Infrastructure | The Undercut Wiki

Infrastructure is the permanent skeleton of the team: nine departments, each a level you buy once and pay for forever. Levels set your car’s baseline performance, its reliability ceiling, how many work units R&D and Marketing and Scouting get every turn, and how big a deal you can negotiate. Everything else in career mode — parts, contracts, drivers — is built on top of these nine numbers.

How to open it

Sidebar / desk Infrastructure icon, or the hub’s Infrastructure entry.

It is only editable in the pre-season. The moment a season is running (season_state exists) the app opens View Only — the title reads “Infrastructure (View Only)” and the +/− buttons vanish. The Driver Academy app is also reached from here.

Controls: Up/Down or a mouse click selects a department, + / Right / Enter buys the next level, − / Left scales one back, Escape closes. On a gamepad the right stick walks the list and the triggers buy/sell.

Layout: a department table on the left (level, progress bar, upgrade cost over running cost, adjust buttons) with your Development Circuit below it, and on the right a Budget Summary (available budget, total running cost, projected budget and upkeep after the selected upgrade) plus a Selected Department panel describing what the next level buys.

The department table

Nine departments, index-ordered. Eight run level 1–10 and start at 1. The Driver Academy runs 0–3 and starts at 0 (CLOSED).

#	Department	What it actually drives
0	Engineering	+0.04 global weight on every part · reliability floor · the Repair/Improve WU pool · repair efficiency
1	R&D	Sets the `potential` ceiling of every part · reliability · the design-project WU pool · design quality
2	Chassis	Body 0.12, Floor 0.08 · Suspension 0.03, Cooling 0.04
3	Handling	Brakes 0.10, Suspension 0.08 · Chassis minor 0.03
4	Aero	Front wing 0.12, Rear wing 0.12 · Floor 0.04
5	Engine	Gearbox 0.12, Electronics 0.12, Fuel system 0.08, Cooling 0.08 — does not touch the engine part
6	Marketing	Income multiplier · negotiation WU · sponsor visibility · deal quality · supplier discount · reputation
7	Scouting	Scouting WU per GP · passive knowledge of every driver
8	Driver Academy	Junior + development driver slots (special: 0–3, expensive, budget-gated)

Build cost and running cost

Eight generic departments:

upgrade_cost(current_level) = 500,000 × (current_level + 1) ^ 1.5
running_cost(level)         = 50,000 × level  per GP

Buying level	Cost	Cumulative from Lv1
1 → 2	$1,414,213	$1.41M
2 → 3	$2,598,076	$4.01M
5 → 6	$7,348,469	$20.95M
9 → 10	$15,811,388	$70.8M

Taking one department from 1 to 10 costs about $70.8M and $500,000/GP of upkeep at the top. Doing it to all eight is not a thing any team achieves.

Driver Academy (deliberately ~8× the curve and ~3× the upkeep):

upgrade_cost(current_level) = 4,000,000 × (current_level + 1) ^ 1.4
running_cost(level)         = 150,000 × level  per GP

Buying level	Cost
0 → 1 (opening)	$4,000,000
1 → 2	$10,556,000
2 → 3	$18,621,000

Downgrading refunds half the price of the level you are giving up (upgrade_cost(level − 1) / 2) and immediately lowers upkeep. It is a real cost-cutting lever, not a mistake-undo — you get 50 cents on the dollar.

The running-cost line lies slightly. The per-department /GP figures in this app and in Finance are printed raw. The TOTAL RUNNING COST readout (and the Net per GP in Finance) applies infra_running_cost_multiplier — default 0.75. Your real upkeep is 75 % of what the rows sum to.

The budget gate on the Driver Academy

Opening the academy (0 → 1) is the only upgrade in the game with a wealth check on top of the price:

ACADEMY_OPEN_MIN_BUDGET = $8,000,000 in the bank

Below that, the button refuses with “Opening a Driver Academy needs a healthy war chest.” Once open, upgrades follow the ordinary can-you-afford-it rule.

Slots by level:

Level	Junior (16–18) slots	Development (18–21) slots
0 (closed)	0	0
1	2	1
2	3	2
3	4	3

Empty slots auto-fill with freshly rolled prospects at the season rollover and when the Academy app is opened; prospect quality scales with level (plus the academy_roll_quality slider). A development driver must leave at 21; a junior graduates after 2 seasons. Signing your own development driver to the race team gets an acceptance boost (academy_dev_sign_boost, default +0.20). Prospects can also be loaned to PF2/PF3 teams, which pays you a per-GP fee and gives them +5 % development.

Engineering (index 0)

The only department that touches every part on the car, plus the pool that pays for repairs and improvements.

part performance     : + 0.04 weight on all 11 parts
default fiability    : 0.85 + norm(eng) × 0.15 + norm(rnd) × 0.10, clamped ≤ 1.15
supplier-backed parts: 0.85 + norm(eng) × 0.10 + norm(rnd) × 0.08 + norm(sup_rel) × 0.12
work-order pool      : 50 + engineering_level × 10  WU   (Lv1 = 60, Lv5 = 100, Lv10 = 150)
repair efficiency    : 0.5 + engineering_level × 0.05
repair WU cost       : ceil(perf deficit / efficiency × 100 × engine architecture multiplier)

(norm(v) = (v − 1) / 9, so level 1 contributes nothing and level 10 contributes the full weight.)

Engineering is the department that makes damage cheap. At Lv1 (efficiency 0.55) a repair costs nearly twice the WU it costs at Lv10 (efficiency 1.0) — and the pool it draws from is 2.5× smaller. Each active supplier partnership adds a flat +5 WU/turn to this pool on top.

R&D (index 1)

R&D never adds performance directly. It raises the ceiling every other department can develop toward.

base_potential = 1.0 + norm(rnd) × 0.15                 # 1.00 at Lv1 … 1.15 at Lv10
per-part        = base_potential ± 0.03 (deterministic per part), clamped 1.00 .. 1.20
design capacity = 20 + rnd_level × 8  WU per GP         # Lv1 = 28, Lv5 = 60, Lv10 = 100
design quality  = 0.6 × responsible_dept + 0.4 × rnd    # feeds the part-quality roll
fiability       = + norm(rnd) × 0.08 .. 0.10

potential is a hard wall: Improve stops at the part’s own potential, and only the head-room a lucky quality roll or a practice development program pushes past 100 display points becomes overdrive. With R&D at 1, every part you design is capped at 100 points — you can develop it to the ceiling in a couple of grands prix and then have nowhere to spend WU.

The design capacity is separate from Engineering’s pool: it is allocated to design projects and locked until they deliver.

Chassis · Handling · Aero · Engine (indices 2–5)

The four specialist departments set each part’s baseline performance and, just as importantly, the improve rate of the parts they own.

perf(part) = 0.896 + Σ (norm(level) × weight × 1.30)        # PERF_SPREAD = 1.30

The 1.30 spread is what makes a top team’s car ~21 % better than a backmarker’s rather than ~17 %.

Improvement per 15-WU step is driven by the responsible department:

improve_gain = (0.004 + dept_level × 0.002) × rnd_dev_rate_multiplier    # slider, default 2.0

Dept level	Gain per step (default ×2.0)	In display points (1 pt = 0.0016 perf)
1	+0.012	~7.5 pts
5	+0.028	~17.5 pts
10	+0.048	~30 pts

Component → department mapping: front wing / rear wing → Aero; floor / body → Chassis; suspension / brake → Handling; gearbox / electronics / fuel system / cooling system → Engine. The engine part itself is driven by your engine supplier’s base_power, not by the Engine department — the Engine department only does the four ancillaries.

So a low-level department hurts twice: a worse baseline part and a slower climb toward its ceiling.

Marketing (index 6)

No car effect at all, and probably the highest-leverage department in the game. Every commercial formula reads it:

Effect	Formula	Lv1 → Lv10
Income multiplier	`1 + (level − 1) × 0.05`	×1.00 → ×1.45
Negotiation capacity	`10 + level × 5` WU/turn	15 → 60
Negotiation speed	`1 + 0.03 × level` per WU	×1.03 → ×1.30
Deal quality (permanent)	`1 − 0.02 × (10 − level)`	×0.82 → ×1.00
Sponsor visibility	brands with `score ≤ 30 + level × 7`	~half the pool → all of it
Supplier client discount	`progress% × 0.02 × level`, cap 20 %	≤2 % → ≤20 %
Sponsor lottery draws	`level × sponsors_per_marketing_level` per pre-season	1 → 10 draws
Reputation, every pre-season	`min(level − 1, 8) / 2`	+0 → +4

The multiplicative stack is the point: at Lv10 you see the biggest brands, negotiate them 30 % faster, sign them at 100 % of face value, are paid 45 % more on the way in, and you unearth ten free pre-negotiated sponsors a year. At Lv1 you see the small half of the market and permanently pay an 18 % haircut on every contract you sign.

Scouting (index 7)

scouting capacity = 10 + level × 5  WU per GP           # Lv1 = 15, Lv5 = 35, Lv10 = 60
knowledge gained  = 1 % per WU allocated per GP

At 20 WU/GP on one driver, 0 → 100 % knowledge takes ~5 GPs. Capacity is split across ongoing observations, so a big network watches several candidates at once rather than one faster.

Level also sharpens the passive knowledge you have of everybody without spending a WU:

baseline = ((20 if racing in a series else 5) + renown × 0.35) × (0.5 + 0.05 × level)
           capped at 60 %

Lv1 scales passive intel at 0.55×, Lv10 at 1.0×. Active scouting is always required past 60 %.

Development Circuit

The panel under the table shows your team’s home test venue — name, turn count, length, mini-map — derived deterministically from your team’s country. It is the same venue the R&D Private Testing programme runs on.

Advanced

Upgrades are pre-season only for you, but not for the AI. An AI team in a Comfortable posture sitting on dormant cash buys one facility level per GP all season (weakest generic department first), and an AI team bleeding money sheds one level per GP (highest non-Academy department). Your nine numbers are frozen from the season opener; theirs are not.
The AI’s rollover budget for infrastructure comes from its finance model’s posture: Comfortable allows 1 (Small) / 2 (Medium) / 3 (Large) upgrades above a cash reserve; Tight allows 0–1 and downgrades one department if net/GP is negative; Critical never invests and downgrades 1–3. The AI Financial Intelligence slider scales it.
Infrastructure is a third of team value. team_value = (sum of all 9 levels) × $1.8M × tier × reputation + results. Every level you buy is also collateral for a bigger loan and a higher share price. Downgrading shrinks both.
Reputation feeds back into infrastructure indirectly. The season-end re-evaluation weighs infra average at 25 % (championship 50 %, income rank 25 %), drifting 20 % of the way toward the deserved figure each year. A serious facility reads as a serious team.
Engine reference infrastructure. A supplier’s engine is derived from a fixed neutral infrastructure (all departments at 5), so every customer of the same motorist gets an identical engine regardless of their own facilities. Your Engineering level still contributes the global +0.04 on top.
Part-quality rolls read 0.6 × dept + 0.4 × rnd. Higher infrastructure biases the loot tier upward (quality_infra_bias, default 0.40 — luck still dominates), which is where overdrive head-room comes from.

Upgrade or spend it on parts?

A rough decision frame, given a ~$70.8M price for one department’s full climb and a ~$10.5M annual car rebuild:

Buy Marketing first, almost always. It is the only department that pays for itself in cash, and its effects compound across every future contract. A single level widens what you can see, sign and be paid.
Buy Engineering before you buy performance. WU are the currency that turns everything else into lap time; at Lv1 you have 60 WU and expensive repairs, which means a damaged weekend eats your whole development budget.
Buy R&D when you are hitting the ceiling. If your fitted parts are sitting at their potential with WU left over, R&D is the only thing that raises the wall. If they are not, R&D buys you nothing this season.
Buy the specialist departments last, and narrow. Aero at 0.12 on two parts and Handling at 0.10 on brakes are the biggest single-part levers; spreading four departments to level 3 each is worse than taking one to 6.
Parts win the current season; departments win the next one. Manufacturing spare wings and improving fitted parts pays out this weekend. A department level pays out from the next pre-season onward — and forever after.
The Driver Academy is a luxury. $4M to open, $33M to max, $450k/GP at Lv3, and its output is prospects who need years. Open it when you are comfortably above the $8M gate and your car is competitive — not as a rescue plan.

Common mistakes

Summing the per-department running-cost rows. They ignore the 0.75 global multiplier; the TOTAL readout does not. You are paying less than the rows suggest.
Expecting the Engine department to improve the engine. It does not. Gearbox, electronics, fuel system and cooling only — the engine part comes from your engine supplier.
Buying R&D at level 1 infrastructure. Potential you cannot reach is wasted money. Raise the department that owns the part (its improve rate) before raising the ceiling.
Upgrading during the season. You cannot — the app is View Only once racing starts. Plan the whole year’s facility spending in pre-season phase 1, before the sponsor and driver bills land.
Using downgrade as an undo. You get half of one level back, and you lose the level immediately. It is a survival lever, not a refund.
Opening the Academy with $8.1M in the bank. The gate checks cash, not sustainability: you will then pay $150k/GP forever, in a season where three red weekends in a row end the career.