The development race: how F1 teams find speed mid-season

The car that lines up for the first race of a Formula 1 season is almost never the car that takes the chequered flag at the last. What launches in winter testing is a hypothesis; what races in the autumn is the answer, rewritten a dozen times over. Between those two points sits one of the most intense engineering contests in professional sport — a development war fought in wind tunnels, supercomputers and back-to-back runs on Friday afternoons, where a tenth of a second can be the difference between the podium and the midfield.

What makes the modern era distinctive is that this war is no longer fought with money as an unlimited weapon. Two regulatory levers now shape almost every decision a design office makes: the Aerodynamic Testing Restrictions (ATR), which ration how much a team can develop, and the cost cap, which rations what they can spend doing it. Understanding how those constraints interact with the raw pursuit of downforce is the key to reading a season correctly — and to explaining why the pecking order you see in March is so rarely the one you see in December.

Where the lap time comes from

Not all lap time is created equal, and not all of it is available to be chased in-season. In broad terms, performance on a modern F1 car splits into three domains, and they are anything but equally exploitable.

Aerodynamics dominates the in-season development budget, and for good reason: it is where the regulations still permit the most freedom and where the marginal return on engineering effort is highest. A ground-effect car generates the overwhelming majority of its downforce from the floor and diffuser, so the floor is where the biggest single gains are found and where teams spend disproportionate resource. Around it sit the classic aero surfaces — the front wing, which sets the tone for how air is distributed down the entire car; the sidepods and bodywork, which manage flow to the rear and the cooling exits; and the rear wing and beam wing, which trade outright downforce against drag depending on the circuit. If you want a refresher on how these elements fit together, our anatomy of an F1 car guide walks through each in turn.

Mechanical performance — suspension geometry, damping, weight distribution, the way the car uses its tyres — matters enormously, but the scope for radical in-season change is narrower. Much of it is baked into the car’s fundamental architecture at design time, and altering it late is expensive and disruptive. Teams tune rather than reinvent.

Power unit gains are the most constrained of all, because the PU is homologated: the specification is locked for defined periods, and manufacturers cannot simply roll out a more powerful engine mid-season to claw back a deficit. Reliability fixes are permitted through a controlled process, but chasing horsepower is largely a between-generations game rather than a race-by-race one.

Domain	In-season development scope	Why
Aerodynamics (floor, wings, bodywork)	High — the primary battleground	Greatest regulatory freedom and highest marginal return
Mechanical (suspension, weight, tyre use)	Moderate — tuning more than reinvention	Largely fixed by the car’s base architecture
Power unit	Very low — effectively locked	Homologated specification; reliability changes only

The practical consequence is simple: when a team talks about “bringing an upgrade,” they almost always mean aerodynamics. The floor and front wing are where seasons are won and lost.

The wind tunnel and CFD

Before any new part is cut in carbon, it exists twice over: once as a computational fluid dynamics (CFD) model living inside a supercomputer, and once as a scale model bolted to the strut of a wind tunnel. These two tools are the twin engines of aerodynamic development, and they work best in tandem.

CFD lets engineers simulate airflow over a proposed geometry numerically — cheaply, in parallel, and without cutting metal. It is superb for exploring a wide design space quickly, screening dozens of ideas and discarding the losers before anyone commits real resource. But simulation is only ever an approximation of a turbulent, three-dimensional reality, and it can flatter a design that behaves quite differently in the physical world.

The wind tunnel provides that physical check. Teams run scale models — typically 60% of full size — on a rolling road that mimics the ground moving beneath the car, which is essential for a ground-effect machine whose performance is dominated by what happens under the floor. The model bristles with pressure taps and load sensors, and the tunnel measures the forces the design actually produces rather than the forces a computer predicts it should.

The development loop, then, is iterative: an idea is screened in CFD, promising candidates are validated in the tunnel, and the best of those are prepared for the track. Crucially, both CFD and tunnel usage are metered — they are the two currencies the ATR rations. A team cannot simply throw infinite compute or infinite tunnel hours at a problem, which turns aerodynamic development into an exercise in efficiency as much as ingenuity.

The Aerodynamic Testing Restrictions (ATR)

The ATR is one of the most deliberately interventionist rules in modern F1, and it is frequently misunderstood. It is a sliding scale: the amount of wind-tunnel and CFD development a team is permitted is tied inversely to its position in the constructors’ championship. Finish higher, and you get less. Finish lower, and you get more.

The logic is unapologetically equalising. The team leading the championship is allowed the fewest permitted runs and the least CFD allocation; the team propping up the standings is granted the most. The intent is to give trailing constructors a structural tailwind to catch up, compressing the field over time rather than letting the strongest teams pull away simply because they can out-develop everyone else.

Two features make the ATR genuinely dynamic rather than a one-off handicap. First, the allocation is recalculated during the season — the standings at defined checkpoints reset each team’s allowance, so a team that climbs the order mid-year sees its development freedom shrink as a direct consequence of its own success. Second, because the scale is pegged to championship position, it creates a subtle strategic tension: winning now costs you development capacity later, which is a real consideration for teams weighing how hard to push a given package.

None of this replaces engineering quality. A well-run team with less tunnel time can still out-develop a poorly run team with more, because the ATR meters opportunity, not competence. But at the margins — and F1 lives at the margins — it demonstrably tilts the playing field toward the chasing pack, and it rewards teams that extract the most learning from every metered run.

Correlation — the silent killer

Here is the uncomfortable truth at the heart of aerodynamic development: the numbers a team generates in the tunnel and in CFD are worthless unless they match what the car actually does on track. That match is called correlation, and when it breaks, a team can spend an entire campaign chasing performance that does not exist.

The problem is that the tunnel is a controlled, idealised environment — a scale model, a smooth rolling road, steady-state conditions. The real car operates in a chaotic one: full scale, variable temperatures, a flexing chassis, a rolling and pitching platform, tyres that deform and heat, and a floor that runs perilously close to a bumpy, undulating surface. If a team’s tools tell it that a new floor is worth downforce that never shows up on the stopwatch, the tools have lied — and the team may not know which of a dozen assumptions is at fault.

The consequences of poor correlation are brutal. A team can develop phantom gains: parts that test beautifully and deliver nothing, or worse, parts that make the real car harder to drive because they shift the aerodynamic balance in ways the tunnel never captured. Worse still, the team’s entire development direction becomes suspect, because every future decision is made using instruments it can no longer trust. Recovering from a correlation problem often means stepping back, running extensive validation to re-baseline the tools, and effectively pausing meaningful development until confidence is restored — an enormously expensive detour in a sport where the calendar never stops.

This is why the best teams treat correlation as a first-class engineering discipline rather than an afterthought. A tunnel that agrees with the track is a licence to develop aggressively; a tunnel that disagrees is a very fast way to spend a season going nowhere.

The cost cap

If the ATR rations how much a team can develop, the cost cap rations what it can afford to. It is a financial limit on relevant spending, and its effect on development philosophy has been profound.

In the pre-cap era, the largest teams could simply out-spend problems: design multiple parallel solutions, manufacture all of them, and pick the best at the track. The cap makes that impossible. Now every upgrade competes against every other upgrade for the same finite pool of money. A new floor that costs a chunk of the budget is a floor’s worth of something else the team cannot do — a rear wing family, a suspension revision, a spare set of components. Prioritisation is no longer a virtue; it is a survival requirement.

This forces brutal trade-offs. Do you spend on the upgrade that helps at high-downforce circuits or the one that helps everywhere a little? Do you accept the cost of manufacturing spares, knowing a crash-damaged part is money you will never recover? Do you commit budget to this year’s car when a rules change looms next year? Every one of these is a genuine dilemma, and the teams that manage the cap best are not necessarily the ones with the cleverest aerodynamicists — they are the ones with the sharpest judgement about where a fixed budget buys the most lap time. The cap has, in effect, made efficiency of spend a core competitive discipline alongside efficiency of airflow.

Upgrade cadence and validation

Upgrades do not arrive at random. Teams plan their development cadence across the season, sequencing packages to arrive when they deliver the most value — and, given the cost cap, when they can afford to introduce and support them.

Some upgrades are circuit-specific: a high-downforce package tailored to a twisty layout, or a low-drag configuration for a power circuit. Others are fundamental: a new floor concept or a reworked front wing intended to raise the car’s baseline everywhere. Teams often stage their season around a handful of major packages, with smaller refinements slotted in between, and they will hold parts back until a track suits them rather than burn budget introducing them where they add little.

Whenever a genuinely new part reaches the track, it must be validated before the team trusts it — because this is the moment of truth for all that tunnel and CFD work. The classic tool is the aero rake: a grid of pitot tubes mounted to the car during practice that maps the pressure field in the airflow, letting engineers compare the real flow structures against what the simulations predicted. Cars also carry extensive sensor arrays — pressure taps, strain gauges, flow-visualisation paint in early running — and teams run back-to-back tests, fitting the old and new specification on successive runs (sometimes split across the two cars) to isolate the upgrade’s true effect from the noise of fuel load, tyre state and track evolution. Only when the track data confirms the expected gain does a part earn its place. This validation step is the practical enforcement of good correlation: it is where the tools are held accountable to reality.

The pivot point

Every season contains a hidden decision that shapes the next one: the moment a team stops developing this year’s car and commits its resources to next year’s. This is the pivot point, and getting it right is one of the most consequential calls a team principal and technical director make all year.

Development on the current car and design of the next car draw from the same finite pools — the same people, the same ATR allocation, the same cost-cap budget. Every pound and every tunnel run spent chasing another tenth in the autumn is a pound and a run not spent on the following year’s concept. Pivot too early and you surrender positions — and prize money, and championship standing — in races you could have contested. Pivot too late and you arrive at the next season with an underdeveloped car, condemning yourself to a year of catch-up.

In a stable regulatory era the calculus is uncomfortable but manageable, because much of this year’s learning carries directly into next year’s car. The problem becomes acute when a rules reset looms. When the technical regulations change fundamentally, this year’s aerodynamic knowledge does not transfer cleanly — a new formula can render hard-won gains irrelevant overnight. That makes the pivot brutal: continuing to develop a car built to rules that are about to disappear can be almost pure waste, so the incentive is to switch focus early and eat the short-term pain. Which brings us to the situation the entire grid now faces.

The 2026 reset changes the calculus

The 2026 season is not a normal development year, because it is built on a new technical formula. When the fundamental regulations change — new power-unit rules, revised aerodynamics, altered car concepts — the usual continuity between one season’s learning and the next is broken, and every assumption about where lap time lives has to be re-examined from first principles. Our guide to the 2026 rule changes covers the specifics; here, the point is what a reset does to the development race itself.

A new formula scrambles priorities in three ways. First, it forces teams to re-establish correlation almost from scratch, because a redesigned car behaves differently in the tunnel and on track, and the confidence built up over years of the old rules cannot simply be assumed to hold. Second, it changes where the gains are: a shift in the aerodynamic or power-unit regulations can move the biggest opportunities from one part of the car to another, and the teams that read that shift correctly early gain a lead that compounds. Third, it raises the stakes on the pivot decision in the seasons around the change — commit early to the new concept and you risk under-developing while rivals score points; commit late and you risk arriving with a car built on the wrong ideas. A reset year is where the development race is at its most volatile, and where the order can be reshuffled more dramatically than at any other time.

Frequently asked questions

Why do teams with less money sometimes out-develop richer ones?

Because both spending and development are now rationed. The cost cap narrows the financial gap between teams, and the ATR actively hands more wind-tunnel and CFD time to constructors lower in the standings. A well-run team that extracts maximum learning from a smaller, ATR-boosted allocation can genuinely out-develop a richer rival that spends its resources poorly. Judgement and efficiency now matter as much as raw budget.

What exactly is a “correlation problem”?

It is when a team’s development tools — the wind tunnel and CFD — predict a performance gain that fails to appear on track, or predict a car balance that does not match reality. Because every future upgrade decision relies on those tools, a correlation problem undermines the entire development programme until the team can re-baseline its simulations against real-world data.

Does the ATR allocation change during the season?

Yes. It is recalculated at defined points based on the constructors’ championship standings, so a team that climbs the order mid-season sees its permitted development time shrink as a result. Success carries a built-in penalty in future development capacity, which is part of the mechanism’s equalising intent.

Why can’t teams just develop a more powerful engine mid-season?

Because power units are homologated — their specification is locked for defined periods, and manufacturers cannot introduce performance upgrades at will. Changes are permitted for reliability through a controlled process, but chasing horsepower is largely a between-generations exercise rather than a race-by-race one. This is why in-season gains are overwhelmingly aerodynamic.

When do teams stop developing the current car?

At the pivot point — the moment they judge that resources are better spent on next year’s car. In a stable rules era this can come late, because learning carries over. Ahead of a rules reset like 2026, teams tend to pivot earlier, because knowledge from the outgoing formula transfers poorly and continuing to develop a soon-to-be-obsolete car can be close to wasted effort.

The bottom line

Modern F1 development is a contest of judgement as much as engineering. Aerodynamics — above all the floor and front wing — is where in-season lap time lives, but chasing it is now bounded on every side: by the ATR, which meters tunnel and CFD time on a sliding scale that favours the chasing pack; by the cost cap, which forces every upgrade to compete for the same finite money; and by correlation, the discipline that keeps all that effort honest. Layer on the strategic knife-edge of the pivot point, sharpened by the 2026 reset, and you have the real story behind the standings. The team that wins the development race is rarely the one that simply spends the most — it is the one that decides best what to build, when to build it, and when to walk away.