Copper connectivity stays in Nvidia's racks longer than anyone modeled, a win for Amphenol and Credo. Per @SemiAnalysis_, the Kyber NVL144 rack is delayed more than 12 months to 2028 because its PCB midplane cannot be made at scale, and the NVL72x2 backup was cancelled. Co-packaged optics will not arrive until the Feynman generation, so the shift away from copper in scale-up moves out with it. x.com 

Credo's interconnect opportunity grows as campus-scale AI clusters require more links to validate and integrate. The OCP 2026 meeting takeaway is that complex requirements put interconnect at the center, with both copper and optical links set to benefit. x.com 

Astera Labs' CXL integration thesis faces a threat if Nvidia builds a CXL-attached memory architecture: Jensen would not use Astera for it. Such an architecture could make training and data sharing across nodes easier. Nvidia's Feynman might also get a better I/O die to avoid switch hops, as Google did with TPUs. x.com  x.com 

Astera Labs gains as AI clusters grow from hundreds to tens of thousands of chips, because that scale demands the high-bandwidth, low-latency interconnect it sells. Its PCIe, CXL, fabric switching and optical products serve hyperscalers and GPU and TPU accelerators, with deep ties to Nvidia, Microsoft and Amazon. The Scorpio line is moving into smart fabrics, opening a merchant scale-up market projected at $20B by 2030, and hyperscaler orders and design wins make revenue highly predictable. … Show more

Inference's move toward edge devices and dense, memory-hungry racks should drive demand for specialized chips, links, power delivery and cooling. McKinsey projects inference at 60% of AI computing demand by 2030. Ambarella makes on-device AI chips; Astera Labs links GPUs, memory and networks; Credo supplies short copper connections; Rambus earns from memory interfaces; Navitas reduces power waste; Modine cools dense racks; and Super Micro builds inference-specific server racks at volume. x.com  … Show more

Credo looks exposed as networking becomes AI's next bottleneck and Nvidia locks down optical fiber for scale-up, scale-out and scale-across links. MicroLED is readily available and a good contender for scale-up and potentially scale-out, with many Asian suppliers pivoting into it. x.com  x.com 

Copper scale-up is dead, making optics hyper bullish, especially Lumentum. Nvidia's Rubin NVLink Nyquist cut from 56 to 47 GHz signals failure: reflections matter, and bidirectional links make them worse. Feynman will have to use 16-wavelength DWDM ring-modulator CPO at roughly 50-64G NRZ. x.com  x.com 

Astera Labs' chip connects every GPU in AI servers, positioning it to ride a data-center buildout that is just getting started. Quarterly revenue reached $392 million in Q2 2026, versus $65 million in Q1 2024. x.com 

Astera Labs carries a huge premium for flawless execution against a CXL market that is small near term. Switch-based CXL memory pooling capturing $15-16B in 2026-2028 is low probability: Marvell guides its XConn switch business to only about $100M by fiscal 2028, PCIe included. Pooling software likely stays immature into 2028, per a Samsung white paper, and Nvidia is contesting the warm KV-cache tier. Astera does CXL best, but SK hynix and Micron offer better risk-adjusted returns. x.com  x.com 

Semtech benefits from Nvidia's strategy of scaling up with copper for as long as possible. At Computex 2026, Jensen Huang said to use copper wherever you can and optics only where you must. Nvidia's Blackwell NVL72 racks use thousands of copper cables rather than fiber to link GPUs within a rack, and keeping data intact over that copper relies heavily on signal-integrity chips from Semtech's active copper cable alliance. x.com