7*24

Hot topic

Market

Activities

On July 14, Euro Stoxx 50 futures narrowed losses, currently down 0.6%, Germany's DAX index futures down 0.6%, and UK's FTSE 100 index futures down 0.2%. (Jin Shi)

[Euro Stoxx 50 Futures Narrow Losses, Currently Down 0.6%]On July 14, Euro Stoxx 50 futures narrowed losses, currently down 0.6%, Germany's DAX index futures down 0.6%, and UK's FTSE 100 index futures down 0.2%. (Jin Shi)

Euro Stoxx 50 Futures Narrow Losses, Currently Down 0.6%

According to Beating, Jenny Wen, the design lead for Claude and Cowork, has announced her departure from Anthropic. She will join Cursor as Head of Design. She previously led the visual redesign of Claude.ai and set the early product direction for Cowork. Before joining Anthropic, she was a design director at Figma, where she led product teams for FigJam and Slides. This move brings Claude's core product experience to a competitor. Cursor and Claude Code are competing in the AI coding agent market, and product design will be part of the competition.

[Claude and Cowork Design Lead Joins Cursor]According to Beating, Jenny Wen, the design lead for Claude and Cowork, has announced her departure from Anthropic. She will join Cursor as Head of Design. She previously led the visual redesign of Claude.ai and set the early product direction for Cowork. Before joining Anthropic, she was a design director at Figma, where she led product teams for FigJam and Slides. This move brings Claude's core product experience to a competitor. Cursor and Claude Code are competing in the AI coding agent market, and product design will be part of the competition.

Claude and Cowork Design Lead Joins Cursor

On July 14, according to Beating monitoring, after DeepSeek completed its first round of financing, founder Liang Wenfeng's personal wealth surged alongside the company's valuation. The Bloomberg Billionaires Index adjusted his net worth from $16.7 billion to $36 billion, an increase of $19.3 billion. According to Bloomberg's statistics, he has surpassed Anthropic co-founder Dario Amodei and OpenAI co-founder Greg Brockman to become the world's highest-valued AI model founder. DeepSeek's latest financing round raised over 50 billion yuan, with the company valued at over $50 billion. Liang Wenfeng personally invested approximately 20 billion yuan, accounting for nearly 40% of the financing amount, making him the largest contributor. The financing structure is also unique. Most investors did not directly invest in DeepSeek but instead contributed funds to a limited partnership managed by Liang Wenfeng. Investors are required to lock in their investment for five years and typically have no voting rights. DeepSeek secured substantial capital, while Liang Wenfeng retained firm control.

[DeepSeek Founder Liang Wenfeng's Net Worth Surges to $36 Billion, Becoming World's Richest AI Model Founder]On July 14, according to Beating monitoring, after DeepSeek completed its first round of financing, founder Liang Wenfeng's personal wealth surged alongside the company's valuation. The Bloomberg Billionaires Index adjusted his net worth from $16.7 billion to $36 billion, an increase of $19.3 billion. According to Bloomberg's statistics, he has surpassed Anthropic co-founder Dario Amodei and OpenAI co-founder Greg Brockman to become the world's highest-valued AI model founder. DeepSeek's latest financing round raised over 50 billion yuan, with the company valued at over $50 billion. Liang Wenfeng personally invested approximately 20 billion yuan, accounting for nearly 40% of the financing amount, making him the largest contributor. The financing structure is also unique. Most investors did not directly invest in DeepSeek but instead contributed funds to a limited partnership managed by Liang Wenfeng. Investors are required to lock in their investment for five years and typically have no voting rights. DeepSeek secured substantial capital, while Liang Wenfeng retained firm control.

DeepSeek Founder Liang Wenfeng's Net Worth Surges to $36 Billion, Becoming World's Richest AI Model Founder

SK Hynix has reportedly begun procuring equipment for the first phase cleanroom of Y1, the first fab in its Yongin Semiconductor Cluster, with initial plans to add 20,000 units per month of 1c DRAM capacity. This was reported by South Korean media ZDNET on July 14th. The activation of Y1's first-phase cleanroom, originally scheduled for May 2027, has been moved up to February 2027. The company plans to begin construction of the pilot production line at that time, followed by formal equipment installation in March to April, expanding its monthly production capacity by 20,000 units. This production line will primarily manufacture sixth-generation 10-nanometer class 1c DRAM, which can be used in AI-related high-value-added DDR and LPDDR products, and is also planned for application in HBM4E, expected to be commercialized as early as 2027. To expedite equipment introduction, SK Hynix has also advanced equipment price contract negotiations, typically held at year-end, to early in the third quarter. SK Hynix is simultaneously accelerating the overall construction progress of the Yongin Semiconductor Cluster. The project, with a total investment of approximately 600 trillion Korean Won, comprises four fabs. The completion target for the fourth fab has been moved up from 2045 to 2033. Construction of Y1's second and third phase cleanrooms is expected to begin in the second half of this year.

[SK Hynix Accelerates Yongin Cluster Construction, Begins Y1 Fab Equipment Procurement]SK Hynix has reportedly begun procuring equipment for the first phase cleanroom of Y1, the first fab in its Yongin Semiconductor Cluster, with initial plans to add 20,000 units per month of 1c DRAM capacity. This was reported by South Korean media ZDNET on July 14th. The activation of Y1's first-phase cleanroom, originally scheduled for May 2027, has been moved up to February 2027. The company plans to begin construction of the pilot production line at that time, followed by formal equipment installation in March to April, expanding its monthly production capacity by 20,000 units. This production line will primarily manufacture sixth-generation 10-nanometer class 1c DRAM, which can be used in AI-related high-value-added DDR and LPDDR products, and is also planned for application in HBM4E, expected to be commercialized as early as 2027. To expedite equipment introduction, SK Hynix has also advanced equipment price contract negotiations, typically held at year-end, to early in the third quarter. SK Hynix is simultaneously accelerating the overall construction progress of the Yongin Semiconductor Cluster. The project, with a total investment of approximately 600 trillion Korean Won, comprises four fabs. The completion target for the fourth fab has been moved up from 2045 to 2033. Construction of Y1's second and third phase cleanrooms is expected to begin in the second half of this year.

SK Hynix Accelerates Yongin Cluster Construction, Begins Y1 Fab Equipment Procurement

On July 14th, renowned US stock KOL Herman Jin stated in his latest market interpretation video, 'Momentum Unwind,' that one of the most crowded trades in the market over the past few weeks involved going long on AI physical assets such as semiconductors, memory, optical communications, neocloud, power, and data centers, while simultaneously shorting software and other old technology assets perceived to be impacted by AI. When deleveraging began, investors were forced to sell the hardware longs with the largest gains and best liquidity, and cover software shorts, leading to a reverse trend of 'hardware down, software up.' Herman Jin believes that SK Hynix's large-scale ADR listing was one of the direct triggers for this round of adjustment. The new supply required capital from funds already heavily invested in AI to absorb it, forcing some investors to reallocate positions. The memory sector's leading decline does not signify the disappearance of HBM or AI demand, but rather that this sector, with its large gains, heavy positions, and strong cyclical attributes, is most susceptible to selling during capital contraction. Herman Jin assesses that the medium-term fundamentals for AI hardware have not yet deteriorated, but the short-term trading structure remains fragile. While the unwinding of the first phase of the long-semiconductor, short-software trade may be subsiding, factors such as pure-long AI funds reducing positions, negative gamma, potential CTA selling, and low summer liquidity could still trigger a second wave of declines. The real risk to be wary of is not in stock prices themselves, but in the credit market. If major cloud vendors continue to issue bonds to sustain AI capital expenditures, leading to increased financing costs, market doubts about the return on AI investments could deepen further, eventually transmitting to orders for GPUs, HBM, optical communications, power, and data centers. Herman Jin is reportedly a former FICC (Fixed Income, Currencies & Commodities) Executive Director at Goldman Sachs Asia, having worked at Goldman Sachs Hong Kong for over 18 years (since approximately 2008).

[Herman Jin: SK Hynix ADR Listing Triggered Sell-off, Credit Market Poses Real Risk]On July 14th, renowned US stock KOL Herman Jin stated in his latest market interpretation video, 'Momentum Unwind,' that one of the most crowded trades in the market over the past few weeks involved going long on AI physical assets such as semiconductors, memory, optical communications, neocloud, power, and data centers, while simultaneously shorting software and other old technology assets perceived to be impacted by AI. When deleveraging began, investors were forced to sell the hardware longs with the largest gains and best liquidity, and cover software shorts, leading to a reverse trend of 'hardware down, software up.' Herman Jin believes that SK Hynix's large-scale ADR listing was one of the direct triggers for this round of adjustment. The new supply required capital from funds already heavily invested in AI to absorb it, forcing some investors to reallocate positions. The memory sector's leading decline does not signify the disappearance of HBM or AI demand, but rather that this sector, with its large gains, heavy positions, and strong cyclical attributes, is most susceptible to selling during capital contraction. Herman Jin assesses that the medium-term fundamentals for AI hardware have not yet deteriorated, but the short-term trading structure remains fragile. While the unwinding of the first phase of the long-semiconductor, short-software trade may be subsiding, factors such as pure-long AI funds reducing positions, negative gamma, potential CTA selling, and low summer liquidity could still trigger a second wave of declines. The real risk to be wary of is not in stock prices themselves, but in the credit market. If major cloud vendors continue to issue bonds to sustain AI capital expenditures, leading to increased financing costs, market doubts about the return on AI investments could deepen further, eventually transmitting to orders for GPUs, HBM, optical communications, power, and data centers. Herman Jin is reportedly a former FICC (Fixed Income, Currencies & Commodities) Executive Director at Goldman Sachs Asia, having worked at Goldman Sachs Hong Kong for over 18 years (since approximately 2008).

Herman Jin: SK Hynix ADR Listing Triggered Sell-off, Credit Market Poses Real Risk

On July 14, over a dozen single-stock leveraged ETFs listed in South Korea at the end of May saw significant declines, with multiple products tracking Samsung Electronics and SK Hynix nearly halving in price. Among them, the 'SAMSUNG KODEX SK Hynix Single-Stock Leveraged ETF,' with assets of $3.4 billion, has fallen approximately 45% since its listing, retreating over 60% from its June peak. The total assets under management of related products amount to about $3 billion. On Monday, SK Hynix recorded a record drop of 15%; on Tuesday, it fell further by over 8% during intraday trading, while the KOSPI index once declined 5% and fell below the 6,500 mark. Jung In Yun, CEO of Fibonacci Asset Management, stated that many retail investors treat leveraged ETFs as long-term investment tools, and the sharp decline has caused them significant losses, potentially weakening their willingness and ability to continue buying semiconductor stocks. South Korea's top financial regulator last month expressed regret over approving the listing of such products, but retail trading enthusiasm has not notably cooled. Over the past month, leveraged and inverse exchange-traded products in South Korea attracted a total of $3.8 billion in inflows, primarily directed toward single-stock products tracking SK Hynix and Samsung Electronics. Jung In Yun expects regulators to strengthen suitability requirements, risk disclosures, and investor education rather than fully suspending the products. On the same day, the South Korean government raised its economic growth forecast for this year from 2% to 3%, citing continued demand for AI chips to offset some of the drag from Middle East conflicts, and projected a record current account surplus of $290 billion this year.

[Korean Leveraged Chip ETFs Plummet 45%, Retail Investors Suffer Heavy Losses]On July 14, over a dozen single-stock leveraged ETFs listed in South Korea at the end of May saw significant declines, with multiple products tracking Samsung Electronics and SK Hynix nearly halving in price. Among them, the 'SAMSUNG KODEX SK Hynix Single-Stock Leveraged ETF,' with assets of $3.4 billion, has fallen approximately 45% since its listing, retreating over 60% from its June peak. The total assets under management of related products amount to about $3 billion. On Monday, SK Hynix recorded a record drop of 15%; on Tuesday, it fell further by over 8% during intraday trading, while the KOSPI index once declined 5% and fell below the 6,500 mark. Jung In Yun, CEO of Fibonacci Asset Management, stated that many retail investors treat leveraged ETFs as long-term investment tools, and the sharp decline has caused them significant losses, potentially weakening their willingness and ability to continue buying semiconductor stocks. South Korea's top financial regulator last month expressed regret over approving the listing of such products, but retail trading enthusiasm has not notably cooled. Over the past month, leveraged and inverse exchange-traded products in South Korea attracted a total of $3.8 billion in inflows, primarily directed toward single-stock products tracking SK Hynix and Samsung Electronics. Jung In Yun expects regulators to strengthen suitability requirements, risk disclosures, and investor education rather than fully suspending the products. On the same day, the South Korean government raised its economic growth forecast for this year from 2% to 3%, citing continued demand for AI chips to offset some of the drag from Middle East conflicts, and projected a record current account surplus of $290 billion this year.

Korean Leveraged Chip ETFs Plummet 45%, Retail Investors Suffer Heavy Losses

On July 14, according to foreign media reports, after DeepSeek completed its latest funding round, founder Liang Wenfeng's personal wealth more than doubled, making him the highest-net-worth founder among global AI large model companies. According to the Bloomberg Billionaires Index, Liang Wenfeng's current net worth is $36 billion (previously about $16.7 billion), surpassing Anthropic co-founder Dario Amodei and OpenAI co-founder Greg Brockman. This comparison only covers companies whose main business and majority of revenue come directly from AI large models, excluding large diversified technology groups such as Alibaba and Tencent, as well as AI industry chain-related companies like data centers and semiconductors. The vast majority of Liang Wenfeng's wealth comes from his controlling stake in DeepSeek. In June of this year, DeepSeek reportedly completed a 50 billion yuan financing round, with the company valued at over 330 billion yuan, and Liang Wenfeng personally contributed 20 billion yuan. According to the Bloomberg Billionaires Index, his shareholding was diluted to approximately 78%. (Jin Shi)

[After DeepSeek's Latest Funding Round, Liang Wenfeng's Net Worth Surpasses Anthropic's Dario and OpenAI's Brockman]On July 14, according to foreign media reports, after DeepSeek completed its latest funding round, founder Liang Wenfeng's personal wealth more than doubled, making him the highest-net-worth founder among global AI large model companies. According to the Bloomberg Billionaires Index, Liang Wenfeng's current net worth is $36 billion (previously about $16.7 billion), surpassing Anthropic co-founder Dario Amodei and OpenAI co-founder Greg Brockman. This comparison only covers companies whose main business and majority of revenue come directly from AI large models, excluding large diversified technology groups such as Alibaba and Tencent, as well as AI industry chain-related companies like data centers and semiconductors. The vast majority of Liang Wenfeng's wealth comes from his controlling stake in DeepSeek. In June of this year, DeepSeek reportedly completed a 50 billion yuan financing round, with the company valued at over 330 billion yuan, and Liang Wenfeng personally contributed 20 billion yuan. According to the Bloomberg Billionaires Index, his shareholding was diluted to approximately 78%. (Jin Shi)

After DeepSeek's Latest Funding Round, Liang Wenfeng's Net Worth Surpasses Anthropic's Dario and OpenAI's Brockman

On July 14, according to Farside Investors monitoring, yesterday the US Bitcoin spot ETF had a net outflow of $424.7 million, of which IBIT recorded a net outflow of $185.5 million and FBTC recorded a net outflow of $245.6 million. The Ethereum spot ETF saw a net outflow of $15.4 million, all attributed to FETH.

[Bitcoin ETF Net Outflow Reaches $424.7M, Ethereum ETF Net Outflow $15.4M]On July 14, according to Farside Investors monitoring, yesterday the US Bitcoin spot ETF had a net outflow of $424.7 million, of which IBIT recorded a net outflow of $185.5 million and FBTC recorded a net outflow of $245.6 million. The Ethereum spot ETF saw a net outflow of $15.4 million, all attributed to FETH.

Bitcoin ETF Net Outflow Reaches $424.7M, Ethereum ETF Net Outflow $15.4M

According to monitoring by Beating, Goldman Sachs estimates that Google, Amazon, Meta, Microsoft, and Oracle will invest $5.8 trillion by 2030 to build AI data centers and computing facilities. Even these tech giants cannot rely solely on cash. The five companies have issued nearly $200 billion in corporate bonds this year and borrowed about $90 billion through joint ventures for data centers. The problem is that although these bonds appear to be backed by big-name companies, the actual guarantees vary greatly. Some projects have leases fully taken over by Google. Others can only collect rent at discounted rates. There are also provisions that if construction is delayed by more than six months, Google and the tenant can exit directly. If the data center is not completed on time, rent payments will not begin, and investors may not get their money back. However, the market currently does not distinguish these risks carefully, and the borrowing rates for different projects are not very different. As more data centers start construction, delays and cost overruns may cause some bonds to default first.

[AI Infrastructure to Cost $5.8 Trillion, Bond Market May Underestimate Abandoned Project Risks]According to monitoring by Beating, Goldman Sachs estimates that Google, Amazon, Meta, Microsoft, and Oracle will invest $5.8 trillion by 2030 to build AI data centers and computing facilities. Even these tech giants cannot rely solely on cash. The five companies have issued nearly $200 billion in corporate bonds this year and borrowed about $90 billion through joint ventures for data centers. The problem is that although these bonds appear to be backed by big-name companies, the actual guarantees vary greatly. Some projects have leases fully taken over by Google. Others can only collect rent at discounted rates. There are also provisions that if construction is delayed by more than six months, Google and the tenant can exit directly. If the data center is not completed on time, rent payments will not begin, and investors may not get their money back. However, the market currently does not distinguish these risks carefully, and the borrowing rates for different projects are not very different. As more data centers start construction, delays and cost overruns may cause some bonds to default first.

AI Infrastructure to Cost $5.8 Trillion, Bond Market May Underestimate Abandoned Project Risks

MakerDAO official tweeted that MakerDAO is reviewing a proposal to extend its existing US treasury bond investments from $500 million to $1.25 billion. The proposal, put forward by Monetalis, the original creator of MIP65, involves investing an additional $750 million into a 6-month US Treasury ladder strategy with bi-weekly maturities. According to the proposal, the strategy is expected to deliver a net annualized yield of 4.6% to 4.5% to the MakerDAO after custody and expected trading costs. To be implemented, the MIP65 debt ceiling extension must go through an on-chain governance cycle, including a Governance Poll and an Executive Vote, to be approved and implemented respectively. Under MIP65, MakerDAO deployed ~$500 million in short-term US treasury bond ETFs. As of January 2023, this investment strategy has brought ~2.1 million in lifetime fees to the Maker Protocol.

Detail

【MakerDAO reviews a proposal to extend its existing US treasury bond investments from $500 million to $1.25 billion】MakerDAO official tweeted that MakerDAO is reviewing a proposal to extend its existing US treasury bond investments from $500 million to $1.25 billion. The proposal, put forward by Monetalis, the original creator of MIP65, involves investing an additional $750 million into a 6-month US Treasury ladder strategy with bi-weekly maturities. According to the proposal, the strategy is expected to deliver a net annualized yield of 4.6% to 4.5% to the MakerDAO after custody and expected trading costs. To be implemented, the MIP65 debt ceiling extension must go through an on-chain governance cycle, including a Governance Poll and an Executive Vote, to be approved and implemented respectively. Under MIP65, MakerDAO deployed ~$500 million in short-term US treasury bond ETFs. As of January 2023, this investment strategy has brought ~2.1 million in lifetime fees to the Maker Protocol.

MakerDAO reviews a proposal to extend its existing US treasury bond investments from $500 million to $1.25 billion

CASHCAT Market Cap Surpasses $100 Million! Robinhood Chain Goes Viral

Behind ANSEM's 300% Weekly Surge: The Recovery and Traps of Solana Meme Coins

Will Zcash’s ‘formal verification’ of Ironwood clear counterfeiting fears and boost ZEC?

Goldman Sachs Report Deconstructs the Competitive Landscape of China's AI Large Models: Who Will Be the Long-Term Winner?

Ethereum uses less power than British Museum – Here’s what changed

AI Overhauled Terence Tao's 30-Year-Old Website, Uncovering Two Bugs Hidden for Over Two Decades in the Process

How to Regulate Single-Stock Leveraged ETFs? On Thursday, the Entire Market Is Watching This Korean Government Meeting

Crude Oil Surges 10% in Intraday Trading! US Military Restores Blockade on Iran, Trump Warns of 'Heavy Blows' Tonight and Tomorrow, Imposes 20% Fee on Strait Shipping

SBI partners with Solana to expand Japan’s on-chain finance, but can it become Asia’s hub?

Wall Street Morning Report: U.S. Stocks Suffer Collective Setback, Apple Hits New High Against the Trend, Tonight's CPI and Waller's Hearing to Determine Interest Rate Path

MakerDAO reviews a proposal to extend its existing US treasury bond investments from $500 million to $1.25 billion

Isenção de responsabilidade：O conteúdo acima não representa as posições da HTX.，A HTX não faz quaisquer recomendações de negociação.。

Todos os comentários0Mais recentePopular

Tópicos em destaque