My point is that it doesn't. A month or two ago, a couple of AM blades might have been 1% of the network, yeah? Now you'd need many, many blades to get 1% of the network.
Do you think that the manufacturing cost is the same? (It isn't). Do you think that the electricity cost is the same? (It isn't.)
As competition ramps up - i.e., the hashrate - the marginal profit per % hashrate will drop precipitously. AM will feel that, both in terms of mining profit and hardware profit (people will simply pay less per TH, which is how manufacturing cost is determined).
You seem to be ignoring the biggest factor: AM also develops their own mining equipment. Your argument works for regular mining bonds but for companies with engineers and R&D teams the whole paradigm shifts. Yes, cost per TH decreases but they release a device that can hash a higher volume of TH. on top of this the whole market knows that AM can deliver! The logistical side of orders, shipping and receiving is already operational.
The only question is will AM continue to outpace the competition in the hardware war and what many of us are thinking is, "Yes." Mainly due to not seeing any formidable competition at the moment, ( though there are some forming on the horizon). The smart bet in regards to mining companies is with AM, I see not compelling evidence otherwise.
In light of this, your position is actually the more risky bet. You are betting that a company will appear that has the hardware, personnel, and infrastructure to do in three months what AM has already been doing and doing well.
I agree I hold the more risky position indeed. That is why I pay about 10% of the strike for my options. I do not think my position is THAT much riskier.
Long-term I have little doubt things will play out as I describe... I am in fact an engineer in the semiconductor industry, and it is my strong opinion the performance increase from ASIC gen 2 will pale in comparison to the performance increase coming off GPUs.
Disclaimer: I'm wrong lots.
.