The price of mining a bitcoin tends to approach the price of a bitcoin, so the question is rather is 21's approach equal to or better than large scale mining.
"Tends to approach" is vague.
In the short term, the marginal cost of mining almost exactly equals the price of a Bitcoin (excluding, possibly, some short transition periods when the price of a Bitcoin is just so high and/or increasing so fast that it is hard to catch it), but if you are below marginal cost, then you are more efficient than the marginal miner and you make a operating profit, which may or may not cover your investment. In the long term, that's the time horizon when investments pay off, or don't. Nothing lasts forever.
No, tend to approach is the correct term. There cost will never exactly reflect the price, due to speculative errors from the miners when they decide to expand or contract, therefore the cost will be different also between miners. So again, can the 21's plan for mining compete with large scale mining? Small scale mining can be and is continually tested, what they bring to the table is a new chip. So can the new chip revive small scale mining?
Edit: You can never know in advance, but I would guess that if the chip is good in a random small scale device, it is even better if you pack a few hundred in a dedicated device with proportionally designed power supply, fans the rest of the stuff that is needed. Is it possible to construct a chip that is better than current mining, and at the same time can not be used large scale?
You can not rely on persons with experience in funding. Stranger projects have been started, a company with million dollar donut cars, selling for 100K per car and losing 900K per car comes to mind.