The result you're seeing is actually correct behavior, but there's a subtle misunderstanding about when Java uses infinity for floating-point overflow.
The key issue is that the addition didn't actually cause an overflow to infinity. Here's what's happening:
Float.MAX_VALUE is approximately 3.4028235E38
1000000000f is 1.0E9
The ratio between these numbers is huge: Float.MAX_VALUE is about 3.4 × 10²⁹ times larger than 1.0E9
When you add a very small number to a very large number in floating-point arithmetic, the result often equals the larger number due to precision limitations. The number 1.0E9 is so small compared to Float.MAX_VALUE that it gets "lost" in the addition.
When Does Infinity Actually Occur?
Infinity occurs when the mathematical result of an operation exceeds what can be represented as a finite floating-point number. Try this example:
float a = Float.MAX_VALUE;
float b = Float.MAX_VALUE;
float sum = a + b; // This WILL be infinity
or
float a = Float.MAX_VALUE;
float b = Float.MAX_VALUE * 0.1f; // Still very large
float sum = a + b; // This will likely be infinity
The Technical Explanation
Float.MAX_VALUE (approximately 3.4028235E38) is already very close to the maximum representable finite value. Adding 1.0E9 to it doesn't push the result beyond the representable range—it just gets rounded back to Float.MAX_VALUE due to the limited precision of 32-bit floating-point representation.
The IEEE 754 standard only produces infinity when the mathematical result genuinely exceeds the representable range, not when precision loss occurs during normal arithmetic operations.