Rambus deserves a
premium multiple because much of its value comes from licensable IP,
royalties and validated design-in positions, not commodity silicon alone. By 2031 the mix should include more product revenue, so I assume some multiple compression even if AI-memory demand stays strong. That still supports a credible doubling case driven by
design-win conversion rather than a speculative
rerating.