Most buyers shop for an injection mold by cavity count and unit price, then get blindsided when the tool starts throwing flash on shot 200,000. The metal inside that mold decides most of what happens next, and it's the one line item procurement teams skip first. We've built molds in Dongguan for two decades, and the steel question comes up in nearly every first call with a new overseas buyer. This piece walks through what injection mold metal really means on the shop floor — not the catalog specs, but what actually moves your cost per part and your scrap rate.
What metal are injection molds made from
Short answer: tool steel, almost always. A mold base might be a cheaper carbon steel like 1045 or S50C, there only to hold everything rigid and square. The parts that touch your plastic — the core and the cavity — are where the grade stops being a detail and becomes the whole job. We run P20, 718, H13, S136, and NAK80 on a regular week, and each one machines, heat-treats, and wears differently once it's clamped in a press.
A pre-hardened block such as P20 already sits at roughly 28–32 HRC when it lands on the mill. You can shape it fast and skip heat treatment entirely, which is exactly why it's the default for prototypes and short runs. Injection molding buyers chasing a quick bridge tool love it for that reason. Yet the moment you need a million clean shots, P20 simply isn't the metal you want holding your tolerances.
People also mix up the base steel with the cavity steel, and that confusion quietly costs money. The base can be plain 1045 and nobody cares, because it never touches plastic or takes the wear. The cavity is the opposite — it's the only thing your customer sees and the only thing that erodes. Heat treatment is where the hardened grades earn their name: we rough the core, send it out to harden, then bring it back to finish the close-tolerance features, because a block shifts a few thousandths when it's quenched. Skip that second pass and you're cutting steel that has already moved on you.
Pre-hardened versus hardened, and why it moves your unit cost
Here's the part most quotes quietly skip. A hardened tool costs more up front — extra machining, a heat-treat cycle, and often a second machining pass after hardening because the block shifts a few thousandths. But it spreads that premium across far more parts. We once had a client from Poland who pushed back hard on a hardened H13 quote for a dishwasher bracket, insisting P20 was "good enough." Six months later his line was dressing the parting line every week and the ribs were already sinking. He came back for a re-tool in H13, and the per-part math finally made sense to him. On that bracket the P20 tool ran about $9k and died near 250k shots; the H13 re-tool was $15k but was still clean past 1.2M. Spread the tool cost across the parts and the 'expensive' steel landed at roughly a third of the price per piece.
The cheat sheet below is the one we actually hand to buyers before they commit:
| Steel grade | Hardness (HRC) | Best for | Typical tool life |
|---|---|---|---|
| P20 | 28–32 | Prototypes, runs under 100k shots | ~100k–300k |
| 718 (pre-hard) | 30–35 | Medium runs, better polish | ~300k–500k |
| H13 (hardened) | 44–52 | Abrasive resins, long runs | 1M–3M+ |
| S136 (stainless) | 48–52 | Medical, corrosive, clear parts | 1M+ |
| NAK80 (pre-hard) | 37–43 | Mirror finish, no EDM marks | ~500k–1M |
Notice the life column isn't a sales number. It's what we see from tools running in real presses with real resins, not a lab bench. A mold making services shop that quotes you H13 for a 5k-piece pilot is either padding the job or doesn't understand your volume — both are worth a second look.
How we pick the steel in our own shop
We start from the resin, every time. Glass-filled nylon at GF30 and above will chew through soft steel like sandpaper, so anything abrasive goes straight to H13 or S136. Clear polycarbonate for a lens wants NAK80 because it takes a mirror without the pitting you get from spark erosion on softer grades. After testing hundreds of molds, we've learned the expensive steel is cheaper when the volume is there — and the cheap steel is honestly the right call for a bridge tool you plan to throw away.
On the ground, the decision is rarely "best steel." It's best steel for this part, this volume, this budget. A factory in Vietnam we support runs a weekly trial mold in P20 to validate a new clip, then commits to H13 only after the design freezes. That two-step habit has saved them a fortune in re-work. Precision mold making lives or dies on exactly that kind of discipline. In abrasive jobs we also plate the wear areas or fit insert sleeves at the gate, because that's where glass-filled resin eats first — a $200 sleeve swap beats a $3k cavity re-cut, and we spec it into the drawing instead of waiting for the wear to show.
Lead time rides on the metal, too
The steel choice also sets your calendar. A P20 tool can ship in two to three weeks because there's no outside heat-treat loop and no return pass. Hardened H13 or S136 adds a heat-treat cycle plus the finish pass back in our shop, so plan on five to seven weeks for the same cavity count. We've had buyers pick the faster steel purely because a launch date left no room for the better one, then regret the tool life later. If the date is tight, tell us early — sometimes a pre-hardened 718 bridges the gap while we harden the production tool in parallel.
Finish is the quiet hero, and mirror EDM does the heavy lifting
Most people assume a shiny surface comes from hand polishing. Sometimes it does. But for the really clean lenses and cosmetic caps, we go straight to mirror EDM — the electrode dresses the steel to a finish a stone can't reach. We run wire EDM and mirror EDM in-house, which means the cavity finish is our call, not a subcontractor's guess. CNC machining gets the geometry close; EDM and polish close the last micron.
One caution worth stating plainly: a mirror finish on the wrong steel shows every flaw. Softer grades pit under the spark, and you end up re-working the cavity you just paid for. That's another reason the steel choice and the surface plan have to be decided together, not in two separate email threads a week apart.
| Surface need | Route we take | Steel that behaves |
|---|---|---|
| Matte, hidden interior | Straight CNC, light stone | P20, 718 |
| Semi-gloss, visible | Stone polish plus bead | 718, NAK80 |
| Optical mirror | Mirror EDM, no stone | NAK80, S136 |
| Textured grip | EDM or chemical etch | H13, S136 |
Why steel stability protects your tolerances
A stable steel is the reason a part holds ±0.01 mm shot after shot. Soft, pre-hardened grades relax a little under clamp pressure, and across a long run the dimensions drift just enough to trip a CMM. Hardened H13 and S136 barely move, which is why automotive and medical programs almost always spec them. We measure every critical dimension on the CMM before a tool ships, and the stable grades are the ones that pass without a fight.
A medical client in Germany learned this the hard way. They'd been running a clear PETG cap in P20 to save cost, and the tool rusted at the water lines inside a year of weekly sterilization. We moved them to S136, and eighteen months later the cavity still mirrors. The resin hadn't changed — only the metal did. Corrosion is the quiet killer most budgets forget to price in. Water and food contact push the same direction: a client running a beverage closure switched from H13 to S136 after the first tool stained at the cooling channels, and the stainless grade shrugged off the citric-acid wash. If your part sees water, steam, or any cleaning chemistry, stainless stops being a luxury.
When a tool is done — and how to stretch it
A mold doesn't die all at once. First the parting line needs more frequent fitting. Then flash shows up at the gate, then at the parting line, then the corners. We log every tool's shot count and repolish the cores on a schedule, which is far cheaper than a surprise re-cut that stops a line. For a hardened tool doing, say, two million shots, swapping a worn insert runs maybe a tenth of a new tool and buys another million. A routine core repolish on our floor costs a few hundred dollars and resets the surface for another 200k to 400k shots — cheap insurance against a five-figure surprise.
We've had automotive clips run past five million on a single H13 core with two insert changes. The trick is catching wear on the CMM trend, not waiting for the operator to notice flash. An honest limit: no steel is immortal, and abrasives win eventually. We tell buyers the expected life up front so the decision never lands on their production manager's worst week. At some point a fresh core costs less than the downtime it's causing — and planning that exit before you start is what keeps plastic injection molding cheap instead of a slow bleed.
Three things to take into your next quote
Pick the steel from the resin and the volume, not from the headline price — that single habit fixes most of the surprises we get called about. Pre-hardened grades win on speed and prototype cost; hardened grades win on life and per-part economics, and the crossover is usually lower than buyers expect. And decide the finish and the steel in the same conversation, because they're really one decision wearing two hats. Get those three right and the tool mostly takes care of itself.
About SHINY Mold
SHINY Mold has built precision injection molds and machined parts in Dongguan since 2003, from a 22,000 m² plant with 120+ engineers and 100+ molding machines. We take a part from DFM analysis to mass production, with in-house CNC, wire EDM, mirror EDM, and CMM so tolerances hold at ±0.01 mm. When you're scoping your next tool, our full range of precision mold making and injection molding work is worth a look.
