After 13 years of gambling, a dying company just surpassed Bitcoin in market capitalization
Author: Zhou, ChainCatcher
Original title: A company that almost went out of business, just surpassed Bitcoin in market capitalization
On June 22, the rise in SK Hynix's stock price led to a market capitalization of 1.35 trillion US dollars, surpassing Bitcoin's total market capitalization of about 1.29 trillion US dollars. At one point in the intraday period, it surpassed Samsung Electronics as the company with the highest market capitalization in Korea.
According to Coinglass data, SK Hynix rose to 16th place in the global asset ranking, while Bitcoin slipped to 18th place.
HBM, and a 13-year gamble
The core driver of SK Hynix's current rise is HBM (high-bandwidth memory). AI training and inference require extremely high memory bandwidth. SK Hynix is Nvidia's main HBM supplier, with a market share of over 60%.
According to financial data, SK Hynix Q1 had revenue of 52.58 trillion won, operating profit of 37.61 trillion won, and a profit margin of 72%. Analysts currently agree that SK Hynix's operating profit for the second quarter is around 62 to 65 trillion won, and some brokers' optimistic forecasts have been raised to over 68 trillion won.
At the beginning of April this year, most of the market's expectations for Q2 were still in the 50 trillion won range. Since then, as memory prices continue to be strong, brokerage firms have generally made drastic revisions. Management said at the earnings conference that the structural memory shortage caused by artificial intelligence will continue for at least a few years, and plans to significantly increase capital expenditure to expand advanced production capacity.
According to reports, SK Hynix began betting on HBM technology in 2009. At that time, the market paid little attention to this complex technology with limited initial demand. From first-generation HBM to HBM3E, this desperate gamble took almost 13 years, and it wasn't until the advent of ChatGPT that it was crowned.
Image Source: AI Generated
SK Hynix has been able to get to where it is today without a critical external aid. After the Internet bubble burst in 2001, Hynix was mired in a debt crisis. At one point, its stock price fell to the level of junk stocks, and even negotiated a sale with Micron Technology, which ultimately ended in failure. For the next ten years, the company was under the control of creditors for a long time.
In 2012, SK Group Chairman Choi Taiyuanli opposed the board of directors, bought it for about $3 billion through SK Square, an investment holding subsidiary, and changed its name to SK Hynix, and injected large-scale R&D capital. It was this investment that allowed the company to continue advancing HBM technology, which was still an unpopular racetrack at the time. SK Square currently holds about 20% of SK Hynix's shares and is its largest single shareholder.
It's worth mentioning that SK Square itself has also tried to enter the crypto market. In 2021, it acquired 35% of the Korean crypto exchange Korbit for about 90 billion won, and plans to issue its own token SK Coin. According to public reports, the market cooled down sharply after the Terra/LUNA crash in 2022, and the SK Coin issuance plan was immediately shelved, and there has been no substantial progress since then.
According to Reuters, citing people familiar with the matter, SK Hynix plans to be listed on the NASDAQ as early as August of this year. This will lower the transaction threshold for US institutions and passive funds, and may further attract capital inflows. Nvidia CEO Hwang In-hoon also recently stated that future cooperation between Nvidia and SK Hynix is expected to bring hundreds of billions of dollars in business opportunities to South Korea.
Why is capital paying for it? Crypto AI in the mirror
In this wave of AI, the market is more willing to pay a premium for links that have already generated actual orders and have visible supply bottlenecks. Computing power, memory, and electricity, these assets directly involved in the AI supply side have been prioritized because revenue can be quantified and barriers can be verified.
HBM production capacity is highly concentrated in SK Hynix, Samsung, and Micron, with a production expansion cycle of 2 to 3 years. This scarcity at the physical level is not built on narratives; it is locked down by production cycles and technical barriers. The storage industry's valuation logic is also shifting from “cycle stocks” to “growth stocks.”
SK Hynix surpassed Bitcoin in market capitalization, a public statement by the capital market on the two scarcity types. The physical layer has formed such a high barrier, and the situation of Crypto AI is worth re-examining.
The Crypto AI circuit has been telling a story for the past two years: decentralized computing power will reshape AI infrastructure, and open networks will surpass closed enterprise data centers. The potential in this direction is real, but in the face of today's SK Hynix market capitalization figures, there are a few realities worth confronting.
According to the IC3 report jointly released by 13 universities including Cornell University, the integration of Crypto and AI is still in its early stages, and the hustle and bustle surrounding this intersectional field has overshadowed actual progress. Decentralized computing power, data markets, and governance are mostly still in the conceptual stage.
At the project level, take Bittensor, the most representative project on the Crypto AI circuit, as an example. Its token, TAO, has declined 20% over the past 3 months. Bittensor co-founder Const wrote on the X platform that the economic incentive layer of the project is still dominated by the core team. They chose to maintain centralization in exchange for rapid iteration. It is estimated that it will take another year and a half to complete the construction of the core mechanism. In other words, their underlying mechanisms are still being patched.
Crypto mining companies that are closer to the hardware layer are also in a difficult situation. According to Galaxy Research data, Bitcoin miners are entering a “surrender period”. Currently, the difficulty of online mining has dropped by more than 20% from its all-time high, making it the biggest pullback since China cracked down on Bitcoin mining in 2021. Some miners continue to withdraw from the network or shut down their devices.
In search of transformation, mining companies such as Core Scientific, TeraWolf, and Hut 8 have announced their entry into the field of AI and high-performance computing. However, according to VanEck, this transformation faces a short-term funding gap of about $50 billion, long-term capital requirements of about $221 billion, and the industry is currently only delivering about 25% of leased AI capacity — companies that have missed construction milestones have faced investor downgrades.
According to the IC3 report jointly released by 13 universities including Cornell University, the integration of Crypto and AI is still in its early stages, and the hustle and bustle surrounding this intersectional field has overshadowed actual progress. Decentralized computing power, data markets, and governance are mostly still in the conceptual stage.
On the financial side, Arthur Hayes pointed out in the recently published article “Reality Test” that since the release of ChatGPT in 2022, the AI industry has accumulated debt issuance of about 1.5 trillion US dollars, which is roughly equivalent to the US dollar M2 increase over the same period — AI has absorbed almost all of the new liquidity, and Bitcoin has never had a chance. Hayes doesn't think this is the logic of “AI goes back to encryption when money falls”. Anthropic and OpenAI's upcoming massive IPOs will further siphon market capital. Once the AI bubble bursts, bank credit contractions will simultaneously tighten liquidity, and Bitcoin will be sold off along with AI.
Since the second half of last year, many traders originally active in the crypto market have turned their attention to US and Korean stocks to chase the AI hardware market. The logic for funding to AI infrastructure is also simple and crude: real orders, physical barriers, quantifiable profit margins.
This kind of certainty is the root cause of current capital's willingness to offer high premiums, and it is this kind of certainty that the AI narrative of the crypto market lacks.
In other words, the dividends of AI infrastructure are currently more likely to be captured by entities with technical barriers and real supply capabilities. In this process, cryptographic networks need to more clearly define their place in the value chain.
Twitter: https://twitter.com/BitpushNewsCN
Compare TG Community: https://t.me/BitPushCommunity
Compare TG subscriptions: https://t.me/bitpush
