For content creators, this site is a gold mine. They'd stand to lose a lot of money if it went away. Are they actually going to stand by and fluoride stare at the KiwiFlare error page without offering to help? It wouldn't surprise me, but it seems like a shitty thing to do.
You're right that a lot of people eat off this site. But the thing stopping us isn't the size of the cheque, and it's worth walking through slowly, because nearly everyone assumes this is a fundraising problem and it isn't.
Start with the line you quoted. Null offered
$2,000 a month for a single gigabit of clean traffic. For reference, a gigabit of ordinary transit in a major city runs somewhere around $100–500 a month. He offered roughly four to twenty times the going rate and still got nowhere. When a man offers to pay several times market price and
still can't buy the thing, price was never the obstacle.
First, why creators can't simply fix this
Most creators' income lives on YouTube, Twitch, Patreon, PayPal and similar. Every one of those can be made to drop a person by the same pressure campaign that's aimed at us. Asking a creator to publicly bankroll this site is asking them to trade their entire livelihood for our uptime. Some quietly help already. Asking them to do it
loudly is asking them to be destroyed, and then there's no creator and no money either.
There's also a problem with taking their money even if offered: we write about these people. A site funded by the subjects of its own threads is a site nobody should trust. The first time a donor asks for a thread to be trimmed, the site is worth nothing. That independence isn't negotiable.
But set both aside, because even a mountain of cash doesn't buy what nobody will sell.
What a DDoS actually is, in plain terms
Imagine you run a shop. One morning ten thousand people show up at the door at the same second. They don't want to buy anything. They just stand in the doorway. Real customers can't get in.
Here's the part everyone misses:
to make those ten thousand people go away, they have to arrive first. You cannot stop them from a distance. You need a doorway wide enough for all ten thousand to walk up to, and enough bouncers to look at each one and say "you're fake, out." You are paying for a door big enough to fit the flood you are throwing away.
So when people say "just filter it," what they're really saying is "buy a pipe big enough to receive the entire attack, then discard it." That pipe is the expensive part.
What building that ourselves actually costs
Rough, real-world ballpark for one filtering site:
Code:
ONE scrubbing site
Rack space + power (2 racks, ~10 kW) $2,500 - 4,000 /month
Cross-connects to carriers (4-6) $1,200 - 3,000 /month
IP transit, 100 Gbps commit $10,000 - 30,000 /month
Remote hands $500 - 2,000 /month
--------------------------------------------------------------
RECURRING $15,000 - 40,000 /month
Border routers that filter at 100G+ $80,000 - 250,000 one-time
Filtering servers (~10, 100G NICs) $60,000 - 120,000 one-time
Optics, cabling, spares $20,000 - 50,000 one-time
--------------------------------------------------------------
UP FRONT $160,000 - 420,000 one-time
And one site is not enough. One site is a single point of failure — if the attack is bigger than that building's pipe, you're down, full stop. You also want the attack to land near where it originates, so traffic from Asia dies in Asia instead of crossing an ocean into your only door. That means three to five sites minimum:
Code:
THREE to FIVE sites
RECURRING $45,000 - 200,000 /month
UP FRONT $500,000 - 2,000,000 one-time
PLUS, once
Our own IPv4 /24 (address space) $10,000 - 16,000 one-time
Our own ASN + registry membership $500 - 2,000 /year
PLUS, people
24/7 on-call network engineers $300,000 - 600,000 /year
Call it
$70,000–250,000 a month, plus roughly half a million to two million dollars before we serve a single page. And that buys us a few hundred gigabits of headroom — against attacks already measured in the hundreds of gigabits and climbing. The attacks we're taking right now are historic in scale; this isn't the usual background noise.
The reason Cloudflare and the like can absorb this is that they have hundreds of terabits and split the bill across millions of customers. We would be one customer paying for all of it, alone.
Now the part money doesn't solve
To filter traffic, somebody has to hand it to us first. The internet isn't one thing — it's thousands of separate companies' networks glued together, and a small handful of very large ones carry everybody else's traffic. Think of them as the motorway operators. If you want the world to reach your building, one of those operators has to agree to build an off-ramp to it. That's called transit.
We don't need them to like us. We need them to
sell us an off-ramp. That is the entire problem.
Because here's what happens. Imagine you rent a shop. You pay every month, on time, in full. You never break a single rule in the lease. But every day a hundred people phone your landlord and say "evict him or we'll make your life hell, and we'll tell your other tenants too." Eventually the landlord reads his own lease, finds the clause letting him end it with 30 days' notice for any reason at all, and uses it. You did nothing wrong. You're still on the street.
That is what an AUP bombardment is, and it happens to us at every layer — the datacenter, the transit provider, their upstream, the payment processor. Nobody has to prove we broke anything. "Acceptable use" clauses let them leave because being associated with us became inconvenient.
Why losing either partner is a catastrophe
- Lose the datacenter and it isn't a config change — it's physically unracking hardware, shipping it, finding a facility that will take us, signing again, re-cabling, and rebuilding carrier relationships in the new building from scratch. Weeks to months of degraded or no service, and at the end of it we have exactly the same 30-day guarantee we had before.
- Lose the transit provider and we need another carrier that is already physically present in that same building. If nobody there will take us, the hardware is stranded — hundreds of thousands of dollars of routers sitting in a room with no road to them. Getting a new carrier to run and light fibre into a facility is months of lead time and tens of thousands in one-off charges, assuming they say yes at all.
So here's the real question
Would you spend $500,000 on routers for a building you might be evicted from in 30 days? Would you sign a three-year rack lease when the carrier feeding it can walk with a month's notice, for no stated reason, and there's no court to appeal to?
That's the actual decision in front of us. Every piece of this requires committing large sums for long periods to partners who can each individually leave at short notice, for reasons that have nothing to do with whether we paid or whether we followed the contract. You cannot build permanent infrastructure on foundations that any one of four separate companies can pull out from under you on a whim.
That's why the ask you quoted is so small and so specific. It isn't "give us a million dollars." It's "let us pay you well above market rate for one gigabit, and don't leave in ninety days." We have the money for that. What we can't manufacture is somebody willing to sign.
So it isn't that creators are standing by doing nothing. It's that this isn't a problem their wallets can reach.
