AMD’s post-Enosemi (May 2025 defensive acquisition) substrate-layer roadmap depends on glass-substrate workflow tooling that doesn’t exist in commercial EDA. ChipletOS ships it today.
MI300 ramped on TSMC SoIC + organic. MI400 candidates target glass interposer for HBM4-stacked AI training. ChipletOS HBM4 cal head closes 90.3% of BEM-vs-Palace gap on 4507 corner cases at 50 GHz (v2 retrain, 2026-05-05). 6/6 regimes calibrated, avg 74% gap closure.
Xilinx-line FPGA-on-glass-interposer roadmap. ChipletOS UCIE cal head 81.6% gap closure at the high-density chiplet-to-chiplet interconnect regime — one of 4/6 elite (>80%) regimes after v2 retrain.
AMD’s Enosemi acquisition (May 2025) was defensive vs Nvidia photonic switch. Glass substrate is the integration substrate for high-density optical I/O. ChipletOS PROV 4 photonics + glass-TGV models cover the workflow.
Defensive acquisition because Nvidia/Marvell were ahead on photonic switch. Same pattern: substrate-layer IP that locks competitors out of the next-gen package. Genesis owns the EDA-adjacent workflow lock on glass substrates.
Up to $5.5B PRE-REVENUE. Glass substrate is the integration substrate. Marvell didn’t wait for Celestial to build a public moat — neither will Nvidia/AMD wait for ChipletOS.
Same defensive pattern. Without ChipletOS, AMD’s Instinct MI400+ packaging team integrates 3-5 separate point tools to cover Cadence Sigrity / Synopsys SiP / Ansys HFSS / Keysight ADS / Siemens HyperLynx. With ChipletOS = single workflow lock across all 5.
$700M, ~10× rev, workflow lock-in premium. The same posture Genesis ships: cross-EDA workflow lock + Genesis-side cert + customer-side native signoff per per-vendor _NATIVE_VALIDATION_DEFERRED.md doc.