Author: Zhou, ChainCatcher
On June 22, a rise in SK hynix's stock price propelled its market capitalization to 1.35 trillion US dollars, surpassing Bitcoin's total market cap of approximately 1.29 trillion US dollars. It temporarily overtook Samsung Electronics during intraday trading to become the highest-valued company in South Korea.
According to Coinglass data, in the global asset rankings, SK hynix has risen to 16th place, while Bitcoin has slipped to 18th.
HBM, and a 13-Year Bet
The core driver behind SK hynix's recent surge is HBM (High Bandwidth Memory). AI training and inference have extremely high demands on memory bandwidth, and SK hynix is the primary HBM supplier for NVIDIA, holding a market share of over 60%.
Earnings report data shows that SK hynix's Q1 revenue was 52.58 trillion KRW, with an operating profit of 37.61 trillion KRW, resulting in a profit margin of 72%. Analysts currently have a consensus for SK hynix's Q2 operating profit in the range of 62 to 65 trillion KRW, with optimistic predictions from some brokerages already revised upward to over 68 trillion KRW.
In early April this year, most market expectations for Q2 were still around 50 trillion KRW. Subsequently, with the continued strength of memory prices, brokerages have generally made significant upward revisions. Management stated during the earnings call that the structural memory shortage driven by artificial intelligence will persist for at least several years and plans to significantly increase capital expenditure to expand advanced capacity.
Reportedly, SK hynix began betting on HBM technology back in 2009, a time when market attention on this complex and initially low-demand technology was virtually non-existent. From the first generation of HBM to HBM3E, this all-or-nothing gamble lasted nearly 13 years, only reaching its crowning moment with the emergence of ChatGPT.
Image source: AI Generated
SK hynix's journey to this point was not without a crucial external intervention. After the dot-com bubble burst in 2001, Hynix was mired in a debt crisis, its stock price once falling to junk levels. It even negotiated a sale with Micron Technology, which ultimately failed. For the following decade, the company remained under creditor control.
In 2012, SK Group Chairman Chey Tae-won, overriding opposition from the board, used its investment holding subsidiary SK Square to acquire it for approximately 3 billion US dollars. It was renamed SK hynix and infused with substantial R&D funding. It was this investment that allowed the company to continue developing the then-niche HBM technology. Currently, SK Square holds about a 20% stake in SK hynix, making it the largest single shareholder.
It is worth noting that SK Square itself once attempted to enter the crypto market. In 2021, it acquired a 35% stake in the Korean cryptocurrency exchange Korbit for about 90 billion KRW and planned to issue its own token, SK Coin. According to public reports, following the sharp market downturn after the Terra/LUNA crash in 2022, the SK Coin issuance plan was shelved, with no substantial progress since.
Reuters, citing informed sources, reported that SK hynix plans to list on the Nasdaq as early as August this year. This move would lower transaction barriers for US institutional and passive funds, potentially attracting further capital inflows. NVIDIA CEO Jensen Huang recently stated that NVIDIA's collaboration with SK hynix could bring South Korea commercial opportunities worth hundreds of billions of dollars in the future.
Why Capital is Buying: The Mirror of Crypto AI
In this wave of AI, the market is more willing to pay a premium for segments with actual orders and visible supply bottlenecks. Assets directly involved in the AI supply side—computing power, memory, and electricity—have received priority allocation due to their quantifiable revenue and verifiable barriers.
HBM production capacity is highly concentrated in the hands of just three players: SK hynix, Samsung, and Micron, with expansion cycles lasting 2 to 3 years. This scarcity at the physical level is not constructed by narratives; it is locked in by capacity cycles and technological barriers. The valuation logic of the memory industry is also shifting from "cyclical stocks" to "growth stocks."
SK hynix's market cap surpassing Bitcoin is a public statement by the capital markets about two types of scarcity. Given that such high barriers have already formed at the physical layer, the situation of Crypto AI is also worth re-examining.
The Crypto AI sector has been telling a story for the past two years: decentralized computing power will reshape AI infrastructure, and open networks will surpass closed corporate data centers. The potential of this direction is real, but standing before the market cap figure of SK hynix today, there are some realities worth confronting directly.
The IC3 report, a joint publication by Cornell University and 12 other universities, points out that the integration of Crypto and AI remains in its early stages, with the hype surrounding this intersection having already overshadowed actual progress. Decentralized computing, data markets, and governance largely remain at the conceptual stage.
At the project level, taking Bittensor, one of the most representative projects in the Crypto AI sector, as an example, its token TAO has fallen 20% over the past three months. Bittensor co-founder const posted on X, stating that the project's economic incentive layer is still dominated by the core team. They choose to prioritize rapid iteration at the cost of maintaining centralization, estimating it will take another year and a half to complete the core mechanism construction. In other words, their underlying mechanisms are still being patched.
Crypto mining companies, which are closer to the hardware layer, are also in a tough spot. According to Galaxy Research data, Bitcoin miners are entering a "capitulation period." The current network mining difficulty has fallen more than 20% from its historical high, marking the largest decline since China's crackdown on Bitcoin mining in 2021, with some miners continuously exiting the network or shutting down equipment.
In pursuit of transformation, mining companies like Core Scientific, TeraWulf, and Hut 8 have announced entries into the AI and high-performance computing fields. However, according to a VanEck report, this transformation faces a short-term funding gap of approximately 50 billion US dollars, with long-term capital needs around 2.21 trillion US dollars. Furthermore, the industry has currently delivered only about 25% of the leased AI capacity—companies missing construction milestones are already facing investor downgrades.
The IC3 report by Cornell University and others mentions that the integration of Crypto and AI remains in its early stages, with hype overshadowing progress. Decentralized computing, data markets, and governance largely remain conceptual.
In terms of capital, Arthur Hayes pointed out in his recent article "Reality Test" that since ChatGPT's release in 2022, the AI industry has issued approximately 1.5 trillion US dollars in debt, roughly equivalent to the increase in the US M2 money supply during the same period—AI has almost absorbed all the new liquidity, leaving Bitcoin no opportunity. Hayes argues this is not a logic of "funds flowing back to crypto if AI falls." The upcoming massive IPOs of Anthropic and OpenAI will further syphon market funds. Once the AI bubble bursts, bank credit contraction will simultaneously tighten liquidity, and Bitcoin will be sold off along with AI.
Since the second half of last year, many traders previously active in the crypto market have shifted their attention to US and South Korean stocks, chasing the AI hardware trend. The logic behind capital flowing into AI infrastructure is simple and brutal: real orders, physical barriers, quantifiable profit margins.
This certainty is the fundamental reason why capital is willing to pay a high premium today, and it is precisely this kind of certainty that the AI narrative in the crypto market lacks.
In other words, the dividends of AI infrastructure are currently more inclined to be captured by entities with technological barriers and real supply capabilities. Crypto networks need to more clearly define their position in this value chain during this process.
