$LITE showing ecosystem forming around near-package optics. Open CPX founders include Ciena $CIEN, Coherent $COHR and Marvell $MRVL, alongside others. Publicly traded contributors: Credo $CRDO, Intel $INTC, MaxLinear $MXL, Applied Optoelectronics $AAOI, Astera Labs $ALAB, Amphenol $APH, Keysight $KEYS, TE Connectivity $TEL, VIAVI $VIAV, Lumentum $LITE.
See original
$LITE showing the ecosystem forming around near-package optics.
Open CPX founders include Ciena $CIEN, Coherent $COHR and Marvell $MRVL, alongside Molex, Samtec and TeraHop.
Publicly traded contributors shown here include
Credo $CRDO
Intel $INTC
MaxLinear $MXL
Applied Optoelectronics $AAOI
Astera Labs $ALAB
Amphenol $APH
Keysight $KEYS
TE Connectivity $TEL
VIAVI $VIAV
Lumentum $LITE
NPO reuses the existing pluggable supply chain, making it much easier to scale.
Image 1 ↗
Reasoning: Fundamentals
@KawzInvests notes OCP Summit data confirms qualification and testing remain the central gating factor for optical transceivers at scale. Celestica numbers show testing windows compressed: 400G 36 months, 800G 24 months, 1.6T 18 months. With GPU architectures refreshing every 12 months, ramping laser fab capacity alone doesn't solve it. Mentions $VIAV $AEHR $COHU $FORM $TER.
See original
Recent data at the OCP Summit confirms what has been clear for a while... qualification and testing remain the central gating factor for optical transceivers at scale.
Celestica's numbers show how testing windows have continuously compressed with each generation:
> 400G: 36 months
> 800G: 24 months
> 1.6T: 18 months
With GPU architectures refreshing every 12 months and 224G per lane leaving no margin for error, the core structural thesis hasn't changed.
Ramping laser fab capacity doesn't solve the problem on its own if transceivers can't be validated and qualified fast enough to match the silicon cadence.
$VIAV $AEHR $COHU $FORM $TER
Image 1 ↗
Reasoning: Fundamentals · Medium term Specific
@michaelsikand highlights a crazy stat from $CLS VP: 70% of engineering resources dedicated to validation. As optics moves closer to expensive AI compute silicon, cost of failure rises, making testing and qualification increasingly critical. Mentions $KEYS $AEHR $VIAV $FORM.
See original
Wow crazy stat from $CLS VP
“70% of our engineering resources are dedicated towards doing validation."
As optics moves closer to expensive AI compute silicon, the cost of failure rises, making testing and qualification increasingly critical.
$KEYS $AEHR $VIAV $FORM
Image 1 ↗ Image 2 ↗
Reasoning: Fundamentals · Medium term Specific
Fund flows or trade records mentioned in posts reflect only the account’s own claims, are marked “Per the post” and are unverified.