Wild thought on $AVGO supply chain chess:
If Broadcom locks up $SIVE laser capacity for the next few years under the excuse of "needing more laser capacity," they could accidentally-on-purpose bottleneck their main ASIC competitors.
Who gets squeezed? $MRVL (Celestial AI), AlChip, MediaTek, GUC (Ayar Labs CPO), and other first-gen co-packaged optics deployments. Oh, and many of these players just happen to be aligned with $NVDA.
So here's the kicker: maybe it's worth Broadcom spending some spare change to lock up the remaining merchant laser supply entirely. Not because they need it all. But because controlling the chokepoint is the move.
This is the kind of vertical integration flex that doesn't show up in earnings calls but shows up in competitor delays 18 months later. If you're building custom AI silicon and your optics supplier suddenly can't deliver, your roadmap just got torched.
Watch the laser supply chain. It's not sexy, but it's leverage.
If Broadcom locks up $SIVE laser capacity for the next few years under the excuse of "needing more laser capacity," they could accidentally-on-purpose bottleneck their main ASIC competitors.
Who gets squeezed? $MRVL (Celestial AI), AlChip, MediaTek, GUC (Ayar Labs CPO), and other first-gen co-packaged optics deployments. Oh, and many of these players just happen to be aligned with $NVDA.
So here's the kicker: maybe it's worth Broadcom spending some spare change to lock up the remaining merchant laser supply entirely. Not because they need it all. But because controlling the chokepoint is the move.
This is the kind of vertical integration flex that doesn't show up in earnings calls but shows up in competitor delays 18 months later. If you're building custom AI silicon and your optics supplier suddenly can't deliver, your roadmap just got torched.
Watch the laser supply chain. It's not sexy, but it's leverage.