USENETPRIME

Not published, which is its own kind of answer. Our US East bench settled at 795 to 818 Mb/s on 50 SSL connections during the 28 July to 2 August window: steady, and a clear step below the UsenetExpress-direct and ThunderNews traces despite riding the same spool. That gap is routing and connection allocation rather than hardware. The 20-connection bonus server is slower again; treat it as a completion tool, not a speed tool.

Advertised at 5,700+ days on the primary, with the bonus server quoted around 5,781 to 5,833. Independent aggregators put the real figure closer to 4,000+. Retention is how far back the spool still holds articles, and when the provider and the outside world disagree by eighteen months the honest publication is both numbers rather than the flattering one. What is not in dispute: the two servers age out independently, so an article missing from one is genuinely worth retrying on the other.

This is the interesting axis. The primary feed is UsenetExpress, the same network as ThunderNews, TheCubeNet and NewsgroupDirect’s core, so no redundancy there. But the XS News bonus block runs on Abavia, which is an entirely separate ingestion and storage chain. Two unrelated backbones behind one login is what most people are buying a second provider to achieve. The score is held down by the primary’s reseller status, not by the bonus, which is the genuinely good idea here.

Claimed at 99%+, rising to 99.98% when the bonus path is counted. Treat that second figure as arithmetic rather than measurement. It is what you get when two independent spools each miss about one article in a hundred and rarely miss the same one. The logic is sound and it is the right reason to like this account; it is still a calculation, not a result. In our checks the bonus server did pick up articles the primary returned empty, which is the behaviour the claim predicts.

US and EU access on the primary plus a separate bonus host, on SSL 443, 563 and 564. Fifty connections on the main feed, twenty on the bonus. These are access endpoints into other companies’ farms rather than hardware UsenetPrime owns, which is fine, provided nobody implies otherwise. Nobody does.

US and EU entry points, with the bonus server adding a Dutch route. No APAC, no documented IPv6. Plaintext ports 23, 80 and 119 are open for reachability on hostile networks. Available is not the same as recommended, and there is no reason to use them when 563 works.

The weakest documentation on this board, and the reason a provider with otherwise strong numbers sits at slot six. SSL is present and works. What is missing is everything around it: ownership is not publicly disclosed, jurisdiction is not stated, and there is no meaningful logging or retention-of-records policy to read. That is not evidence of bad practice. It is an absence of evidence either way, which is a different problem and not one you can resolve at checkout. If your threat model is your ISP, the SSL is doing its job. If it is anything more than that, buy from someone who will tell you who they are.

None, and none advertised. Given how thin the rest of the disclosure is, that is arguably the correct call. A bundled VPN from a company that will not name itself is not a privacy feature. Your NNTP session is encrypted regardless; a VPN would hide that the session exists, and you can buy that separately from someone accountable.

Visa, Mastercard, Amex, Discover and PayPal. No crypto documented, which is a notable gap for a provider whose main appeal is a bonus backbone rather than a brand. Advertised August 2026: $10/mo unlimited, $75/yr at an effective $6.25/mo, or the Preferred annual at $50/yr with a 750 GB XS News bonus block, which is the plan the whole review is really about. Blocks are non-expiring; a 30-day refund window is cited. Checkout pages were partly obscured during research, so confirm the recurring rate rather than the promo rate before committing.