Data Centers: The Pain Point of POWER

The multi-billion-dollar Data Communications market is a particular focus for POET. Data centers are enduring an excruciating pain point in terms of power. Energy Management costs are approaching 50-60% of the cost of operating a Data Center today. Each watt of heat that does not have to be rejected from the rack could be worth about $10 per month in operating costs for a data center. A single copper direct attach cable burns about 3W of power per end – so a total of 6W per cable. Lets take a single mega datacenter with 10000 servers and 100000 copper links. If you can save 5W of power per copper link used in this one Data Center, this could easily translate into 500,000W of saved energy or $50 million per year in saved operating expense. And that is for a single mega data center.

POET Technologies - POET Platform

As a 6 year Sn POET investor I find it troubling and irresponsible that this misleading (if and could) information of $50 million per year in saved operating expenses is still posted on POET's website.
It should read: If you can save 5W of power per copper link used in this one Data Center, this could easily translate into 500,000W of saved energy per hour or  $500,000 thousand per year in saved operating expense. And that is for a single mega data center.
In Poet's (if and could) example the $50 million per year in saved operating expenses is inflated 100 times.                      

Do the math.

10,000 servers multiplied by 10 links per server = 100,000 links per data center.
100,000 links multiplied by 5W = 500,000 W of power saved per HOUR.
500,000W divided by 1000 ( one kilowatt per hour ) = 500 kWh.
500 kWh multiplied by 24 hours = 12,000 kWh per day.
12,000 kWh multiplied by 365 days per year = 4,380,000 kWh saved per year.
Factor in the % of cooling costs of a Data Center that would apply to the cables. Negligible.
Multiply  4,380,000 kWh saved per year by the average Industrial or Commercial (take your pick) price per kWh in the USA.

EIA - Electricity Data

Poet's (if and could) example of saving 5W out of 6W per copper link, which translates to an impressive 83.3 % savings over copper links is nothing but conjecture at this time.
They can't even be bothered to simply correct the absurd and misleading information in their website.
This shareholder is enduring an excruciating pain point in terms of granularity.
Sound familiar?

Drunkenlogger