Instagram has quietly stopped talking about hashtags at all
Instagram's creator guidance, posted August 23, 2026, pushes captions, not hashtags, as the discovery lever: captions "add relevant context, spark conversation, and tell search engines what your content is about." That guidance doesn't mention hashtags once, consistent with years of Instagram downplaying them as a reach driver. One uncorroborated blog claims Instagram capped hashtags at five in December 2025; treat that number with suspicion until a second source confirms it, and lead with captions.
TikTok treats hashtags as metadata, not a multiplier
TikTok's ranking explainer groups hashtags with captions and sounds as one "video information" signal, a categorization input, not a discovery trigger. Confident rules like "use exactly five hashtags" show up in uncited SEO content, not in anything TikTok has published.
X's (Twitter's) own disclosed ranking code shows no hashtag-weighted signal
X open-sourced its ranking model in May 2026. The scoring system matches on semantic embeddings, not literal hashtag strings, so there's no explicit hashtag-weighted factor in the code. That's absence of evidence, not a denial, but it supports the long-running "hashtags barely matter on X" folklore.
LinkedIn is the one platform where hashtags still do something different
LinkedIn hashtags work partly as followable topic channels, a genuinely different mechanic from Instagram or TikTok. Long-standing guidance suggests roughly three to five relevant hashtags; anything more specific is unconfirmed until LinkedIn states it directly.
Threads hasn't built a hashtag culture at all
Threads' entire 2025–2026 product investment (Dear Algo, Your Algo, the Discovery and Communities hubs) is building non-hashtag mechanisms for feed control. There's no mature hashtag guidance to speak of, which is itself the guidance: don't import an Instagram hashtag strategy wholesale.
The pattern underneath all five platforms
TikTok's captions and sounds, Instagram's caption-first push, and X's embedding-based matching are all moving discovery toward understanding your actual words, and away from literal hashtags. Hashtags aren't dead, but treating them as a distinct lever in 2026 misreads where these systems have gone.
A hashtag routine that survives scrutiny
- Write the caption to be found first. Put your actual keywords in real sentences; that's what discovery systems now read.
- Use three to five accurate, specific hashtags as a minor assist, not the main event. Any round number works; the precision some guides claim isn't in what platforms have published.
- Treat LinkedIn hashtags differently: as topic channels worth following, not just tags.
- Don't copy a hashtag block across platforms unchanged. What reads as normal on Instagram reads as spam on X or LinkedIn, which is the specific problem Postey's AI rewrite tools exist to solve.
Any hashtag rule stated as an exact number, with no source, is very likely someone repeating a nine-year-old Twitter blog post. Write for the caption, use hashtags lightly, and demand a source for any number.
