Exactly. Say you go to a fusion laboratory that employs the top ten fusion scientists in the world, with a current budget of a few hundred million dollars, and you say to them "I've got 2 billion here in additional funding for you. You need to spend it in the next few months because I'm trying to stimulate an economy right here."
Does that hire you forty more world class fusion scientists? Are there forty more world class fusion scientists?
Are those scientists going to go "Fantastic! We'll just order up three more of these experimental fusion reactor rigs, here are the plans, we'll start hiring engineers and buying up electronic components from the local silicon fab down the road driving all that money into economic productivity for you!" - unlikely. They already have one experimental fusion reactor. Three more the same aren't going to get them anywhere any faster. They won't know what the next one should look like until they've finished getting results out of this one.
Maybe they just spend it all on repainting the lab, buying new office furniture, getting in catered lunches - you might see a marginal improvement in fusion research productivity, and certainly there's a local economic stimulus which was the idea there, but you're not getting the valuable long term capital value of, say, using the money to rebuild a few bridges.
Even if it turns out your ten fusion scientists have a ready proposal to build the next generation experimental fusion reactor, plans are drawn up and costed, and all they need is you to hand over that check for two billion, if they then go ahead and order the parts from China, subcontract a bunch of German engineering firms to build the cooling systems, and get the heavy steelwork welded together in Korea, your stimulus effect just disappeared overseas.
Of course, if that experimental fusion reactor turns out to be THE ONE that creates viable economic fusion power, maybe that's worth the 2 billion. But that's not stimulus spending, that's gambling on an investment.