How the two differ in kind
The difference is one of sequence, not hardware. A cast-in anchor is located before the concrete exists: it hangs in a template, the concrete is placed around it, and its position is whatever survived the pour. A post-installed anchor is located after the concrete is hard, off the thing it actually has to line up with, and then has to make a grip or a bond in a hole it cut for itself. Almost every practical consequence in this comparison is downstream of that one ordering decision.
What cast-in buys is the load path. A headed bolt bears on steel deep inside concrete that was cast tight around it — no drilled annulus, no dust on the bearing surface, no bond to lose — which takes pullout largely off its failure menu and leaves the concrete breakout cone as the thing to check. It is also generic: the basic breakout strength of a single headed cast-in anchor is a published expression in concrete strength and embedment, and the code hands cast-in anchors a higher coefficient in that expression than a post-installed anchor gets by default — with no brand attached to any of it. Post-installed anchors are products before they are details. Their capacity belongs to one manufacturer's evaluation report, at one qualified diameter, embedment, hole size and cleaning procedure, and none of it interpolates between the tabulated depths or transfers to the anchor a different supplier sends instead.
So the trade is about where the risk sits and who can see it. Cast-in concentrates everything into the hours before the pour, where it is a surveying and bracing problem, and where an error announces itself loudly at erection when a plate will not drop over the bolts. Post-installed spreads the risk into the installation itself, where it is a hole-cleaning, depth and torque problem, and where a bad job is often invisible for the life of the building. Neither risk is worse in the abstract. The honest question is which one your project is actually set up to manage — and on most jobs the answer is decided by whether you will be standing on the deck before the concrete arrives, not by anything in the capacity tables.
The factors that actually differ
Show
| Cast-in anchor bolts | Post-installed (drilled) anchors | |
|---|---|---|
| When the position is fixed | Before the concrete exists, off a drawing. The template has to be set out, braced and re-checked once placing starts, because a bolt group walks under the pressure of wet concrete and a vibrator. | After it is hard, off the thing the anchor has to line up with. Layout error largely stops being a category — you drill where the hole in the plate actually is, not where it was drawn. |
| How load reaches the concrete | Bearing. A head, nut or plate presses against concrete cast around it, so there is no annulus, no drill dust on the bearing surface and no bond that could have been compromised at install. | Grip or bond. A wedge, sleeve or undercut presses outward on the hole wall, or adhesive bonds along it. Both add pullout and bond failure to the menu, and both are decided by how the hole was drilled and cleaned. |
| Where the capacity number comes from | The code, generically. Concrete strength and effective embedment drive it, under the coefficient the code assigns cast-in anchors and a cracked-versus-uncracked adjustment on top of it — a detail any engineer can run without a catalogue open. | One product's evaluation report, read for far more than the capacity table: qualified embedments, minimum member thickness, edge and spacing minima, cracked-concrete and seismic qualification, permitted orientation. An anchor rated for the load and disqualified by the geometry is the common surprise. |
| Reinforcement and services in the way | Competes with the steel for space but cuts nothing. It can be tied to the reinforcement, which matters because hairpins and U-bars crossing the failure surface only exist if somebody ties them before the pour. | Has to cut past whatever is there. Scanning before drilling is not optional in a congested or post-tensioned member — a severed tendon is a structural incident, not a delay — and no anchor reinforcement can be retrofitted around a drilled hole. |
| Embedment you can afford | Nearly free. Going deeper costs a longer piece of rod, and member thickness is the only real limit. That matters more than it sounds, because breakout capacity climbs faster than embedment does. | Bought with drilling time. Depth costs the bit, the dust, the extra cleaning cycles and the report's minimum member thickness, so drilled designs live under quiet pressure to stay as shallow as the numbers allow. |
| Whether a mistake is visible afterwards | Yes, always. Position and consolidation are witnessable before the concrete covers them, and a template out of tolerance is discovered when the steel arrives — painful, but never silent. | Frequently not. Cleaning, drilled depth, torque and cure temperature all leave the same view from above — a nut on a plate. Nothing distinguishes the anchor that was done properly from the one that was not. |
| Fixing a mistake | Ugly and engineered. A group out of tolerance becomes slotted plate holes with plate washers, a re-detailed plate, or chipping out and post-installing supplementary anchors — decisions taken with a crane standing idle. | Routine. A hole that hits a bar or blows out at the surface is abandoned, filled and re-drilled a short distance away, subject to the spacing the report demands from the abandoned one. |
| Inspection and installer qualification | A pre-placement hold point: position, embedment and bracing checked while they can still be corrected, with the survey and the photographs as the permanent record. | Heavier exactly where it matters most. An adhesive anchor installed horizontally or upwardly inclined to carry sustained tension attracts both a certified installer and continuous special inspection under most adopted codes; other post-installed anchors are inspected periodically, against the printed installation instructions rather than the drawing. |
| Where the cost sits | Generic hardware — rod, nut, plate washer, a template — with the real cost in coordination on the pre-pour critical path. It scales with anchor GROUPS rather than anchors: a large bolt pattern is barely dearer to set out than a small one. | Per hole, and close to linear in the anchor count: proprietary hardware at a premium over plain rod, adhesive, scanning, and drilling labour that grows with depth and diameter. In exchange it is spent whenever access allows, in nobody else's pour window. |
Which one, and when
Choose cast-in anchor bolts when…
- The connection is meant to yield its steel rather than break out its concrete — a moment baseplate, a column base, a seismic hold-down. That design is bought with embedment, and embedment is cheap before the pour and dear after it.
- Anchor reinforcement is part of the answer. Hairpins and U-bars taking tension across the failure surface have to be tied in with the rest of the steel; nothing retrofits them into hardened concrete.
- The member is thin, congested or post-tensioned, and placing steel around the obstructions beats drilling through them and hoping the scan told the truth.
- You control the pour, or the pouring trade will accept a surveyed template and brace it properly. Cast-in's risk is manageable precisely when that sentence is true, and unmanageable when it is not.
Choose post-installed (drilled) anchors when…
- The concrete already exists. Retrofit, an existing slab, a wall somebody else poured — the comparison is over before it starts, and the work becomes getting the drilled installation right.
- The layout is not settled. Fabrication drawings arriving late, positions taken off steel already erected, a rail or sill that must follow what was built rather than what was drawn.
- Many small fixings rather than a few large ones — sills, rails, ledgers, pipe supports. Templating each of them before a pour is a survey exercise out of all proportion to what any one anchor carries.
- The pour window is not yours to book. If pre-pour access depends on another trade's programme, cast-in's cheap hardware is bought with a scheduling dependency you cannot enforce.
Now run your own numbers
This page holds no prices on purpose — a national average is wrong for almost every real project. Quantify both options with your dimensions and your local quotes.
Frequently asked questions
- Is cast-in simply stronger?
- Not mechanism for mechanism, and the shortcut is worth unpicking. At the same diameter and the same effective embedment in the same concrete, a properly installed undercut or adhesive anchor can match or beat a headed cast-in bolt, because breakout is the same failure surface in both cases and a product's evaluation report can qualify it for the coefficient a cast-in anchor gets by right. The code's own default is a lower coefficient for post-installed than for cast-in, so parity there is earned by the specific report rather than assumed. Two further things make cast-in stronger in practice rather than in principle. First, depth is nearly free before the pour and expensive after it, so cast-in details simply get specified deeper, and capacity rises with embedment faster than in proportion. Second, the failure menu is shorter: a headed bolt cast into concrete has no bond to lose and no hole that could have been cleaned badly.
- The cast-in bolts are out of position. What are my options?
- Survey the as-built positions before anyone touches anything, and send the survey to the engineer rather than making the plate fit on site. In ascending order of pain: slotting or oversizing the plate holes with plate washers, which changes how shear transfers and is the engineer's call rather than the fitter's; a re-detailed plate made to the as-built pattern; and abandoning bolts to post-install supplementary anchors, which then has to be checked as a mixed group. What is not on the list is heating and bending the bolts to reach. It damages the steel, and the bearing surface no longer sits where the design put it.
- Can one baseplate use both?
- Yes, and repairs routinely end up that way — but the result is a mixed group and it stops behaving like a uniform one. The two types have different load-slip stiffness, so they do not share load in the proportions a simple pattern check assumes, and the geometry has to be re-checked treating the whole set as one group with overlapping failure surfaces rather than as two independent patterns. Both of those are engineering decisions with a signature on them, not site arithmetic.
- Why is hole cleaning talked about so much?
- Because with an adhesive anchor the wall of the hole IS the load path, and the dust left on it behaves like a release agent between the adhesive and the concrete. The loss is not marginal, and nothing about the finished job reveals it. The cycle count and brush size in the printed installation instructions are not a suggested standard of care either — they are the conditions the anchor was qualified under, so departing from them abandons the number the design was based on. Two field realities do most of the damage. A brush that has worn down no longer scours the surface it was bought to scour, and it gives no sign of having stopped. And a hole cut with a diamond core barrel comes out smooth rather than rough, which suppresses bond so severely that many adhesives are not permitted in one at all — so a cored hole needs the report checked rather than being treated as a tidier version of a drilled one.
- Can I push anchors into wet concrete instead of setting a template?
- No, and it is the tempting middle ground that delivers neither side's advantage. Wet-setting — pushing and wiggling an anchor into freshly placed concrete — is excluded outright by most product approvals, because it leaves a void down one side of the shank, the concrete is never properly consolidated against the bearing surface, and nothing about the resulting embedment was tested. It also fails on its own terms: the position ends up no better than a template would have given, since nothing holds the anchor while the slab is screeded and finished.
- Which works out cheaper?
- There is no honest single figure and this site will not invent one, but the shapes are clear enough to plan against. Cast-in hardware is generic and cheap, and nearly all of its cost is coordination — setting out, bracing, and a re-check after the vibrator has been through. That cost attaches to anchor groups rather than anchors, so a large pattern is barely dearer to set than a small one, and it lands on the pre-pour critical path, where delay is at its most expensive. Post-installed cost is per hole and close to linear in the count: proprietary hardware at a premium over plain rod, adhesive, scanning and drilling labour that grows with depth and diameter, spent whenever access allows. The decisive asymmetry is not the install at all — it is the cost of being wrong. A misplaced cast-in group is repaired by engineers with a crane waiting; a spoiled drilled hole is abandoned and re-drilled a hand's width away.