A tire has to roll cheaply, grip a wet road and not wear away — and the three fight each other. Two of them can be measured by an outside lab. The third is graded by the company selling the tire.
Soft rubber grips beautifully and wears out fast. Hard rubber rolls cheaply and lasts, and slides when the road is wet. The industry calls it the magic triangle: improve one corner and you usually pay for it in one of the other two.
This is not a business decision. It is a property of the material — the stickiness that holds a wet road and the softness that wastes energy are the same behavior in the same rubber. Chemistry has bent the triangle. Nobody has abolished it.
So a tire is a compromise, and somebody chooses where the compromise sits. Until this vote, in America, that somebody was whoever bought the tire, mostly without knowing they were choosing.
On 17 August 2026 the California Energy Commission adopted the country's first efficiency standard for replacement tires, 5–0. A tire built for sale in the state has to come in under 9 on the scale that measures its drag, and from 2033 under 7.1 — roughly the average of the tires car makers already fit at the factory. Wet braking has a floor of 1.0, which means exactly as good as a standard reference tire.
The state's own accounting: $1.50 more per tire in the first phase, $6.50 in the second, against about $180 of fuel saved over the life of a set. And the commission's finding on everything else — no adverse trade-offs.
Rolling resistance goes on a machine: the tire is pressed against a steel drum, spun up, and the drag is read off directly under an international procedure that fixes the load, the pressure and the temperature. Hand the same tire to a lab in another country and the same answer comes back.
Wet grip works by comparison. That floor of 1.0 does not mean one of anything — it means exactly as good as the reference tire, a specific standard tire with a part number that exists for no other purpose than to be what everyone else is measured against. Your tire and that tire run the same wet track on the same day, and the number that comes out is a ratio.
A limit is only a limit if somebody outside the company can check it. On these two corners, they can.
Tread life is measured in America by a system built in 1979. The manufacturer takes a convoy of cars to a 400-mile loop of highway near San Angelo, Texas, runs its own tires alongside a standard reference tire for 7,200 miles, measures the rubber lost, extrapolates the rest of the tire's life, and assigns the grade.
The manufacturer runs the test. The manufacturer extrapolates. The manufacturer assigns the number that goes on the sidewall. Seven thousand miles is a small fraction of a tire's life, so most of that number was never measured — it is a projection, made by the company and checked by nobody.
The federal agency sets the procedure and does not verify the result before the tire goes on sale. There are no routine audits. Reporting a grade lower than you measured is expressly allowed; reporting one higher is not. So a 400 from one brand and a 400 from another are not the same claim about the world.
None of that is hidden. It is how the federal rule is written, and the industry has known it for 47 years. California's own staff report says it plainly: manufacturers are able to under-report the results.
California wrote strict numbers on the two corners with an outside referee. On the third it could not, because there is nothing to write against. But the rule still depends on that third number — it leans on it harder than before.
The rule has ways out, and they are sensible ones. A tire built to last a very long time gets a looser rolling resistance limit: 0.4 of slack in the first phase, 0.7 in the second, and up to 1.4 for the longest-lasting. That is good design; it admits the triangle exists.
But the gate to that slack is the tread life grade. The self-run, self-assigned, under-reporting-permitted grade. The looser your limit, the more it rests on a number you wrote about yourself.
Winter tires sit outside the rule entirely: no limit, no reporting. Qualifying used to require a judgement call about whether a tire was really meant for winter. In July the commission struck that, which is ordinarily good practice — it replaces an official's discretion with an objective test. What remains is a snow traction figure and the three-peak snowflake on the sidewall.
So light truck all-terrain tires carrying that snowflake are now outside the rule altogether — a segment the commission's own testing put among the worst it measured, on drag and on wet grip alike. And that snowflake is awarded on the manufacturer's own test as well.
The wet floor can be checked: anyone who doubts it can buy the tire and run it against the reference. Nobody can do that with the third number. The hard number gets written where there is something to measure with, and the compromise migrates to the corner where there is not.
The law that gave California this power, passed in 2003, carries an instruction: the commission has to show its standards do not shorten average tire life. Not with a wear test — no controlled tread life testing was run for this rulemaking. The data came in as sidewall grades and one retailer's private mileage model. Pooled, it showed no meaningful relationship between wear and efficiency, and the commission concluded there was no adverse trade-off.
And there is one more sentence, in the cost analysis. The model that produced the savings figure assumes a four year tire life — in the world with the rule, and in the world without it. The same four years in both. Which means the model cannot show the harm it is being used to rule out. That is not a wrong answer. It is a question that was not asked.
The savings arrive as fuel, so you collect them by driving miles. At the state's assumed 10,413 miles a year, the tires pay back the markup in under seven months of a four-year life. Drive less and you still collect — just slower.
Europe is not a swap for California's approach: it is a low floor with a label above it. Mandatory minimums have been in place for years — a wet grip floor and a rolling resistance ceiling. But that ceiling sits at 10.5, on the same scale where California is heading for 7.1. It cuts off the worst tail and nothing else. Everything above is sorted by the label: a grade from A to E for fuel and for wet grip, on the price tag, so the buyer picks.
Price: somebody has to read it. By the European Commission's own report this year, 39% of buyers recalled seeing the label at all, down from 50% in 2017, and three quarters of British drivers take whatever the fitter recommends. It works, slowly — but a label moves the market at the speed of attention, and hands its savings to the people who were already going to read the small print.
Fix the actual defect: give the third corner a referee — an independent, repeatable way to measure how long a tire lasts, held by someone other than the manufacturer. International regulators are inching there, writing requirements for the first time for worn tires rather than new ones.
Price: building a test like that takes years, and every tire pays for it in its price. And tread life is where the compromise goes when the other two corners are pinned. Pin the third as well and the space engineers work in closes from all three sides at once — that is the version where tires really do leave the shelves. California also cannot do this one alone: the tread life system is federal. A state can write its own standard on top of that measurement, but it cannot repair the measurement. It is building on a ruler it is not allowed to fix.
The 2003 statute asks that replacement tires be at least as efficient as factory tires — on average. You could hold each manufacturer's California sales to that average, the way fuel economy rules hold a car maker's whole fleet, instead of holding every single tire to a line. Then the grippy, short-lived, brilliant tire survives, as long as the volume around it is efficient.
Price: you have to audit sales rather than products, which is a larger machine than testing rubber. The worst tire stays on the shelf, so the argument that a floor protects the buyer who does not know what to look for goes away. And an average is manageable: discount the good model, move more of it, and the number comes right without the bad model changing at all.
None of this makes California's rule a bad rule. The fuel savings are real, most tires on the road already clear the first step, and the commission pushed both deadlines back when the industry objected — which is what listening looks like. It is a rule with a known dependency: two of its numbers can be checked by anyone, the third cannot, and that is where the ways out are.
Which points at the cheapest design of all, and the only one California can build alone: leave the rule as written, and require that the grade opening the slack be confirmed by an independent lab rather than by the applicant. Somebody has to pay for that convoy in Texas. The state cannot repair a federal measurement — but it can decline to let an unchecked number through its own door.
So when any rule names a number, two questions sit underneath it that almost never make the news. Who runs the test? And who checks the answer? If the answer to both is the party being regulated, what you are looking at is not a limit. It is a promise, written in the language of a limit. It may well be kept. But you now know which number to ask about first.