A CAPTCHA is a challenge designed to distinguish human visitors from automated ones, and proxy traffic is disproportionately likely to trigger it.
Why proxies specifically get flagged
Free and datacenter proxy IP ranges are heavily associated with bots and scraping, simply because so much automated traffic legitimately comes from them. A site seeing a request from a known proxy range reasonably treats it as higher risk than an average residential IP.
It's not personal
Getting a CAPTCHA doesn't mean a site has specifically identified you — it means the IP or traffic pattern matched a broad risk profile that the site challenges by default, regardless of who's actually behind it.
Reducing how often it happens
Slower, more human-like request timing and using less heavily-used proxy IPs (residential over datacenter, less-scraped free proxies over frequently listed ones) both reduce, though don't eliminate, how often CAPTCHAs appear.