Top Research Companies

A maker is ranked here on its research robots only. Not on its global rating, and not on its size. A company that is strong in one category and weak in another ranks differently on each board — that is the point of having one board per category.

Rank = best robot, plus credit for repeating it, plus news. The evidence part is the maker's best eligible research robot plus up to 5 points once it fields 8 of them. The credit is capped, so a large catalogue cannot buy a rank — catalogue size partly measures how much of a maker we have collected, which is our business, not theirs. Every part is printed above: the best robot, the depth credit, the news, and the total they sum to.

Deployments and funding are not in this score. Seven in ten companies have no deployment record on file and funding is blank for exactly the makers a buyer trusts most, so ranking on them would rank our own reading rather than the market. Nothing here can be bought, claimed, or subscribed to.

A maker's total here is not its RoboScore. It is its best eligible robot on this board, plus capped credit for depth, plus news. The company's own RoboScore and RoboIndex — across every category — are on its page. News can raise a rank by up to 15 points and can never lower one, so a maker with live events is ordered above one on the same evidence. Both parts are printed on the row, so a company sitting above a higher score always says why. A quiet maker keeps its full score: we have not measured silence, only failed to hear anything, and the two are not the same.

How we rank — the evidence behind every robot score these standings are built from.