# Where can an AI startup advertise without a Series A budget? > AI startups bid against the best-funded companies online for the same words. The channels that still work on a few hundred dollars, with honest costs for each. Published 2026-09-01 · 7 min read · advertise-an-ai-startup, advertising, ai-startups Canonical: https://bidsurvivor.space/blog/where-to-advertise-an-ai-startup --- There is a specific arithmetic problem behind the question of where to advertise an AI startup: the words that describe your product are the same words being bought by companies that raised nine figures last quarter. WordStream's benchmark set puts the all-industry average cost per click at roughly **$5.42**, and the AI-adjacent end of the software category sits comfortably above that. On a budget measured in hundreds, buying those clicks is not a test. It is a donation. What follows is the full map of what is left, cheapest first. (Disclosure: published by [BidSurvivor](https://bidsurvivor.space), which appears in the paid tier, labelled.) ## The free tier is unusually good in this category For most products, "submit to directories" is filler advice. For AI tools it is genuinely the highest-return hour available, because the directories rank. **AI tool directories.** There's An AI For That, Futurepedia, Toolify and a long tail of smaller ones own a large share of the "AI tool for X" queries that people still type into a search box. The filter is simple and worth applying ruthlessly: open the directory, search for a category you know, and see whether the directory itself appears in the results. If it does not rank, listing on it is bookkeeping, not marketing. Budget an evening, do the twenty that rank, ignore the two hundred that do not. **Launch platforms.** [Product Hunt](https://www.producthunt.com/launch) still puts AI products in front of an audience that came specifically to find new ones. It is free, and it is one shot per product — which means the preparation matters more than the day. Have the landing page, the demo and the first comment written before you schedule anything. **The communities around what you build on.** Whatever model, framework or workflow your product sits on top of has a subreddit, a Discord and a forum full of people solving the problem you solved. Post the output, not the pitch. This is slow, it does not scale, and it produces the only early users who reliably stay. None of these is advertising in the sense of buying inventory. All three are the reason a small AI company can be visible at all before it has money to spend. The same free tier exists a layer down, for a mobile AI app rather than a web product — store listing, directories and communities, mapped out in [where to advertise a mobile app for free](/blog/where-to-advertise-a-mobile-app-for-free). ## Why the default paid channels punish this category Two things have happened at once, and they compound. The first is price. Search advertising allocates by auction, so the clearing price for a keyword reflects the deepest pocket bidding on it. In a category where dozens of funded companies are all describing themselves with the same eight nouns, that pocket is very deep. The arithmetic of a small search budget is worked through in [what $200 buys you on Google Ads](/blog/what-200-dollars-buys-you-on-google-ads), and every line of it gets worse when the keyword has "AI" in it. The second is that the click itself is worth less than it was. Search Engine Land's reporting on AI Overviews found [informational queries losing a large share of their clicks](https://searchengineland.com/google-ai-overviews-cut-search-clicks-report-471497) once an answer is generated above the links. That hits AI startups twice over: it thins the organic traffic you were relying on to make the paid channel affordable (the [AEO checker](/tools/aeo) tests whether your pages are structured to survive that shift), and it means the paid slot is competing for a smaller pool of attention at a higher price. Paid social has a different failure. The creative conventions of AI marketing converged about two years ago — the same gradient, the same "10x your workflow" line, the same demo GIF — and an audience that has scrolled past four hundred of them treats the fifth as furniture. Nielsen Norman Group has been documenting [that avoidance reflex since 1997](https://www.nngroup.com/articles/banner-blindness-original-eyetracking/); it is not a creative problem you can out-design at three figures of spend. Both channels work. They work at five figures, with a person whose job it is to run them, over a quarter. They are not a way to find out whether your positioning is right. ## The paid channels a three-figure budget survives **Niche newsletters.** AI newsletters multiplied faster than their advertisers did, which is good news for a buyer. Below the famous ones, placements in the **$50–300** range reach lists of a few thousand people who opted in to reading about exactly this. Marketplaces like [Paved](https://www.paved.com) list rates openly; individual writers will often quote you a lower one by email. Do the division before you buy: price divided by (expected visitors × your signup rate) is your cost per signup, and you can compare that to anything. The [ad budget calculator](/tools/ad-calculator) does that arithmetic for any channel. **Reddit.** Sub-dollar clicks with community-level targeting, if — and only if — the creative can survive being read by people who resent advertising. Our verdict, including the failure modes, is in [Reddit ads for startups](/blog/reddit-ads-for-startups). **A time-boxed slot — ours, so weigh it accordingly.** [BidSurvivor](https://bidsurvivor.space) is one board with two slots a day, twelve hours each, sold by open auction. The first brand into an empty slot pays **$0**, with no account and no card. Taking a slot somebody already holds starts at $1.00 and no bid can exceed $5,000. Every visitor starts with $100 of house credit before signing up for anything, so a bid inside that costs nothing out of pocket. Opens and click-throughs are published per brand, which is the actual point for an AI startup: you can see what everyone else's placement did before you spend on your own. The scale is small and the numbers saying so are public. Use it to find out which one-line description of your product earns a click, not to fill a funnel. If the shape of the problem is "Google prices me out", the fuller comparison of substitutes — Bing, long-tail search, newsletters, directories — is in [Google Ads alternatives for small budgets](/blog/google-ads-alternatives-for-small-budgets). This piece is the map of the category; that one is the map of the escape route. ## What to measure, and in what order The mistake specific to AI startups is measuring signups. Signups are cheap to get and tell you almost nothing, because "try a free AI tool" is a low-cost decision that people make idly. Measure three things instead, in this order. **Click-through on the pitch line**: does the sentence describing your product make a stranger click? That is testable this week for nothing. **Second session**: of the people who signed up, how many came back within seven days without an email prompting them? **A completed piece of real work**: did anyone produce an output they kept? A channel that produces the third at any price is worth more than one producing a hundred of the first. ## The sequence Directories this week, because they compound and cost only time. A prepared launch when there is something worth launching. A free slot or a small newsletter placement to settle which sentence about your product actually earns attention — the [rules of the board](/blog/how-bidsurvivor-works), ours included, are written down so you can check them before you bid. Then, and only then, paid search, on the exact keywords your cheaper tests already proved. The order matters more than the budget. Every channel above gets cheaper once you know your own pitch works, and every one of them gets ruinous if you buy it to find that out.