How to Remove Old Epoxy and Paint
The material decides the method, and how it behaves under attack decides the price. The counterintuitive part is that a floor which was installed well is the expensive one to take off.
Settle two questions before you price anything. First, does it have to come off at all, or is it sound enough to abrade and recoat? Second, how does it behave when you attack it: brittle coatings shatter and suit blasting or grinding, soft or thick ones smear and need tooling that shears rather than abrades, and a thin, well bonded coating does neither and simply grinds off slowly. A ten minute test patch answers the second question for this floor. Then quote removal and profiling as two operations, because removal tooling leaves a surface, not a profile, and the bare slab underneath still needs everything a bare slab needs.
First decide whether it has to come off
Removal is expensive and slow, and it is not always the job. A coating that is genuinely sound, bonded across the whole floor and compatible with what is going over it can be abraded to a profile and recoated, and that is a much smaller piece of work.
Prove it rather than assume it. Score a cross hatch through the coating to the concrete in several places across the floor, press tape over it and pull: what lifts tells you what is holding. Tap the floor and listen for the change in note that says an area is hollow underneath. Check the places the floor gets used hardest, not the corner nobody walks on, because a coating fails where the load is.
What does not work is a partial answer. A floor letting go in patches is a floor that is failing everywhere at different speeds, and leaving the areas that have not lifted yet means bonding the new coating to something already on its way off. Once meaningful areas are coming up, the honest scope is the whole floor, and saying that at quote time is much cheaper than discovering it at week two. Why epoxy floors peel covers how to tell which failure you are looking at, since the cause decides whether the replacement will last any longer than the original.
What you are removing decides everything
You can usually find out more from a knife than from the building’s paperwork. Cut a small notch through to the slab and look at the cut edge under a light. It shows how many layers there are, roughly how thick they are, and where the weak plane sits: between two coats, or right at the concrete.
Multiple layers matter more than people expect, because a floor that has been recoated three times is three materials with three behaviors, and the machine meets them one after another. It is also the usual reason a removal rate that looked fine for the first hour collapses in the second.
What you will often not discover is the actual chemistry, and the useful realization is that it rarely changes what you do. The method is chosen by behavior, not by name.
The three behaviors, and what each one needs
It shatters
- Thin, hard and brittle: old floor paint, thin epoxy, a coating that has already gone chalky with age
- It breaks into flakes and chips rather than deforming, and it often lifts ahead of the tooling
- Shot blasting is efficient here, and grinding is fast; this is the cheap end of removal
- The risk is complacency: it comes off so easily in the test patch that nobody budgets the profiling pass
It smears
- Soft, flexible or thick: urethanes, elastomeric coatings, heavy build systems, adhesive residue
- It deforms instead of cutting, packs between the segments and blinds the tooling within minutes
- Abrasive diamond and shot blasting both fail on it for the same reason, so neither is the answer
- It needs tooling that shears: polycrystalline diamond, or a scarifier where the build is heavy
It is simply stuck fast
- A thin, well installed coating on a properly prepared slab, which is the job somebody did right
- It neither shatters nor smears. It comes off as dust, slowly, at whatever rate the tooling cuts
- This is the expensive one, and it is the case people most often underquote
- There is no clever method for it. There is a rate, measured on the floor, and an honest number
That last one is worth sitting with, because it inverts the intuition. The better the original installation, the more the removal costs. A delaminating floor is half off already and comes up in sheets. A floor that was ground, primed and bonded properly has to be cut away from the concrete a fraction of a millimeter at a time.
The methods, and where each one stops
Grinding with removal tooling
- The general answer, and the only one that reaches edges, corners and anywhere a machine cannot drive
- Polycrystalline diamond shears coating off rather than abrading it, which is what defeats soft films
- Tooling choice here is not the tooling you would profile with, so plan on changing sets partway
- Slow on thick build, and it is the step everything else eventually hands back to
Scarifying or planing
- Cutters on a rotating drum, hammering material off rather than cutting or abrading it
- The right call for heavy build, thick industrial systems and coatings a grinder cannot get under
- Aggressive enough to damage the slab if run too deep, which turns preparation into repair
- Always leaves grooves. It is a removal step and never a finishing one
Shot blasting
- Fast across a large open field on a brittle coating, and close to dust free when maintained
- Fails on soft and thick coatings, which drive the abrasive in rather than fracturing
- Cannot touch an edge, a corner or anything it cannot drive over, so it never finishes a floor
- The full comparison is in grinding vs shot blasting
Chemical stripping
- Worth it for detail work, stair nosings, plant rooms and floors no machine can be brought to
- Slow and messy on open floor, and the waste is a disposal question rather than a sweeping one
- Residue left behind is a bond breaker, so the removal is only done when the floor is genuinely clean
- It wets the slab, which puts you behind the moisture gate before you can coat
Removal and profiling are two operations
This is the single most expensive misunderstanding on a removal job, and it is easy to make because both operations look like a man behind a grinder.
Removal tooling is built to get a coating off. What it leaves behind is uneven, commonly gouged, and often still carrying residue in the low spots where the tooling bridged over. It is not a prepared surface and it is nowhere near the profile the next product specifies. Frequently it is too coarse rather than too fine, which costs coating as well as time.
So a removal job has two rates, two tooling sets and two lines on the quote: take it off, then prepare the slab underneath exactly as you would prepare any bare slab. The prep pillar covers what that second stage still has to satisfy, including the moisture gate, which an old coating may have been hiding for years.
Price the removal from a rate you measured, not a rate you assumed
Run a test patch, time it, and use that rate here for the removal stage. Then run the calculator again at your normal rate for the profiling passes that follow, because those are a separate operation on the same floor.
The seeded rate is for grinding a bare slab. Removal has its own rate, and only the floor can tell you it.
Open the live calculator →Why removal cannot be quoted by area
Preparation on a bare slab is reasonably predictable, which is why a rate per square foot works for it. Removal is not, and the spread between an easy floor and a hard one is not a percentage, it is a multiple. Five things move it, and none of them are visible from a floor plan.
- How well it is bonded. The counterintuitive one. A failing floor comes up in sheets and a well installed one does not come up at all until it is cut away
- How thick, and how many layers. Each recoat is another pass, and the layers can behave differently from one another
- Which behavior it has. Shattering, smearing and grinding away as dust are three different rates on the same machine
- How much of the floor a machine can reach. Edges, corners and plant rooms are slow, and on a removal job they are slower still
- What has to happen to the dust. Containment, an occupied building or a regulated waste stream all change the job rather than decorate it
This is also the honest footnote to the redo arithmetic on the peeling page. That example prices a removal as one more grinding pass, and calls it the cheapest possible redo for exactly this reason: it is the delaminated case, where the coating is already half off. A floor that was installed properly and failed for some other reason costs considerably more than that to take back to bare concrete.
Before the machine starts
Two questions belong before the work rather than during it, because getting them wrong does not produce a bad floor, it produces a contamination event.
Lead and asbestos. Lead was not banned from residential paint in the United States until 1978, and lead containing coatings were used industrially as well. Some pre 1980s resilient flooring adhesives, black cutback mastic in particular, contain asbestos. If the age and history of the building make either plausible, have it tested before anything disturbs the surface. A positive result changes what the job legally is, who is allowed to do it, and how the waste leaves site.
Dust and containment. Removal generates far more material than preparation does, and it is not only concrete dust: it is the old coating as well. Extraction has to be matched and running, the area has to be separated from the rest of an occupied building, and the debris has a destination that is decided before it is on the floor rather than after.
Where to go next
If what you are pulling off is flooring rather than a coating, the material underneath has a name you can usually establish, removing carpet and tile adhesives covers why that changes the tooling: soft adhesive and black cutback smear, while thinset is a cement and grinds like the slab.
If what you are looking at turns out to be thin rather than built up, this is the wrong page for it. Removing sealers and curing compounds covers why the two operations rule above does not carry across to a film that is thinner than the depth the profile pass cuts to anyway.
Once the old floor is off, the bare slab starts from the beginning, and how to prep concrete for epoxy covers the four gates in order, including the moisture test that an old coating may have been masking. For the tooling decision under this one, bond vs concrete hardness explains why the set that removes well is rarely the set that profiles well.
For planning the hours, how long concrete grinding takes turns passes into a schedule, and grinding cost per square foot turns that schedule into a priced line. And for the failure that puts most floors in front of this decision in the first place, why epoxy floors peel separates the four causes that look identical on the day.
Frequently Asked Questions
Can I just coat over the old epoxy instead?
Sometimes, and it is always worth asking before you price a removal. If the existing coating is genuinely sound, well bonded across the whole floor and compatible with what is going over it, it can be abraded to a profile and recoated. Check adhesion in several places rather than one: score a cross hatch through to the concrete and try to lift it with tape, and tap the floor for hollow areas.
What does not work is a partial answer. A coating that is failing in patches is failing everywhere, just at different speeds, and leaving the parts that have not let go yet means the new floor is bonded to something that is on its way off. Once you are removing meaningful areas, the honest scope is usually the whole floor.
How do I know what the old coating is?
You often cannot know exactly, and the useful discovery is that it rarely matters. Cut a small notch through to the concrete with a sharp knife and look at the edge: it shows you how many layers there are, how thick they are, and whether the failure plane is between coats or at the slab.
What actually decides the method is not the chemistry, it is the behavior. Does it shatter, does it smear, or is it simply stuck fast? A ten minute test patch answers that for any coating, named or not, and it answers it for this floor rather than for floors in general.
Why does my grinder just smear the coating and gum up?
Because the coating is soft, thick or flexible, and abrasive tooling is the wrong instrument for it. Instead of being cut, it deforms, packs into the gaps between segments and blinds them, and the machine stops cutting entirely while looking like it is working.
The answer is tooling that shears rather than abrades, which usually means polycrystalline diamond, or a scarifier where the build is heavy enough to plane off. Shot blasting fails on the same material for the same reason: the abrasive drives into a soft film rather than fracturing it, and the machine loads up.
Is chemical stripping worth it?
On small areas and detail work, sometimes. Around fixed equipment, in a stair nosing, or on a floor a machine cannot be brought to, a stripper may be the only route. On open floor it is slow, it is messy, and the waste is a disposal question rather than a sweeping one.
The two failures worth knowing about are residue and water. Softened coating and stripper residue have to be removed completely, because anything left behind is a bond breaker under the new floor. And a stripped and rinsed slab has taken on water, which puts you behind the moisture gate before you can coat.
Could the old paint contain lead, or the adhesive under it contain asbestos?
Both are possible on an older building, and both are questions to settle before a machine touches the floor rather than after. Lead was not banned from residential paint in the United States until 1978, and lead containing coatings were used in industrial settings as well. Some pre 1980s resilient flooring adhesives, particularly black cutback mastic, contain asbestos.
If either is plausible from the age and history of the building, have it tested. A positive result does not merely add a step, it changes what the job legally is, who can do it, and how the waste leaves site. Grinding it first and asking afterwards turns a survey question into a contamination event.
Do I still have to grind after the coating is off?
Yes, and budgeting as though you do not is the most common way a removal job loses money. Removal tooling is built to take a coating off, not to leave a surface a new coating can bond to. What it leaves is uneven, often gouged, frequently still carrying residue in the low spots, and nowhere near the profile the next product specifies.
So removal and profiling are two operations with two rates and two lines on the quote. The bare slab you are standing on at the end of the removal still needs the same passes any bare slab needs.