A $0 newsletter about AI tools: the beehiiv stack and the three ways it earns

A newsletter is the only asset in this stack you own outright. A follower count belongs to a platform that can throttle you overnight. A list of email addresses goes with you.

This is the full build for an AI-tools newsletter that costs $0 to run: beehiiv on the free plan, one repeatable weekly issue, one short-form platform pointed at one signup link, and the three ways newsletters in this category actually earn.

One thing up front about the money. Every figure below is a public rate for this category, taken from published sponsorship rate cards, from what newsletters visibly charge for a paid tier, and from how beehiiv describes its own ad network and paid recommendations. They are an anchor so you do not invent one. They are not a forecast of what you will earn.

What you'll have when you're done

• A live newsletter on beehiiv's free plan with a hosted signup page and a working send
• A five-block issue template you can fill in about 90 minutes a week
• A twelve-week topic bank, so week four is not the week you stall
• One growth loop: one short-form platform, one link, one CTA, repeated
• The three revenue mechanics set up in the order they actually switch on
• [SCREENSHOT: the beehiiv dashboard after the first send, subscriber count and open rate visible]

Before you start

• A beehiiv account on the free plan. It carries a hosted signup page, a newsletter website, unlimited sends and a subscriber ceiling in the low thousands. Plan limits move, so open the pricing page before you build anything around a number.
• One thing you are genuinely doing with AI tools. Not a topic you researched. A thing you use. The whole format below is a byproduct of work you were doing anyway, which is the only version of this that survives to week twelve.
• One short-form platform you will post on three times a week. One. Not four.
• About two hours to launch. Then about three hours a week: 90 minutes writing, 90 minutes on the loop.
• Optional: a custom domain, roughly $12 a year. Skip it for the first month. Nobody unsubscribes over a beehiiv subdomain.

The honest shape of 90 days

Ninety days is enough to build a list and switch one revenue mechanic on. It is not enough to replace an income, and any version of this promising otherwise is selling you the plan rather than running it.

The reason issues 1 and 2 come before any growth work: a signup page with an empty archive converts badly, and you will not know your own format until you have written it twice.

Step 1: Launch the publication in beehiiv

Create the publication, and spend the setup time on exactly three things. Everything else in the settings can wait.

The name and the one-line promise. The promise is what makes someone subscribe from a landing page they will read for nine seconds. Make it specific and falsifiable. "AI news" is not a promise. "One AI tool I actually used this week, what it replaced, and what it cost" is.

The signup page. beehiiv gives you a hosted landing page on the free plan. Put the promise at the top, the cadence directly under it, and one email field. Do not ask for a first name. Every additional field costs you signups and you have no use for the name in week one.

The welcome email. One email, sent on signup, doing three jobs: confirm what they get and when, link the best thing you have already published, and ask one question with a reply-to that reaches your inbox. Replies to a welcome email are the highest-signal thing a small newsletter gets, and they teach you the format faster than any analytics screen.

Step 2: Lock the weekly issue format

The reason most newsletters die around week five is not effort. It is that every issue is a blank page. A fixed format turns writing into filling in blanks.

Five blocks, in this order, every week:

"One I'd skip" is the block that does the disproportionate work. Every AI newsletter is enthusiastic about everything, so an honest negative is the rarest thing in the category and it is what makes the positives worth anything. It costs you nothing except the discipline to actually try things you end up not liking.

The subject line is written last, from the "one thing" block, and it is the concrete noun rather than the category. "I replaced my video editor with a 40-line script" travels. "This week in AI" does not.

Step 3: One platform, one link, one CTA

The growth loop is the part people over-engineer. The version that works is embarrassingly small: one short-form platform, one signup link, the same CTA every time.

Pick the platform where you already post, or the one whose format you can stand. Then:

1. The link goes in the bio. One link, to the signup page, not to a link tree with six destinations. Every hop loses people.
2. Every post ends with the same line. The same words each time. Repetition is what makes a CTA get learned rather than read.
3. The content is the issue, cut up. Your "one thing" block is a short-form video. Your "one I'd skip" block is another. You are not producing content for two channels, you are producing once and cutting twice.
4. Tag the link so you can see what worked. Put UTM parameters on the signup URL and beehiiv will show you which source the subscriber came from.

https://yournewsletter.beehiiv.com/subscribe?utm_source=instagram&utm_medium=bio&utm_campaign=launch

Change utm_source per platform if you eventually add a second one. Do not add a second one before day 60.

Why one platform: at a small size, three platforms is not three times the growth, it is one third of the reps on each. Reps are what makes short-form work. Pick the one you can post on three times a week for twelve weeks without negotiating with yourself.

Step 4: The three ways this category earns

Here is what each mechanic is, what it requires, and the public rate behind it. They switch on in roughly this order.

Mechanic 1: beehiiv's own ad network and paid recommendations

Two separate products, both built into the platform.

The ad network places sponsors into your newsletter for you. You do not sell, negotiate, or invoice. beehiiv matches an advertiser to your audience, the placement runs in your issue, and you are paid on the performance of that placement. The catch is a qualifying bar: beehiiv has historically required a subscriber floor in the low thousands of active subscribers before a publication is eligible. Check the current requirement in your own dashboard, because it is the kind of number that moves.

Paid recommendations, which beehiiv ran under the name Boosts until it renamed the feature, is the two-sided one, and it is the more interesting mechanic at small scale. As a recipient, you pay other newsletters a set amount per verified subscriber they send you, which turns money into list growth at a price you set. As a Boost provider, you recommend other newsletters to your readers and are paid per verified subscriber you send them. Provider-side paid recommendations, the feature beehiiv used to call Boosts, are the earliest thing on this list that can pay a small list, because you are being paid on referrals rather than on your own size. Which plan tiers can use which side of paid recommendations has changed before, and beehiiv has renamed the feature at least once,, so confirm it on your plan before you build a month around it.

Mechanic 2: a paid tier

The standard shape in this category is a free weekly issue plus a paid tier at $5 to $10 a month, or $50 to $100 a year. That band is not a secret; it is what you see when you look at the pricing pages of newsletters that run one.

The arithmetic is the useful part, and it is unforgiving in a way worth seeing before you start:

Conversion from free to paid in this category is typically low single-digit percentages, so 100 paying readers is usually a free list of a few thousand. That is a year-two number for most people, not a day-90 number.

Two constraints. First, paid subscriptions on beehiiv require a paid beehiiv plan, so this mechanic is the one that stops the stack being $0. Do not turn it on until the free list justifies the plan cost. Second, a paid tier obliges you to produce something extra every single week, forever, for people who have paid. That is a real commitment and it is the reason to make this the third switch rather than the first.

Mechanic 3: affiliate links to tools you already use

The one that works from subscriber one, because it pays on a purchase and not on a size.

You are already naming tools in every issue. Some of those tools run affiliate programmes. Apply, use the tracked link when you name the tool, and disclose it. The mechanics that matter:

• Only link tools you actually use. In a newsletter whose entire premise is honest first-hand reporting, an affiliate link to something you have not run is the fastest way to be worth nothing.
• Disclose it in the same breath as the link. Not in a footer, not in a separate page. One plain line directly under or beside the link saying you earn a commission. This is a legal requirement in several jurisdictions and it is also just the thing that keeps a reader trusting the "One I'd skip" block.
• Keep the "one I'd skip" block free of affiliate links, always. The moment your negative reviews and your commissions interact, the newsletter's one differentiator is gone.

Direct sponsorships, when they come

Not one of the three switches, because it is a conversation rather than a mechanic, but it is where the numbers get bigger and you should know the band. Newsletter sponsorships are quoted as a CPM, and the unit matters: rate cards price per thousand opens, not per thousand sends, so a list with a weak open rate is worth less than its subscriber count suggests. beehiiv works its own example at a $25 CPM on a 10,000-subscriber list, which is $250 for one placement. Narrow, high-intent tech and business niches sit above that.

Run the arithmetic at your real size, not an imagined one:

To get your own row, multiply subscribers by your open rate. A 2,000 subscriber list opening at 40 percent is 800 opens, so it is the first row, not the second.

This is the number people do not run before they start, and it is why the first year of a newsletter is a list-building exercise rather than a revenue one. It is also why the niche matters more than the size: a list of 2,000 people who buy AI tooling for a company is worth more to a sponsor than 20,000 general subscribers, and CPMs in that shape go well past the band above.

Step 5: The 90-day plan, week by week

WEEK 1     Publication live. Signup page. Welcome email. Issue 1 sent.
WEEK 2     Issue 2. Link in one bio. First 3 short-form posts.
WEEK 3-4   Weekly issue. 3 posts a week. Apply to 2 affiliate programmes
           for tools you already use.
WEEK 5-8   Weekly issue. 3 posts a week. Add the reply question to every
           issue and read every reply. Adjust the format from the replies,
           not from the analytics.
WEEK 9     Check the beehiiv dashboard for ad network eligibility. If the
           list is not there yet, note the gap and keep going.
WEEK 10    Turn on paid recommendations as a provider. Recommend only newsletters you
           would read.
WEEK 11-12 Weekly issue. 3 posts a week. Write the sponsor one-pager:
           subscriber count, open rate, who reads it, what a placement
           costs at your real CPM.
DAY 90     One mechanic live. A real open rate. Twelve issues in the
           archive that prove you ship.

The instruction that matters most in that plan is the one in weeks 5 to 8: adjust from the replies, not the analytics. At a few hundred subscribers your open rate is noise and one honest reply telling you which block they read first is worth more than a month of dashboards.

What kills these in the first 90 days

Four failure modes, in the order they happen.

Launching on a topic you do not do. A newsletter about AI tools written by someone who does not use AI tools daily runs out of material in five weeks, because there is nothing behind it but other people's announcements. The format in step 2 only works if the "one thing" block is a real thing you did.

Adding a second platform in week three. The loop needs reps on one surface. Three platforms at a third of the reps is worse than one platform at full reps, and it feels like progress the whole time it is failing.

Optimising the send time before you have a format. At 300 subscribers there is no send-time signal in your data. There is a format signal in your replies.

Turning on monetisation before there is anything to monetise. A sponsor slot in an issue that goes to 200 people is worth about six dollars and costs you the one thing that made the issue worth reading. The order in this guide is the order for a reason.