The weakness of net zero is not in its ambition. It is in its form. The goal is written as a balance: emissions on one side, reductions and removals on the other, the target being that the two cancel. The trouble starts the moment a goal is written as an equation, because an equation cares only about its sum. It does not ask how the sum was reached.
Begin with the part most people reach for, then the part that matters more.
Real decarbonisation has two kinds of route. One is to reduce the activity that produces the emissions. The other is to keep the activity and balance it, through a technological fix or by buying offsets that count against the total. Reduction is the route people resist, because it asks them to give something up, and demand-side change is consistently the slowest and least popular lever to pull. It is not that no reduction happens. Efficiency standards, modal shift and process change are all real. It is that reduction is chronically under-supplied relative to what is needed, because it is the route that costs comfort rather than money. So the weight falls on the routes that can be bought.
This is where offsets get misused, and it is worth being precise about the fault. Offsetting was never designed to deliver net zero on its own. In the frameworks that define the term, including the Science Based Targets initiative, offsets are meant to mop up a small residual of genuinely hard-to-abate emissions after deep absolute cuts have been made. In practice they become the first resort rather than the last, because buying a credit is easier than changing behaviour. The concept is sound. The use of it is not. That distinction matters, because it locates the problem exactly: not in net zero as designed, but in the way a balance metric invites you to satisfy it by the cheapest available means.
That brings us to the more important point. Because net zero is a ratio, it makes every route that reaches balance look equally good. The accounting sees a tonne and a tonne. It does not see how the tonne was handled. But the routes are not equally good. You can reach balance through means that are quick, buyable, and harmful in other ways: non-additional credits, plantations that displace ecological cost elsewhere, removals that will not hold for as long as the carbon stays in the air. Or you can reach it through means that are slow, difficult and often more expensive, but genuinely sound: real reduction, durable removal, change that leaves the system healthier than it found it. The metric treats these as identical. Worse, it tilts toward the harmful route, because cheapness and low quality travel together, and the equation only asks to be satisfied, not to be satisfied well.
A common objection is that offsets simply move the problem to another part of the world. In pure climate terms that is not quite right, and conceding the point makes the real critique stronger. A genuinely additional, permanent removal does not relocate anything. It removes. A tonne is fungible in radiative-forcing terms wherever it is handled. What is not fungible is everything the tonne-count ignores: permanence, additionality, and local ecological and social harm. Stated as geography, the objection is answerable. Stated as quality, it is not. The balance hides precisely the dimension on which the routes differ.
So the danger is not that net zero is the wrong goal. It is that a goal expressed as a balance is a proxy, and a proxy can always be met by the path that games it most cheaply. This is the mechanism set out in proxy-capture: once the measure becomes the target, the measure and the territory come apart, and effort flows to the number rather than the thing the number was meant to stand for. It is also a case of overoptimisation: the system improves its own headline metric while shedding the capacity that actually mattered. The ledger reaches zero. The atmosphere does not notice the difference between a real removal and a clever one, but the equation cannot tell them apart, and so neither, in the end, can the people relying on it.
The fix is not to abandon the target. It is to stop trusting the sum on its own. A balance that cannot distinguish a good route from a bad one is not a measure of progress. It is a measure of bookkeeping.
Further reading
from-optimised-to-resilient examines what it means to pursue resilience rather than a single optimised number, which is the deeper shift this argument points toward.
Garden notes
proxy-capture - When a measure becomes the target, the measure and the territory separate. Net zero is the climate version: the balance becomes the goal, and the balance can be gamed.
overoptimisation - Systems that narrow to the measurable shed unmeasured load-bearing capacity. Reaching net zero by the cheapest path is the same move applied to carbon accounting.