Is Nvidia Assembling the Parts for Its Next Inference Platform?