It's helpful to disaggregate the parties involved a bit--not by nation or colony, but by sector. When the statement "colonialism was a net drain on European powers" is made, it sometimes has an implicit agenda, but it also tends to focus on state revenues. If you look entirely at metropolitan state revenues, few colonies were very profitable, and those that were did not generally remain so.

If you expand it outwards to include private concessionaires, trading companies, shipping lines, and so forth, the picture changes and they look far, far more profitable. The state (colonial or imperial) did not control company outlays or reap their revenues, but only set up the spaces and policies under which they operated. For example, it was evident that British interest in expansion and consolidation in southern Africa was driven by the discovery of minerals, even though almost none of that revenue actually went to the Crown--it went to private companies like De Beers (and Consolidated Gold Fields, also Cecil Rhodes's baby); for them, empire was tremendously, even scandalously profitable. The Congo Free State / Belgian Congo is another example. Leopold II wasn't ruling for Belgium, however he might represent himself as being the state; it was the fiefdom of a private company, his own, and he became quite rich through ruthless, brutal behavior in a temporary market situation (high demand for natural rubber) but made no real investment on the ground and put very little of the money into Belgium's coffers--he "merely" built things. When the Belgians took over, they continued to operate some of the profitable industries, but again through concessionaries like the Union Muniere (imagine the diacriticals there) which were not the same as the revenue of the colony itself. Some of that money might have come back to the Empire or to the individual colony/dominion, as it did with De Beers or the Randlords in South Africa, but usually that didn't count as direct revenue from the colony itself--they were duties, fees, or taxes paid by private entities in some other context. [edit: Yes, these entities often paid for their concessions--but they paid very little, and that wasn't technically colonial revenue unless paid beyond the colony itself.]

The problem with all of this is that it becomes impossible to know exactly how much wealth empire did generate, for whom, and where it went. We can get some figures, but other entities are still quite private, or didn't report. Leopold II, for example, famously destroyed a tremendous amount of the paperwork on the Congo Free State. The various East India Companies also made a tremendous amount of money for shareholders, who took their dividends while the companies themselves eventually sank when they couldn't balance their books anymore--leaving the home government to hold the proverbial bag. One could argue that India was always profitable to the British government during the era of the Raj until WWII (so 1858-1942), but a fair bit of money did go back into India itself to support further colonial infrastructure as was the case with most other colonial regions. Few colonial powers expected their possessions to pay in a direct way, which explains why initial administration of overseas territories so often devolved on concessionary companies that were willing to front the capital and see how it might turn out. From working in the British archives, I can tell you that the philosophy of empire on the cheap was alive and well in 19th-century southern Africa, in part because state revenue was so weak. At the same time, one of the world's richest human beings--Rhodes--was busy building his own little Company empire north of the Limpopo with a private army, private diplomats, and the works, all with the blessing of the Crown but little direct engagement. This arrangement wasn't strictly mercantilist by the late 1800s, but it did involve certain limited monopoly concessions, much as still exist for various companies within many independent postcolonial states today. Then, as now, the position of government treasuries tends to obscure the broader picture.

edit: Robert Rotberg's The Founder is nice if you want to see how Rhodes himself operated in this web of government and private capital--he held political office in one hand (Prime Minister at the Cape, 1890-1896) while running his private empire in the other. We're still looking for the absolute smoking gun, but his labor policies and trade policies (not to mention his attempt to topple the semi-independent Boer government in Pretoria at the time) show this interplay pretty effectively. The Cape pleaded poverty all the time, but was run by an inordinately wealthy man. Hochschild's King Leopold's Ghost may still be the most accessible for the Congo Free State. The era of the companies is in Bown's Merchant Kings: When Companies Ruled the World 1600-1900, which is very broad but it may be a helpful starting point. Huw Bowen's The Business of Empire is much better for a closer look at the EEIC/BEIC.