Goldman Sachs invests $250M in compressed air energy storage

Interesting question, but seems unlikely.

If you're thinking of a direct mechanical linkage, you wouldn't want to put it up on the tower because the mass to effectively store megawatt-scale power is way more than you'd want to support up high. Also, since the wind blade usually swivels to the wind, you'd want to avoid the rotational inertia. So, you'd need a mechanical linkage to transmit the power to the ground, maybe installing it under the tower base.

You would not want to use a direct mechanical attachment, i.e., resembling a combustion engine - transmission flywheel, because this would impair the ability of the wind turbine to start. Generally, turbines are designed for minimum inertia to easily start in low-wind conditions. A direct connection would impede that.

Now, we're adding a clutch-sort of mechanism, and that has its own additional complexity, weight, and energy.

Moreover, considering that we're trying to store megawatt-scale energies, we are at a large mass spinning very fast, and probably spun up with an electric motor.

So, it would seem the best way to do that would be to make a flywheel farm, with the flywheels below ground to contain failures. At this point, why locate it in the probably inconvenient location where the wind turbines are located, and instead put it somewhere more convenient, such as nearer to the consumption areas?

Just cheeked some levelized utility scale costs for the US:

Advanced Nuclear is of $72/MWh.

Wind onshore $30/Mwh

Solar photovoltaic (PV) $30/Mwh

Lithium-ion battery storage is roughly $180/Mwh.

At the moment, if you want to provide baseload better than nuclear using renewables, the cost must of storage must not exceed $30-$40/Mwh.

(grid baseload is the minimum level of demand on an electrical grid over a span of time, something unvarying power plants are best suited for. Renewables require storage and advanced grid to provide baseload.)

You are spot on - the real answer is readily available: nuclear. We have reactor designs that if the active systems fail the reactors shut down naturally instead of running away. We have designs that can be burning all that "spent" fuel we are currently trying to bury for 10,000 years (!). Modular reactor designs in particular have incredible promise to provide for infinite scaling (just add more modules as needs increase), deal with the inefficient building, maintenance and decommissioning costs of todays light water reactor designs, and being modular provide far more granularity in power delivery further negating the need for energy storage to smooth out spikes and troughs in demands; allowing energy producers to far better match generation to power requirements.

Sadly I think it will be far more feasible to develop energy storage technologies than hope we can ever get enough people to be rational about nuclear. I mean how asinine is it for Germany to have turned off perfectly working and economically viable nuclear plants to then be dependent on RUSSIA for natural gas? How dumb do you have to be to be to think that's a REMOTELY good idea? Yet here we are (thankfully France seems to be far more rational - you don't hear them even suggesting they plan to be equally idiotic).

Many battery storage studies don't account for replacement/maintenance, or the infrastructure that would be required to sustain replacement/maintenance at grid scale, let alone the resources required to scale up to handle initial deployment of batteries in sufficient quantities to support the entire grid. Where are we going to get the manufacturing capacity, let alone materials? The scale here is pretty mind boggling if you start to do the math.

Our current battery tech is nothing more than a transitory technology - they are far from sustainable long term. Personally I think super capacitors will be our ultimate solution, but in the meantime if these guys really have dealt with the seepage issues of compressed air, this is are next best bet for MASS energy storage at scale. You only have the privilege of worrying about efficiency if you have a working system. Wood driven steam sucked, but it drove the industrial revolution because there was literally no other viable alternative at the time that could have remotely scaled in the same way.

Remember grid scale solutions require MASSIVE energy capabilities. The scale here is ridiculous compared to home or vehicle energy requirements. If you don't think Goldman Sachs took all of this into account before they made their investment - well, you are grossly mistaken. You don't get to be their size by continually placing bad bets.

The heat still has to be managed, otherwise Kyshtym[1] happens. 160 tons of flying concrete is no joke.

https://en.m.wikipedia.org/wiki/Kyshtym_disaster

What would be even cooler is zeppelins with swimming pools and jacuzzis while you take in the views of the sky.