Your custard curdled because the egg proteins in it bonded to each other faster than the sugar, the dairy and, in a pastry cream, the starch could keep them apart. Once those proteins link too tightly they stop holding water and squeeze it out. That is the lumps in a thin pool you are looking at.
The heat was too high, or too uneven, or it went on too long. Which of the three depends on which custard you were making, because a stirred custard sauce and a pastry cream fail at genuinely different points, and the advice for one will ruin the other.
One boundary first. This page explains the texture failure, not whether what you cooked reached a safe temperature. That is a separate question and the authority on it is the USDA Food Safety and Inspection Service. Every temperature below is a texture threshold from a culinary testing source, and none of them is a safe-cooking figure.
Allergens. Custard is an egg and dairy preparation, and both are on the FDA's major food allergen list. This page gives no substitutions, because a substitution that is wrong on an allergen is a safety problem rather than a texture one.
What Curdling Actually Is
Egg proteins sit coiled up. Heat makes them move, they knock into each other, and the weak bonds holding those coils closed break. The unfolded proteins then bond to each other and build a mesh that traps liquid. That mesh is what a custard is.
Keep heating and the same process goes one step too far. America's Test Kitchen puts it plainly: "If they are overheated, too many bonds form and the proteins clump" (Fixing Broken Custard, read 2026-08-24). The mesh contracts, wrings out the liquid it was holding, and a sauce becomes solids and whey.
This is the same failure that makes a meringue bead, which is why a weeping meringue and a curdled custard look unrelated and are not. In both cases an egg protein network has tightened past the point where it can hold its water.
Bare egg protein sets low. The American Egg Board gives egg white as coagulating between 144°F and 149°F (62.2°C and 65°C) and egg yolk between 149°F and 158°F (65°C and 70°C), read 2026-08-24. A custard is nowhere near that fragile, and the reason is the three things you mixed into it.
Three Things Are Holding the Proteins Apart
Dilution. Milk and cream put distance between the egg proteins. Proteins further apart take more energy to find each other, so the mixture sets higher than the yolk alone would.
Sugar. The American Egg Board lists sugar among the factors that change coagulation, specifically as something that "raises the temperature for coagulation," while acids do the reverse and lower it. A sweet custard tolerates more heat than an unsweetened one, and a custard with citrus juice or fruit puree in it tolerates less.
Starch, and only in some custards. Cornstarch or flour in a pastry cream does something the other two do not. America's Test Kitchen states that "the starch interferes with the cross-linking of egg proteins, thus preventing coagulation" (Perfecting Pastry Cream, read 2026-08-24). It physically gets in the way of the mesh forming.
Take one of those away and the margin narrows. That is the whole story of this failure.
What Tempering Actually Does, and What It Does Not
Every page on this subject tells you to temper the eggs: dribble the hot milk into the yolks before returning everything to the pan. The instruction is fine. The reason usually given for it is wrong, and that changes what you should do when you are in a hurry.
The usual explanation is that you are warming the yolks gently so they are not shocked. America's Test Kitchen tested that and concluded otherwise: "Tempering doesn't work because of the gradual heating of the yolks; it works because the addition of liquid dilutes the uncooked egg proteins, making it harder for them to link up and form firm clumps when heated" (Do You Need to Temper Eggs?, read 2026-08-24). In their test the batch whose yolks were warmed separately before meeting the liquid showed signs of curdling, while the batch where everything went into the pot at once and was stirred constantly came out identical to the control.
So the rule is not "warm the eggs." It is get liquid into the eggs before heat gets into them, and never stop stirring. Pouring a whole pan of scalded milk onto three yolks in one go fails not because it is hot but because for a second those yolks are still concentrated.
Why a Custard Sauce and a Pastry Cream Fail at Different Points
This is where most of these disasters come from. The reader applies the rule from one custard to the other.
A stirred custard sauce, a crème anglaise, has no starch in it. Yolks, dairy, sugar, nothing else. Dilution and sugar are the only protection those proteins have, and the window is narrow. America's Test Kitchen takes crème anglaise off the heat "when the mixture registers 175 to 180 degrees" and notes that "egg-based puddings and custards can curdle if cooked beyond 185 degrees" (Knowing When Puddings and Custards Are Done, read 2026-08-24). Those are texture numbers, not safety numbers. A few degrees of inattention is the entire margin.
A pastry cream has starch, and it is not only allowed to boil, it has to. There is a second reason for that beyond swelling the starch, and almost nobody mentions it. Egg yolks contain amylase, an enzyme that digests starch. America's Test Kitchen states that "pastry cream must be heated to a temperature high enough to destroy the amylase enzyme present in egg yolks, which would otherwise break down the starch and make the pastry cream runny," and gives the finished target as 200 degrees with "three or four bubbles" bursting at the surface.
So the two fail in opposite directions. Cook a crème anglaise like a pastry cream and it scrambles. Cook a pastry cream like a crème anglaise and it will not scramble at all, it will go slack in the refrigerator hours later while the enzyme you never destroyed eats the starch you were relying on. If your pastry cream was thick in the pan and soupy the next morning, that is what happened, and more cornstarch next time will not fix it.
The Ordinary Causes Underneath All of That
The pan. A thin base makes hot spots, and the custard touching the metal is always ahead of the custard in the middle. Heavy base, moderate heat, and a flat-edged spatula that reaches the corners a whisk does not.
Stirring that stops. Constant means constant. The film sitting still against the base scrambles first, and once those lumps exist they will not dissolve back in. There is also no reward for speed: a custard that sets fast has set unevenly.
Acid. Lemon, passion fruit and other acidic additions lower the temperature at which the proteins set, per the American Egg Board note above. Add them off the heat where the recipe allows it.
Leaving it in the hot pan. A pan keeps cooking what is in it after the burner is off. Get the custard out and into a cold bowl as soon as it is done. How long you then keep it, and at what temperature, is a food safety question and is not answered here.
Can You Rescue It
Sometimes, and only when the lumps are small.
America's Test Kitchen's tested rescue is an immersion blender: "A quick buzz effectively broke down the clumps, restoring a perfectly creamy texture," pulsing in five-second intervals for 15 to 45 seconds depending on lump size, then straining through fine mesh. They warn against a countertop blender or food processor, which whip in too much air and leave you with froth.
If it has gone fully to scrambled egg, blending gives you smooth scrambled egg. Nothing brings back a protein network that has completely contracted.
Two things this rescue does not do. It does not make a curdled custard safe or unsafe, because texture is not a safety signal in either direction. And it is not the fix for a grainy custard, which is usually undissolved sugar rather than broken protein, the same distinction that separates a split from a gritty buttercream.
What This Is Not
Not a split buttercream. That is a broken emulsion of fat and water, it happens cold as often as hot, and it usually comes back with temperature and beating. See why buttercream splits.
Not a deflated egg foam. A meringue that collapsed lost its air, a mechanical failure rather than an over-set one, covered in why a meringue deflates.
Frequently Asked Questions
Why can pastry cream boil when crème anglaise cannot? Because pastry cream contains starch, and starch physically interferes with the egg proteins cross-linking. Crème anglaise has only dilution and sugar protecting its proteins, so it sets and then over-sets within a few degrees.
Does tempering actually stop the eggs curdling? It helps, but not for the reason usually given. America's Test Kitchen's testing found the benefit comes from diluting the concentrated egg proteins, not from warming them gradually. Constant stirring from the very start does the same job.
My pastry cream was perfect in the pan and runny the next day. What happened? It almost certainly never reached a proper boil. Egg yolks contain amylase, which breaks down starch and is destroyed only at high enough heat. The starch was digested after the custard came off the heat.
Is a curdled custard safe to eat? Curdling is a texture failure and tells you nothing either way about safety. Whether an egg custard was cooked to a safe temperature is a separate question and the authority for it is the USDA Food Safety and Inspection Service. A smooth custard is not evidence of anything, and neither is a curdled one.
The Short Version
- Curdling is egg protein bonding too tightly and squeezing out the water it was holding.
- Sugar and dairy raise the temperature at which that happens. Acid lowers it. Starch blocks it.
- Tempering works by dilution, not by gentle warming. Stir constantly from the first second.
- Crème anglaise has a window of a few degrees. Pastry cream has to boil, or the amylase in the yolks thins it later.
- None of this is a food safety answer. That question belongs to USDA FSIS and to its own page.