[cap_it letter=’A’]s cryptocurrency values tank, anyone interested in deploying GPUs, as most commonly found in graphics cards, will be able to grab a bargain, with prices now hovering around MSRP levels.
Outside of the world of gaming, GPUs are most usually to be found in cryptocurrency mining rigs of specific flavors (see below) but are used increasingly in arrays powering AI algorithms.
Towards the end of 2017 and the first month or two of 2018, the tech and mainstream press was full of news and features about cryptocurrency: LiteCoin, Bitcoin, Ripple, Ethereum, and the like.
The fall in values has quietened interest in the currencies, although other uses of encrypted public ledgers (blockchains) continue to remain significant points of interest for the business press.
The decline in the price of cryptocurrencies has been likened by some as, if not the crytpocurrency bubble bursting, then at least the gas leaking out pretty darn quickly.
Tech-oriented investors have taken a severe kicking, as prices dropped more quickly (up to 15 times more quickly, according to Morgan Stanley) than the value of stocks after the dot.com bubble burst of 2000.
Whether it is in the world of finance, healthcare, or transportation, #deeplearning has an incredibly wide spectrum of applications to make the world a better place. Read more about some of the most commonly used deep learning phrases and terminology: https://t.co/H83SEzWYcl pic.twitter.com/aQIg0q9OhU
— NVIDIA AI (@NVIDIAAI) July 5, 2018
Bitcoin, the most widely recognized non-fiat currency, has shed around 70 percent of its value, and some 800 different tokens (currency types) are currently worth less than a single cent, according to CNBC.
There’s even a dedicated site, deadcoins.com, that tracks ICO scandals, and coins that never made the grade.
It is difficult to pin down reasons down for a decline in value of any investment, of course, but cryptocurrencies’ decline has been linked to several factors: traditional profit-taking, well-publicised crypto exchange hacks, rafts of worthless ICOs, and a tightening of governmental attitudes to cryptocurrencies in influential areas of the globe, the US and China being but two.
One of the results of the lower currency values has been that crypto-mining has become significantly less lucrative in recent months.
The high investment in hardware and running costs (read, electricity) of a decent mining rig have always made profit margins razor thin, even at the time of inflated exchange rates.
Lower cryptocurrency values mean that mining is less attractive, and that means demand for mining hardware falls.
Depending on the currency mined, hardware tends to be either ASIC- or GPU-based, the latter required when, for instance, currency devisors prefer to construct a token’s infrastructure so only internal computer processors or GPUs can be used to chew hash, as it were.
With fewer GPUs purchased, prices of graphics cards have fallen from the inflated levels of the last year or so, meaning the PCI cards are no longer the eye-watering outlay they were at the height of the cryptocurrency bubble.
Downward pressure on prices is increased by rumors that Nvidia is about to release the next generation of its GeForce range. Plateauing semiconductor production costs and lower silicon prices have also helped.
Depressed (or normal) prices could mean a more significant use of GPUs in AI compute matrices; outside of gaming rigs and crypto-mining setups, artificial intelligence and its variants (machine learning, deep learning, cognitive computing, etc.) can benefit from the hardware’s high number-crunching vs. power drain ratio.
