack using light instead of copper. That specific piece got pushed to 2028. That is what SemiAnalysis is referring to.

Now think about this from a $SIVE and Ayar Labs point of view.

Ayar Labs joined NVLink Fusion this june on a completely SEPERATE track.

NVLink Fusion is not about connecting Nvidia GPUs to each other. It is about connecting third party custom chips like Amazon Trainium into Nvidia's ecosystem using optics.

Two completely different things. The delay does not touch the Ayar Labs connection at all.

Now here is where it gets potentially bullish.

Nvidia just admitted they have no proven way to scale up their own GPU clusters beyond copper for now.

That leaves a gap.

I believe hyperscalers still need to scale and fast... So instead of scaling up inside a rack they have to scale out across racks using more networking and more optical connections

That scale out side is exactly where Ayar Labs and $SIVE sit.

This delay might actually be accelerating the exact use case the $SIVE and Ayar Labs connection is built for.

[revoAIution @revoAIution]

@jeongpark0509 Yes, honestly I think Jabil alone entering mass production could do that.

[revoAIution @revoAIution]

Heading into this week I think $SIVE at $1.7 billion is wrong and deeply undervalued.

And before anyone says it. Yes they have very little revenue right now. The shorts will tell you that is how to value this company. They are wrong. That is not what you are buying.

You are buying the $799 million pipeline. The six confirmed threads pointing toward hyperscalers. None of it converted yet.

Not one of its biggest partnerships has turned into real volume revenue yet. Zero. Every relationship is confirmed to exist and confirmed to be moving toward production. But almost none of it has hit real numbers on the AI and hyperscaler side yet.

That is exactly why I think the price is wrong right now.

Six confirmed breadcrumbs. Six threads pointing toward some of the most powerful companies on earth. And the stock just got cut nearly in half.

$GFS to $AMD.
Ayar Labs to $NVDA NVLink Fusion.
$JBL.
O-Net and Enablence
$POET to Lumilens
$AAPL?

If even one of these converts into real volume revenue the current valuation looks silly. If three of them land with initial orders I genuinely think a $10 billion market cap happens faster than most people expect.

That is not all six. That is not perfection. Just three of six starting to convert.

From $1.7 billion to $10 billion that is roughly a 6x.

With institutional money already in at 57 SEK from an oversubscribed raise that closed overnight.

Personally the most interesting stock in the markets right now.

[revoAIution @revoAIution]

@jeongpark0509 Not sure how you think this is baked into the current valuation?

On speculation alone (most are confirmed) this should be at least $3-4b before a single hyperscaler order lands. Just compare to other photonics companies and you'll get a better view on it.

[revoAIution @revoAIution]

@goatedmarylin Watching this closely but I think we are a bit too far out to solidify anything yet.... but agreed!