The most direct way site speed affects AI citation is through crawl completeness. When an inference engine crawler sweeps a site, it allocates limited time per domain. A site that responds slowly serves fewer pages in the same crawl window. A site that responds in under 50ms consistently serves the complete sitemap in a single session.
The IEO Engine flat-file PHP architecture is optimized for crawl completeness. When ClaudeBot ran its 130-page deep sweep, every page response completed before the crawler's timeout. A WordPress site under the same crawl load would likely produce timeouts on some pages, leaving those pages uncrawled in that session.
A slow or timeout-producing site introduces uncertainty into the citation decision. If a retrieval system cannot reliably access a page, it is less likely to cite that page — the uncertainty about whether the content will be available when the citation is followed reduces citation confidence.
A site that responds consistently and completely produces higher citation confidence. The inference engine has never encountered an access failure on this domain, so there is no uncertainty about whether the content will be available.
Site speed has a floor below which it starts to hurt and a ceiling above which additional improvement produces no measurable citation benefit. GoDaddy shared hosting with a flat-file PHP architecture typically produces response times in the 30-100ms range — well above the floor. Premium cloud infrastructure with the same architecture produces 10-30ms responses. The difference is not meaningfully distinguishable for AI citation purposes.
The IEO Engine documented deployments demonstrate this: the methodology produces documented citation outcomes on GoDaddy shared hosting. The hosting is adequate. The architecture is what matters.