The phosphate anion originates from multiple sources, and is responsible for some of the high anion gap metabolic acidosis associated with uraemia.
The inorganic phosphate in your extracellular fluid is but a pale reflection of the total body phosphate. Intracellularly, phosphate is the most numerous anion, available predominantly in the form of inorganic complexes (eg. in phospholipids, phosphoproteins and as a part of ATP). But this is all discussed elsewhere. Suffice to say that phosphate prefers to remain inside cells, and extracellular phosphate is merely in transit, on its way to a cell somewhere.
This transient pool of extracellular inorganic phosphate originates from multiple sources. Much of it is a byproduct of the metabolism of bone hydroxyapatite, but a large proportion must also surely come from signalling and catabolism. Phosphate is in everything and is released during routine cell maintenance, eg. where phospholipids are being degraded, during cell death, or in the course of normal function (for example where phosphate-rich ADP is used as a signalling molecule in clotting, vascular reactivity, etc).
Whatever phosphate is not bound to protein gets filtered freely by the glomerulus, and then most of is reabsorbed in the proximal (70%) and distal(10-20%) tubules. The resorption occurs with the help of a sodium-phosphate cotransporter (one of the various sodium-organic anion contransporters), and it is driven by the sodium gradient.
Obviously, if your level of phosphate increases, negative feedback mechanisms adjust the rate of its excretion (by suppressing its resorption). When the GFR decreases to around 20% of normal, phosphate reabsorption is thought to be maximally suppressed. In this situation, virtually no resorption takes place, and all of the filtered phosphate escapes with the urine. Sadly, at this stage glomerular filtration is so poor that even with maximally suppressed resorption the rate of phosphate loss cannot keep up with phosphate production, and hyperphosphataemia develops.
On average, it seems there is a daily oral intake of about 40mmol of phosphate, and there is an additional daily generation of about 5mmol in the process of metabolism of phospholipids and phosphate-contaning proteins. Of the total daily phosphate loss, about 30mmol is excreted renally, and 15mmol is lost via stool.
Imagine the CRF patient with zero urine output. Imagine them eating a nice meaty phosphate-loaded steak. If the daily consumption of phosphate remains the same, at the end of the day the patient ends up with an additional 20-30mmol of phosphate on board. Considering what we know about the distribution of exogenous phosphate (i.e. with over half of it staying extracellular), one may expect the serum phosphate levels to rise by 1.0-1.5mmol per day in anuric dialysis dependent patients. Renal physicians attempt to counteract this by limiting the bioavailability of oral phosphate (for example, by feeding their patients rare-earth metal phosphate binders.)