Based on a report by Susquehanna, the lead time for what can be called ancillary semiconductors, i.e. the kind that are paired up with processor, SoCs and GPUs, now have a lead time of over six months. This is based on data from the distribution channel and as such, it's unlikely to affect large companies that have long-term contracts with their suppliers. However, smaller companies, or semiconductor manufacturers that solely rely on the distribution channel for sales of the parts, are not in a good place right now. Obviously this doesn't affect everything equally and many parts are also in stock with the big distributors, but sometimes at inflated prices compared to a couple of years ago.

Susquehanna is pointing towards several reasons for the increase in lead times, although the big ones include the Russia's war on Ukraine, an earthquake in Japan (on the 16th of March) as well as the more recent lockdowns in several cities in the PRC which are key to the production of everything from semiconductors to finished consumer goods. At the latter half of 2020, the lead time was less than 14 weeks, but has since then increased to almost 27 weeks and it looks like it's likely to continue to increase for the time being. The worst hit components are said to be analogue chips, so things like signal amplifiers, power control oscillators and the like, which had the lead times increased by 18 over the space of a month. The report says that Broadcom has increased its shipping times by as much as 30 weeks and its backlog of orders just keeps growing. This is quite surprising, as earlier reports mentioned that the WiFi makers were reasonably unaffected by the increase in lead times, but Broadcom does make a wealth of other products too, so this could be unrelated. On a more positive note, it seems like the lead time for passive components has improved by a couple of days, with an average lead time of 25 weeks.

Crypto is going nuclear in the not so distant future, as US company Talen Energy revealed plans to construct a cryptomining data center in the immediate vicinity of the Susquehanna Steam Electric Station in Pennsylvania. Initial power consumption of the installation is expected to settle around 164 MW, with that figure climbing up to 300 MW once the infrastructure is complete - powered by dual 1+GW nuclear units and two independent substations. Talen Energy estimates maximum on-site power in the order of 1 GW.


The idea is to begin a narrative reversal around the environmental cost of cryptocurrency mining - if power is provided by cleaner technology, cryptocurrency mining's carbon footprint is bound to be reduced, in turn increasing attractiveness for environmentally-conscious businesses. Elon Musk, for one, made waves in both news outlets and cryptocurrency markets (and value) when he announced the decision for Tesla to cancel acceptance of Bitcoin as payment for its vehicles, citing environmental concerns regarding power sources keeping Bitcoin's blockchain secure. Another company, Energy Harbor Corp, announced a five-year collaboration with Standard Power to power its Bitcoin blockchain mining center in Coshocton, Ohio, starting December this year.

Industry specialists with various analysis groups have stated that they expect the world's current chip supply shortages to not only fail to be mitigated in the first half of 2021, but that they might actually last well into 2022. It's not just a matter of existing chip supply being diverted by scalpers, miners, or other secondary-market funnels; it's a matter of fundamental lack of resources and production capacity to meet demand throughout various quadrants of the semiconductor industry. With the increased demand due to COVID-19 and the overall increasingly complex design of modern chips - and increased abundance of individual chips within the same products - foundries aren't being able to scale their capacity to meet growing demand.

As we know, the timeframe between start and finish of a given semiconductor chip can sometimes take months. And foundries have had to extend their lead times (the time between a client placing an order and that order being fulfilled) already. This happens as a way to better plan out their capacity allocation, and due to the increased complexity of installing, testing, and putting to production increasingly complex chip designs and fabrication technologies. And analysts with J.P. Morgan and Susquehanna that are in touch with the pulse of the semiconductor industry say that current demand levels are 10% to 30% higher than those that can be satisfied by the fabrication and supply subsystems for fulfilling that demand.