Nvidia's winter dilemma: New graphics card sales slow amid cryptocurrency bear market

On February 15, 2019, Beijing time, NVIDIA (NASDAQ: NVDA), an internationally renowned graphics chip company and artificial intelligence company, released its Q4 2018 financial report. NVIDIA's revenue for the quarter was US$2.205 billion, a year-on-year decrease of 24%.

Artificial intelligence analyst Wang Shuyi told a reporter from China Business News that the collapse of digital cryptocurrency valuations has affected the mining machine market, which has had a significant negative impact on Nvidia. The market has deteriorated since the third quarter of 2018, leading to very negative industry expectations for 2019. Cautious expectations have affected orders, which has dragged down Nvidia's market performance.

At the same time, the reporter found that since the second half of last year, a large number of second-hand graphics cards have appeared on the market. The main customer groups of these graphics cards are gamers or computer assembly enthusiasts, but it is worth noting that many second-hand graphics cards that were once used for mining have also flowed into the market and are disguised as ordinary second-hand graphics cards for sale. The recycling and sale of cheap "mining cards" has also affected the sales of new graphics cards.

However, in the view of Yang Xiao, the founder of Panda Miner, Nvidia still has room for development in graphics card mining machines. As Nvidia founder and CEO Huang Renxun said: "Q4 was an unusual, extremely turbulent and disappointing quarter. But looking forward, we are confident in the company's strategy and growth momentum."

The reporter asked Nvidia China's public relations manager about relevant issues, but no response was received as of press time.

Chain reaction of "mine disaster"

Wang Qi (pseudonym) is a salesperson at a digital product store. He recently posted a post on Xianyu to buy back various second-hand "mining cards" at a price of 111 yuan. He told reporters that he "buys and sells" mining cards, buying them at low prices through online platforms such as Xianyu, and then reselling them. "I usually have hundreds of cards on hand, but I have collected fewer in the past few days," he said.

The so-called mining card refers to the graphics card installed in the graphics card mining machine. Because the graphics card mining machine is basically in a 24-hour working state when it is turned on, its performance status has been questioned. Many consumers who trade second-hand graphics cards online will mark that they do not want mining cards or non-mining cards.

The mainstream types of mining machines include graphics card mining machines and chip mining machines (i.e. ASIC mining machines). The type of mining machine used by miners mainly depends on the type of virtual currency they mine. Most miners use chip mining machines to mine Bitcoin and graphics card mining machines to mine Ethereum.

There are many resellers like Wang Qi. On February 13, 2019, the reporter went to a digital product store in Zhongguancun, Beijing to buy a second-hand graphics card. A salesperson recommended a Gigabyte GTX 1070 graphics card to the reporter, but after the reporter questioned whether it was a mining card, the other party stopped selling this product. Another salesperson privately told the reporter that 90% of the second-hand GTX 10 series cards (1050, 1060, 1070) are mining cards. "Now the graphics cards in the hands of miners are piled up like mountains." He said.

An article on the website of the American magazine Forbes stated that second-hand mining cards are one of the factors affecting the sales of Nvidia graphics cards.

GTX is a series of products launched by NVIDIA and was launched in October 2016. Driven by the bull market of digital currencies in 2017, a mining boom was set off, and the prices of mining machines and even the core components of mining machines also rose sharply. A gamer posted on the forum of NVIDIA's official website on January 20, 2018, "If the graphics card prices continue to rise like this, I'm afraid there will be no more computer gamers that you (NVIDIA) care about." The player said in the post that the price of GTX 1070 was nearly $1,000 at the time, and the player thought it was worth at most $450.

The prosperity of graphics cards and mining machines continued until the first half of 2018. As the digital currency market entered a bear market, Nvidia's stock price continued to fall after reaching its highest point of $289 per share on October 1, 2018. The lowest point so far was on January 3 this year, when the stock price was $127 per share, a drop of more than 50%.

In a shareholder letter, Nvidia lowered its revenue target for Q4 2018 and explained the reason as a sharp drop in demand for digital currencies and inventory backlogs from Pascal Mid-Range GPUs. Nvidia is digesting the inventory by postponing the production cycles of several planned new products, which are expected to be digested between February and April 2019.

Zhang Meng (pseudonym), a staff member of a mining farm in Sichuan, told reporters that the funds of his mining farm increased dozens of times in 2017. However, since the last batch of goods was delivered in May and June last year, his mining farm has stopped purchasing new graphics cards.

Yang Xiao said that inventory backlog is the current situation in the industry, but due to the sluggish market of digital currencies, the current sales of new graphics card mining machines are still relatively sluggish.

It is an indisputable fact that the “mining disaster” has had an impact on Nvidia, but the question is, how big is the impact?

Yang Xiao told reporters that AMD (hereinafter referred to as A card) currently has a larger share in the graphics card mining machine market, because A card has a better affinity for the algorithm of mining Ethereum, so the existing graphics card mining machines in the market are dominated by A cards. "Currently, the graphics card models corresponding to the graphics card mining machines in circulation in the market are mainly A card 470, 570, 580, N card (ie NVIDIA) P1060, but with the upgrading of the requirements of small currencies for graphics cards, the demand for high-end graphics cards such as P1080, P1080t, vega64 is gradually increasing." He said.

"However, for other small currencies mined with graphics cards, N cards have a higher cost-effectiveness than A cards due to their graphics card affinity and healthier and richer underlying language ecosystem. With the recent popularity of mining small currencies such as Grin and Beam, the market demand for N cards is increasing." Yang Xiao said.

"Artificial Intelligence" Cools Down

On February 6, 2019, at SoftBank's earnings briefing, Masayoshi Son reviewed the situation of Nvidia, an internationally renowned chip giant invested by SoftBank's Vision Fund in 2017, and how its stock price rose rapidly in 2016 and began to plummet in the second half of 2018. He finally stated that he had withdrawn his investment in Nvidia.

Nvidia lowered its revenue target in its Q4 2018 performance guidance, but the earnings report of $2.205 billion was slightly higher than the target of $2.2 billion. Its stock price rose instead of falling on February 15, eventually rising slightly by 1.08%.

After Nvidia released its third-quarter earnings report on November 16, 2018, its stock price plummeted by 18.76%. On January 28, 2019, Nvidia lowered its Q4 earnings forecast for fiscal year 2018 from the previous forecast of $2.7 billion to $2.2 billion, and Nvidia's stock price fell by nearly 14% that day.

Wang Shuyi said that from the perspective of stock prices alone, the current decline is normal. Although the decline from the 2018 high point is 50%, if we look at it over the past 10 years, Nvidia's growth is still very impressive. He personally believes that SoftBank's exit at this time is also a high-level cashing out.

Nvidia has two main types of businesses, including GPU and artificial intelligence deep learning related products and services. GPU is the core chip of the graphics card (the hardware responsible for outputting display graphics in the computer), and Nvidia sells GPU to graphics card manufacturers such as Gigabyte to generate revenue.

In the field of artificial intelligence, Nvidia is also promoting GPU computing for deep learning (the process of machines being trained to mimic the mechanisms of the human brain to interpret data) and developing new computers and software to apply AI to terminal devices such as self-driving cars, autonomous robots and medical equipment.

Nvidia's downturn is also affected by a variety of factors. In the sluggish macroeconomic environment, the growth rate of its GPU main business revenue has slowed down, of which the slowdown in the sales of gaming graphics cards has had a greater impact on it. In Q4 2018, Nvidia's gaming revenue was US$954 million, down 45% from last year's US$1.739 billion. However, in Q1, Q2, and Q3 of 2018, the growth rates of gaming revenue were 68%, 52%, and 13%, respectively.

Nvidia's gaming revenue accounted for 43% of its total revenue of $2.205 billion in Q4 2018. The decline in gaming revenue also made it more difficult for other projects to boost the overall revenue. Revenue from the data center business (mainly accelerated servers for deep learning in artificial intelligence) increased by 12% from $606 million in Q3 2018 to $679 million in Q4. In Q3, Nvidia said revenue was also lower than expected because customers became "more cautious," resulting in "many expected transactions" not being completed last month.

The macro reason for being “more cautious” is related to the changes in the overall economic situation and the turmoil in US technology stocks since the second half of last year. Artificial intelligence has been a hot field in the past two years, with frequent financing activities and media reports. However, people have found that although artificial intelligence startups have raised a lot of funds, few have been successful.

NVIDIA, known as the "first stock in artificial intelligence," launched its product NVIDIA DRIVE in 2015, marking NVIDIA's official entry into the field of deep learning, according to its official website. This product focuses on autonomous driving, which is also a well-known project of NVIDIA.

Nvidia's stock price has also been rising since 2016, and the title of "the first artificial intelligence stock" has been closely associated with it, although it is essentially a semiconductor company.

After experiencing a wave of AI craze, there is a reflection on this concept. Wang Shuyi told reporters that some AI companies that came in later were keen on hyping up the concept, but Nvidia still has applications, such as data centers and autonomous driving, and has won orders from quite a few car manufacturers.

Some people believe that it is precisely because of the revenue Nvidia has earned in the field of artificial intelligence that latecomers also want to join in and share in the pie. "It is difficult to compare Nvidia with 'PPT companies'. Of course, during this round of the biggest bubble, Nvidia must have also contributed to the flames." Wang Shuyi concluded.