You'll need a minimum of 6 layers to fan-out most modern BGAs, and if you've got other peripherals or a lot of power domains, you may need a lot more (the i.MX6 REX project is done on 12 layers, for example). If your SoC is smaller than 0.8mm pitch, you'll need via-in-pad technology, laser-drilled microvias, and/or 3mil traces.
Manufacturing and assembling prototype quantities (10 or so) of a 6- or 8-layer board with a custom stack-up, plus assembly and BOM could cost $10k or more, depending on board complexity, turn-around time, and if you're going to China or not. Considering you'll probably need to run two or three prototypes to get the design nailed, you're talking about a $30k investment in materials.
Designing and routing a DDR memory bus by someone new to high-speed routing could take 50 hours or more of engineering time. Budgeting $100/hr, means you're at $5000 in personnel costs, just for the memory bus.
And you're quibbling over the price of an EDA tool that costs less than $10k? Just buy Altium and do it the right way.