Short-term versus long-term gains. Businesses have the short term covered, but they're rarely going in on high-risk/low-reward (monetarily) research.

So supply and demand can get stuck at these local maxima. Central planning can think longer term. Maybe it doesn't pay off, but sometimes it does. And when it does, it can shake us free of that maximum.

This works, and is why Bell Labs and its peers existed. For reasons I don't fully understand, they stopped doing this, largely. Maybe because it's difficult to quantify those long-term gains.

Finally, those scientists get to choose where they are employed. So if the private sector wants them to do this work, they can hire them.

Using an inaccurate phrase like “central planning” calls your motives into question. That’s designed to conjure up scary images of Cold War-era Russia or China, but it’s really not how things work in a democracy where democratically-elected representatives allocate money based on their constituents’ interests, and hire experts in their fields to disburse those funds to private researchers based on the recommendations of other private experts, and the resulting research is often greatly profitable for private companies competing in the market (nobody bought Air Force microchips or had to license CRISPR from the president’s nephew).

One thing all of the major advances which shaped the 20th century have in common is that they required huge upfront spending which even large companies couldn’t afford. Smart people will contribute to society in various ways but creating those opportunities allows that to happen more frequently and on a larger scale.