The cost of (buying) DNA varies dramatically depending on the length and the purity of a sample, largely because DNA samples of different length are produced by different methods. The cost purification can also add up to more than that of the production. The Nature article specifically mentions common assumptions about the requirements for DNA components that can be used in programmed DNA self-assembly – essentially they have demonstrated that this can be done a lot cheaper than people have assumed.
The short "staples" used in the origami project are synthetic single-stranded DNA (also called oligonucleotides or oligos). These can be ordered on-line, and the price is typically described as 1$/base. Each oligo is synthesized by adding one base at a time and each time there is about 1% chance of an error – the exact number will depend on the process, but one vendor (IDT) explains that for a 27-mer about 30% of the sample will contain mistakes. Purifying the sample to keep only the correct 70% of the oligos can make it 2-3 times more expensive, hence the article specifically mentions that the proposed method works with cheap unpurified staples. But still, a hundred of different staples each about 30 bases long will be about 3000$.
The same method clearly can not be used to produce the origami template (scaffold), which is 1000's of bases long, so they use a natural sequence that can be multiplied by cloning – a relatively cheap process, although with extraction and purification the costs add up – to 80% of the total cost according to the article.
Adding up the above estimates, my educated guess is that the total cost of materials in this work adds up to about 10k $ quite easily – custom synthesized and purified DNA is still more expensive than gold on a per gram basis, but then again, so are most nanomaterials. The saving grace is that in some cases you need very little of them to improve or make a product, but until we find a way to make custom-designed nanomaterials on an industrial scale, none of this stuff will be affordable for projects outside of research labs.