South Korea's stock of semiconductor chips dropped more than it has since 2014. This big decrease shows that customers are buying chips faster than companies can make them, as they need more equipment for developing artificial intelligence (AI) technology. Official data released on May 31 revealed that in April, chip inventories fell by 33.7% compared to a year earlier - the largest drop since late 2014. This is the fourth month in a row that inventories have declined, while at the same time South Korea's exports of semiconductors have gone up again. Additionally, South Korea's production of chips rose 22.3% in April, which is less than the 30.2% increase from the previous month. Shipments from factories grew 18.6%, also lower than March's 16.4% growth.

South Korea is home to the two biggest memory chipmakers in the world (Samsung and SK Hynix), and they are competing to supply chips to NVIDIA, the latest having an insatiable appetite for more and more chips. These two Korean companies are in a race to develop a more advanced and more profitable version of high-bandwidth memory, or HBM. During the memory chip boom from 2013-2015, inventories didn't increase for about a year and a half. In the 2016-2017 cycle, inventory declines lasted nearly a year. A report from South Korea's central bank expects the latest surge in chip demand to continue at least until the first half of next year. This is because the "artificial intelligence boom" is driving up demand similarly to how cloud servers caused an expansion in 2016, and now mostly forgotten crypto-mining fever. South Korea will release its latest export data on June 1.

It appears that some of the current chip shortages might be artificially induced by one or multiple companies in the chip supply chain, according to an article by TIME Magazine. The article is taking a look at the role TSMC is playing in the global chip production industry and TIME has interviewed TSMC chairman Mark Liu among others in the industry.

Mark Liu is quoted as saying "But I told them, "You are my customer's customer's customer. How could I [prioritize others] and not give you chips?"" when asked about the complaints by car makers, since they were among the first to suggest TSMC was one of the issues. Due to the various allegations against TSMC, Liu had a team collect data points to try and figure out what was going on and to see which customers were truly running low on stock and which customers that might be stockpiling for a rainy day.

NVIDIA's add-in card partners are ordering fresh stocks of GeForce RTX 2070 graphics chips even as the performance-segment of the GPU market has changed with AMD's introduction of its Radeon RX 5700 series "Navi," according to a Gamers Nexus report citing sources among NVIDIA partners. NVIDIA partners are expecting a slump in AMD's RX 5700 series graphics card inventories, particularly that of the RX 5700 XT, to create a price-point at which to sell the RTX 2070. NVIDIA partners expect RX 5700 XT inventories to run slim as supply of the 7 nm "Navi 10" chips from foundry-partner TSMC may not satiate the SKU's reportedly high demand.

NVIDIA's current product stack has the original RTX 2060 at $349, the RTX 2060 Super at $399, and the RTX 2070 Super at $499. The RTX 2070, which is outperformed by the $399 Radeon RX 5700 XT, was practically phased out from NVIDIA's product-stack as it was succeeded by the RTX 2070 Super at its $499 price-point. With the RTX 2070 making a comeback, it would be interesting to see what its price-point will be. There is a gap between the $399 RTX 2060 Super, and the $499 RTX 2070 Super, although the performance gap between the RTX 2060 Super and the RTX 2070 is a paltry 4 percent, which is easily closed by moderately overclocking the RTX 2060 Super. As of this writing, both pricing and availability of the RX 5700 XT appear normal.

According to DigiTimes, NVIDIA and AMD partners are doing their best to digest unsold graphics card inventory via promotions and discounts. The idea here is that they can achieve increased amounts of revenue and move a lot of the graphics card stock they accumulated following (and counting on) the crypto craze. This move will certainly affect their bottom line when it comes to profits, but that's just what these companies have to do. Hardware sold at a tiny profit is always better than that which stays in the warehouse simply deprecating, and these companies know it best.

DigiTimes cites the example of AMD-partner TUL corporation which manages the PowerColor brand, saying that they achieved, via promotions, an increase of 115% in revenues on January (over their December values). This increase in revenue still compares negatively YoY, where it's still 85.7% lower compared to January 2018. And despite the increase revenue, profits declined to the red: the company had net losses of NT$10.31 million in January 2019 and EPS of negative NT$0.31. Some hard times could be coming for AIB partners, who will have to bite the bullet on pricing to move their stockpiles of older generation graphics cards.

Even as we achieve consumerism in scales hitherto unseen, tech companies always want to sell more - there's "always" increased production, there must always be increased, projected demand from customers. However, when demand isn't there, and growth slows down or even stagnates, production takes its time to adjust - and already manufactured products have few opportunities other than going on towards a swelling inventory.

This is what is happening with a myriad of tech companies, such as Apple, Samsung, Xiaomi, Intel, Hon Hai (Foxconn), among others. We could even take a page from our own PC industry and look at NVIDIA's Pascal inventory that is in need of clearing up - and which has resulted in bottoming prices of previous-gen cards as we look towards the new RTX 20-series. Which, coincidentally, have been launched with increased pricing over the previous generation. Perhaps another way of moving old inventory?