One pool, every crawler
Google added a line to its crawl-budget documentation stating that the crawl capacity limit is shared across all crawlers. The consequence it spells out is that high demand from one crawler, with Googlebot-Image given as the example, reduces the capacity available to the others.
That reframes a familiar problem. Crawl budget usually gets discussed as though the crawler fetching your HTML works from its own allowance, which made image weight a page-speed topic and nothing more. If one pool feeds every Google crawler on the property, a catalog serving thousands of oversized product images is spending capacity that would otherwise reach new and updated pages.
The scoping matters as well. Google's crawl-budget guidance names enterprise sites above a million pages and medium sites above ten thousand pages that change rapidly. A three-hundred-product store is not the audience for this document.
What a 304 does
The added recommendation reads: "Use HTTP caching: Support 304 (Not Modified) HTTP status codes. If a page hasn't changed since Google last crawled it, returning a 304 code tells Google to reuse the cached version, saving your server bandwidth and resources."
A 304 is a response with no body. The server is telling the client that the copy it already holds is still current, so there is nothing to send. Googlebot skips the download and moves on, and the request costs a fraction of a full fetch.
The catch for most stores is ownership. Returning 304 correctly depends on the server honoring conditional requests, which on a hosted platform is decided by the platform and whatever CDN sits in front of it. That is a configuration question for a host, not a toggle in a store dashboard.
What to do about it
Ask your host what it returns for an unchanged page
Conditional-request handling is a server behavior, so on BigCommerce, Shopify, or any managed host the answer belongs to the platform and the CDN in front of it. Send the question to support rather than hunting for a setting. If full 200 responses go out every time, you now have a documented Google recommendation to attach to the request.
Count what image crawling is spending
If crawl capacity is shared, image requests compete with product pages for it. Open the Crawl stats report in Search Console and read the by-file-type and by-response breakdowns. A store fetching far more images than HTML has a candidate explanation for slow discovery of new products.