Get institutional grade news at retail prices.

Every wire, scored the instant it breaks.

Most desks are reading news that has been sitting in someone else's inbox for three minutes. This is the other side of that, and it is not a website: a native application on your machine, all four wires in one stream, a direction and a magnitude on each line before you have finished the headline.

in development

168 ms

--:--:-- ET

TIME · ETAITICKERHEADLINE

Summary · public reporting

An 8.5% stake makes Markiplier GoPro’s largest individual holder

A Schedule 13G dated 20 August put Mark Fischbach — the YouTuber who has gone by Markiplier for the better part of two decades — on the register with 8.5% of GoPro, roughly $9.3 million at the time of filing, and made him the largest individual holder of a company whose cameras he has been using for years.

He is not a fund and he is not an activist. By his own account he looked at where the stock was trading and decided the market had the business wrong, then spent months quietly building the position and learning the disclosure rules that came with it.

The company he bought into is in trouble. GoPro is down sharply over the past year and worse over five, said in May that it was weighing strategic alternatives, and has since put going-concern language in its own filings. None of that deterred him; he kept adding.

Fischbach frames the investment as part of a wider interest in lowering the cost of making films. He has also backed Strada, which builds tooling for moving large footage between collaborators.

He has some standing to make the argument. His self-distributed feature Iron Lung took roughly $51 million against a $3 million budget this year, released without a studio behind it.

GoPro cameras shot the behind-the-scenes footage on that film, though he did not think they were good enough for the film itself. A teaser for a new interchangeable-lens body is what changed his mind about the hardware — and, it seems, about the company.

The market took the filing at face value. GPRO closed 31 August up 46% on 220 million shares, and kept going after the bell.

GPRONASDAQ · 31 Aug 2026 0.87+0.27  +45.80%

0.60 0.87

O 0.61H 0.87L 0.60VOL 220.0M

GPRO headline and tape: real Nasdaq trading, 31 August 2026, replayed from five-minute closes. The story beside it is our own summary of publicly reported facts, not a reproduction of any outlet’s article. The other releases in the feed are simulated and their issuers fictional. Every score shown is illustrative of the model’s output, not a record of it.

Why press releases move markets.

Most traders never realise the news they are trading on has been sitting in someone else's inbox for three minutes. By the time Yahoo, Google or your broker shows it, the move is already over. Here is the same headline reaching four audiences.

Desks that do this properly pay five figures a month for wire-speed news. pr.live puts the same four wires, at the same instant, in front of a retail trader for $99.

A release crosses one wire, and you don’t pick which.

The company chooses. Whatever it pays for is the only place the release appears at the moment it breaks — whole, but in that wire’s own format, with its own metadata and its own way of naming the issuer. Watch three of the four and you simply miss whatever crossed the fourth. pr.live takes all four, normalises them into one stream, and scores each release as it lands, while the first sentence is still being read elsewhere.

Time is not the same everywhere.

Close enough to something heavy, an hour costs seven years, and nobody living through it feels late. Distance from the wire works the same way: everyone reading three minutes behind believes they are current.

The feed is the floor.

The whole product is in beta, so here is the honest split: what the model and the feed already do, what the API already lets you do yourself, and what we simply have not built.

You see it first. The call is still yours.

A release reaches you about 170 ms after it crosses, with a direction and a magnitude already attached. That is the whole product: the score is an input to your decision, not a replacement for it, and you can always pass. Five releases cross below and you get five seconds on each. Beside your result sit the three you are actually competing with — the desk that saw it three minutes late, the one that was every bit as fast and read it wrong, and an algorithm running on the same feed you have — quick, often right, and never able to sit one out.