The claim
"Foundry compatible" is one of the most overused phrases in semiconductor startup decks. It is often meant to reassure investors that the company can scale without owning a fab.
But compatibility is not binary. A process can look compatible in a slide and still fail at design-rule integration, tool availability, thermal budget, contamination rules, yield learning, test methodology, reliability, or business access.
Warning signs
Be skeptical when the claim is based on generic CMOS language, a university cleanroom flow, a single module demonstration, or a vague statement that "any foundry can run this."
The harder questions are specific: Which foundry? Which node? Which module? Which design rules? Which tools? What masks change? What reliability stress is required? Who owns the yield-learning loop? What is the foundry's incentive to support this?
What good evidence looks like
Good evidence includes a real process integration plan, credible design-rule assumptions, known tool compatibility, contamination and thermal-budget analysis, test strategy, reliability path, and evidence that a foundry or manufacturing partner would actually engage.
Foundry compatibility must also be commercially compatible. A technically possible flow is not enough if the economics, volumes, qualification burden, or support model make it unattractive.
TIGRE lens
TIGRE pressure-tests whether a foundry claim is real, partial, aspirational, or misleading. The question is not whether the process can be imagined in a fab. The question is whether it can support a manufacturable, qualified, sellable product.