Assuming a sufficiently high marginal tax rate, what do you mean?
By mainstream Keynesianism I am referring to the neoclassical synthesis that dominates mainstream macro today. The best-known advocate for this school of thinking is undoubtedly Paul Krugman.
Basically this is the "synthesis" of Keynes' theory of aggregate or effective demand with the classical efficient-market models of the economy. Thus, the mainstream synthesis says that "schocks" can temporarily push the economy out of equilibrium, but the natural tendency of the market toward equilibrium remains generally unquestioned.
The problem with this of course is that Keynes showed that the economy could easily reach equilibrium in a state of mass unemployment. Post-Keynesians go further and say that the economy isn't a system that tends toward equilibrium at all; drawing on behavioral psychology and other extra-economic fields of study they argue that the economy can and frequently does go completely off the rails*! This is one predictive area where the post-Keynesians have the neoclassicals beaten hands-down: left to itself the market does not demonstrate any tendency toward equilibrium, quite the opposite. This leads neoclassicals to systemically underestimate the extent and duration of financial crises.
The worst tendencies of the neoclassicals are embodied in the "real business cycle" theory which completely ignores the role of finance and claims that the business cycle is rational utility-maximizing responses to real changes in the economy.
Anyway: traditional Keynesian pump-priming is trickle-down in the sense that it's giving money to 'investors' and hoping they will invest it. Minksy said that it would inevitably be inflationary without price controls, and he favored a program of public employment instead. The way to avoid trickle-down in public policy is by not giving any money to for-profit companies: give the money to the people who might buy things from those companies instead. Let consumer spending drive investment rather than the other way round.
I suppose in fairness I should mention that mainstream Keynesians do tend to support countercyclical programs like unemployment insurance that do exactly what I'm talking about here. But the "stimulus" as passed under Obama was mostly trickle-down which seriously diluted its effectiveness. If it hadn't been accompanied by the automatic stabilizers it would not have come close to "succeeding".
*they actually claim something closer to "there are no rails except those imposed by policy"