Gate on a Real Constraint

Scarcity reads as manipulation unless the limit is actually real.

Every invite-only launch makes the same bet: restrict the door, let the waitlist do the talking, and hope nobody asks why the door is exactly this wide. Most founders reach for the tactic because it worked for somebody else, without doing the one thing that made it work for that somebody else, tying the limit to a number they could actually defend. Manus AI ran this experiment in public in March 2025, and the backlash it took over the following weeks is the whole lesson in miniature. If you cannot name the resource you are rationing, do not gate on it.

What to do: Find the resource that is actually capping how many users you can serve well right now, compute, support headcount, onboarding calls, manual review capacity, and gate access to that number, not a round one that sounds exclusive. State the number and the reason on the same page where people request access, so the claim is checkable the day you make it.

Why it works: A constraint you can name and defend reads as an operations decision, while one you can't reads as a marketing trick, and the first time someone catches you on the second, every invite you send afterward inherits the suspicion.

Example: Manus AI's invite-only launch in March 2025 sent its access codes into a resale market on eBay and Xianyu, with real listings running from $1,000 up to $13,797 for a single code. Product partner Zhang Tao told the public the gate was "due to genuinely limited server capacity at this stage," not a growth play, because the team had built infrastructure for a demo and got launch-day demand instead.

Walk it through

I could not pull a clean screenshot out of this one. Manus's March 2025 signup flow is gone from the live site, and the Wayback Machine's archived copy just renders a client-side error, a React app that never rehydrates outside its own domain. So here is the sequence as it actually ran, from contemporaneous reporting, plus the part you can still check today.

1. March 6, 2025: the gate goes up. Manus, built by the Chinese startup Butterfly Effect, launches invite-only. No paid media behind it, according to the company. Just a locked signup and a code you had to already know somebody to get.

2. Within days: a resale market appears. Codes start moving on Xianyu and eBay. Prices run from $1,000 to $13,797 for one code at the high end. That price is not curiosity, it is a market pricing in real scarcity, whatever is actually causing it.

3. Zhang Tao names the constraint, in public. Manus's product partner posts that the invite system is "due to genuinely limited server capacity at this stage," and admits the team built for demo-level traffic, not the demand that showed up. That is a specific, checkable claim: servers, not hype.

4. The public does not buy it, also in public. Coverage from the South China Morning Post and others reports open accusations that Manus was running deliberate "scarcity marketing tactics," with the same news cycle carrying both the claim and the counter-claim. That is the actual risk of gating without receipts, you do not get to argue your case privately.

5. The resolution is the real proof. By 2026, Manus dropped the invite requirement. Registration is open to anyone, and the old codes now just add bonus credits instead of blocking the door. A real capacity constraint gets fixed and the gate comes down. A manufactured one usually needs a new excuse instead.

Now run the same test on your own gate before you ship it. Write down the actual number, the resource it maps to, and the metric or date that will let it open wider. If you can't fill in all three, you are gating on vibes, not on a constraint.

The read

Name the resource or don't gate at all. "Server capacity" survives scrutiny because it's checkable, compute has a real cost and a real ceiling. "Limited spots" with no unit behind it is a marketing sentence wearing an ops costume.
The gate needs an exit. LinkedIn opened past Reid Hoffman's personal contact list once density existed elsewhere. Gmail opened past invite-only in 2007 once storage got cheap. Manus opened past invite codes once server capacity caught up. Real constraints resolve. Manufactured ones rotate to a new story instead of closing out.
A resale market is a demand signal, not a legitimacy signal. Codes changing hands for $13,797 proves people wanted in badly. It proves nothing about whether the reason you gave them was true, and your reputation only rides on the second fact.

Steal it

Before you gate anything, find your actual ceiling. If it's support, count how many onboarding calls your team can run well in a week and cap signups at that number, then publish it. If it's compute, say which budget or model tier you're capped on and let people do the arithmetic themselves. If it's review, say how many applications a human actually reads in a day. The gate should feel like an ops constraint because it is one, not a growth trick wearing one as a costume.

Defend it by staying ready for the question everyone eventually asks: which server, whose calendar, what number. Manus took that question in public and had an answer, a small team that built for a demo and got hit with real demand. If your honest answer is "we picked a number that felt exclusive," don't ship the gate. Fix the actual constraint until it's real enough to name, or don't gate at all.

Gotchas

A named number can still get disbelieved. Zhang Tao named servers specifically and still got accused of manufacturing hype, because stating a claim precisely does not make it automatically believed. Be ready to show your work, not just state it once.
A constraint that never resolves starts to smell like the fake kind. If your gate sits at the same number a year later with no infrastructure update, no headcount change, nothing, people will conclude you never had a real one.
If the constraint is real, fix it fast. A gate that protects genuinely scarce capacity is still turning away real demand every day it stays shut, and the growth story only looks good in hindsight. The company that both defends the scarcity and races to open the door wins twice.