Modern reverse-osmosis (RO) desalination is an energy hog: seawater must be squeezed against a semipermeable membrane at roughly 60–80 bar, and the high-pressure pumps eat 3–4 kWh per cubic meter of fresh water produced. The pressure is the fundamental cost — but pressure is just gravity times a column of liquid. What if the "pump" were simply height?

The physics of the osmotic column

Seawater has an osmotic pressure of about π ≈ 27 bar at 35 g/kg salinity. To force water backward through an RO membrane you need hydrostatic pressure exceeding π. A column of seawater generates pressure P = ρgh. Solving:

h_min = 27 × 10⁵ Pa / (1025 kg/m³ × 9.81 m/s²) ≈ 269 m

A 269-meter seawater column just barely matches osmotic pressure. To drive useful flow, we overshoot. Consider a 1000-meter-tall tower — think Burj Khalifa (828 m) plus a bit — filled with seawater:

P_bottom = 1025 × 9.81 × 1000 ≈ 100 bar
Net driving pressure = 100 – 27 = 73 bar

That's above typical industrial RO operating pressure. Install a bundle of spiral-wound membranes in the basement. Fresh water bleeds through into a return riser, buoyantly climbing because it's 2.5% less dense than the surrounding brine. How high?

ρ_fresh · g · h_fresh = 73 × 10⁵ Pa
h_fresh = 73 × 10⁵ / (1000 × 9.81) ≈ 744 m

So fresh water spontaneously rises 744 meters up a companion pipe — delivered to a city reservoir at cliff-top elevation with zero pumping downstream.

Where the energy really comes from

There is no free lunch. You still had to lift the seawater 1000 m to start. That's 1025 × 9.81 × 1000 = 10.05 MJ/m³, or 2.79 kWh/m³. At 50% recovery (typical for RO), each cubic meter of fresh water represents 2 m³ of seawater lifted — 5.58 kWh/m³ raw. Grim, until you notice: the concentrated brine still sits in a 1000 m column. Draining it through a Pelton turbine on the way back down recovers 1050 × 9.81 × 1000 = 10.3 MJ/m³ ≈ 2.86 kWh/m³. Round-trip efficiency of 85% on the lift and 90% on the recovery turbine yields a net ~1.4 kWh/m³ — and the fresh water arrives at 744 m elevation, worth another 2.03 kWh/m³ in avoided distribution pumping. On paper, the tower beats conventional RO.

What would break first

• Membrane compaction. Polyamide RO membranes creep under sustained 70+ bar. You'd need ceramic or graphene-oxide membranes — a real 2020s materials frontier.
• Column pressure at the base. 100 bar is the pressure inside a scuba tank. The tower's lower walls need pre-stressed concrete like a penstock — roughly 1.5-meter-thick walls at the base for a 10 m diameter column (hoop stress σ = Pr/t ≈ 33 MPa, fine for high-strength concrete).
• Biofouling. A stagnant, sunlit seawater column is an algae paradise. The tower must be opaque and continuously flushed.
• Structural gravity load. A 1000 m × 78 m² water column weighs ~80,000 tonnes — roughly one Eiffel Tower's worth, all sloshing in the wind.

Practical site: a coastal cliff. Draw seawater to the top with wind or solar, let gravity do the pressurizing, harvest brine's PE on the way down. The tower isn't really magic — it's a mechanical battery whose "current" happens to be drinkable.