[plutokras]

They have fixed costs per query. Unlike VC-backed companies burning cash for market share, they need sustainable unit economics to stay profitable. This makes regional pricing hard to pull off.

I know I'm speaking form a position of privilege and this will be unpopular; but I'm not fond of subsidizing other users by paying premium prices for my subscriptions.

[rurp]

That's largely under Kagi's control though. They don't seem to be working hard to reduce their expensive dependencies, instead choosing to focus on a number of other features/products. Maybe that's the better business strategy overall but it does limit their appeal to many potential customers.

[tough]

I hadn't thought about that

TBH, if i where them i'd be trying to serve open source models from my own infra, much cheaper to pay per GPU's per hour and batch process all your users prompts, than leave that big 95% fat margins to OpenAI and Anthropic

But I guess they have customers who want those APIs anyways, idk, again, i thought they where a search service, not an ai company, so this sub for llms business deal is weird from that POV? like great that it works for them to get money/customers but that doesnt seem their main point of existing?

[mkayokay]

Also with regional pricing one must ensure the lock out all those people who want the service cheaper than in their country and try using VPNs or other means. Otherwise you loose even more money.

[mongol]

Does not neccessarily have to be done that way. It could also work like you pay for Kagi in country X, and get it localized for that country, using that language and prioritizing sites in that country and so on

[chilldsgn]

I agree with you. I understand it's incredibly challenging to implement regional pricing, especially if your company is based in a 1st world country, and dependencies are also priced for 1st world countries.