NeutralVRTVertiv Holdings, LLCVertiv Holdings, LLCAMDAVGOBroadcomBroadcomBEBloom Energy CorporationBloom Energy CorporationCRDOCredo Technology Group HoldingCredo Technology Group HoldingINTCIntelIntelMRVLMarvellMarvellMUMicron TechnologyMicron TechnologyNBISNebius Group N.V.Nebius Group N.V.NVDANVIDIANVIDIASNDKSandisk CorporationSandisk Corporation

9/23/2026 on

Best watchlist in this market IMO $NVDA — AI GPUs $AMD — AI CPUs/GPUs $MU — Memory (DRAM/HBM) $SNDK — Flash Memory $INTC — CPUs/Foundry $AVGO — Custom AI Silicon $CRDO — Networking DSP $MRVL — Custom Silicon/Networking $NBIS — Neocloud $BE — Data Center Power $VRT — Data Center Cooling/Power $AAOI — Optical Transceivers $AXTI — Semi Substrates $AEHR — Test Equipment $SPCX — space exploration

NeutralVRTVertiv Holdings, LLCVertiv Holdings, LLCAMDCRDOCredo Technology Group HoldingCredo Technology Group HoldingMRVLMarvellMarvell

9/20/2026 on

Watching these 10 names across AI compute, edge AI, networking, optics, manufacturing and power: 1. $CRDO My first name to investigate for the AI connectivity angle. AECs, optical connectivity and high-speed interconnects are becoming increasingly important as AI clusters scale. Credo says its AECs can materially reduce power consumption versus laser-based alternatives. 2. $AAOI Higher risk, but potentially much higher operating leverage if its optical products gain share as AI data-center bandwidth requirements increase. This is more speculative than CRDO. 3. $LITE A direct way to play optical infrastructure, including optical circuit switching and photonics. Its collaboration with Marvell is specifically targeting next-generation AI infrastructure. 4. $MRVL Less speculative than several names above, but still interesting because it sits directly in the connectivity layer. Marvell is targeting switching, optical interconnect, PCIe and CXL memory solutions across AI infrastructure. 5. $COHR Another photonics/optical play. Potential upside comes from increasing bandwidth requirements and next-generation optical architectures. 6. $CLS The less-obvious infrastructure play. It participates in the manufacturing and systems layer behind hyperscale/AI infrastructure. Worth watching for continued AI-related revenue growth. 7. $QCOM This is the edge-AI + data-center convergence play. Qualcomm recently announced a multigenerational collaboration with Amazon involving customized AI silicon and optical connectivity for AWS. 8. $AMD A broader AI-compute opportunity spanning data centers and increasingly AI PCs/edge devices. The upside case doesn’t depend entirely on edge AI. 9. $VRT Not an edge-AI company, but potentially one of the strongest picks-and-shovels around the physical constraint: power + cooling. As AI racks become increasingly power-dense, thermal and electrical infrastructure become more important. 10. $GEV The power-grid side of the thesis. AI data centers need electricity, and GEV has exposure to generation and electrification infrastructure. S&P Global expects AI/data-center demand to contribute to increased investment in power infrastructure.