2026-08-05 Quarta

Notícias de cripto - Página 231

Mantenha-se a par do mercado de cripto. Notícias em tempo real, análises, preços, histórias em alta e análise de especialistas — tudo num só lugar.

On the Eve of Its U.S. Journey, SK Hynix Plummets Sharply

Just before its highly anticipated U.S. listing, SK Hynix saw its share price plummet dramatically, losing over 14% in a single day. The sell-off was triggered by market fears of a potential slowdown in AI capital expenditure. This followed a news report suggesting Meta might sell "excess AI compute," which was later amended to remove the word "excess." The initial phrasing sparked a chain reaction in investor sentiment, linking it to a potential peak in AI demand. Despite the sharp downturn, the article argues this is likely an overreaction driven by market sentiment and structural de-leveraging, rather than a fundamental reversal of the AI trend. The author points out that even if Meta proceeds, it could be an optimization of existing assets, not a systemic demand contraction. SK Hynix is in the final stages of its U.S. IPO via an ADR listing on Nasdaq, aiming to raise approximately $29.4 billion—one of the largest such offerings ever. The funds are earmarked for expanding domestic Korean production capacity for HBM (High Bandwidth Memory) and advanced packaging. A key motivation for the U.S. listing is to achieve a valuation re-rating, escaping the so-called "Korea discount" and tapping into the higher valuation multiples typically given to AI-related semiconductor stocks in the U.S. market. In conclusion, the article views the current price drop as a potential buying opportunity, suggesting the long-term industry fundamentals for SK Hynix—particularly its leading position in the crucial HBM market—remain strong. The significant capital raised from the IPO is also seen as a factor that could provide underlying support for the stock post-listing.

Odaily星球日报07/02 09:48

On the Eve of Its U.S. Journey, SK Hynix Plummets Sharply

Odaily星球日报07/02 09:48

World Cup Upsets Keep Coming, the 'Dumb Money' in Prediction Markets Got Me Laughing

The 2026 FIFA World Cup has been marked by frequent upsets, turning prediction markets into a high-stakes game of chance. Odaily Planet Daily examines several high-profile cases where "smart money" bets went disastrously wrong, questioning if these losses offer any contrarian insights. A major upset occurred when underdog Cape Verde held football powerhouse Spain to a 0-0 draw. A trader, betting $1 million on a Spanish victory at 0.92 odds to earn $85,000, instead lost their entire principal. This match set a precedent for underdogs stifling favorites. Similarly, Portugal, despite featuring star Cristiano Ronaldo, was held to a 1-1 draw by debutants DR Congo. A trader with a 49% win rate lost over $243,000 predicting a Portuguese win. The article highlights the case of a notorious "anti-indicator" address, @Zzzz87. After initially losing over $620,000 (with a sub-40% win rate) by betting on underdog upsets, the address switched strategy. It began backing favorites in the knockout stages, reportedly turning a $269,000 profit in a week, despite being down $255,000 over the past month. This exemplifies the market's volatility and the difficulty of establishing a consistent strategy. The core conclusion is that football's inherent unpredictability defies simple logic based on player valuations or national rankings. Whether following "smart money" or betting against "dumb money," the only certainty is uncertainty. The article advises enthusiasts to enjoy the games while remaining adaptable in their approach to the prediction markets.

Odaily星球日报07/02 09:41

World Cup Upsets Keep Coming, the 'Dumb Money' in Prediction Markets Got Me Laughing

Odaily星球日报07/02 09:41

UK Crypto Regulation Enters Countdown: What Changes Are in the FCA's New Rules?

UK Crypto Regulation Enters Countdown: Key Changes in FCA's New Rules The UK Financial Conduct Authority (FCA) has published its final crypto asset regulatory policy, marking a shift from consultation to implementation. The new regime, set to fully expand from October 25, 2027, moves crypto businesses from registration to authorisation. A critical window from September 30, 2026, to February 28, 2027, allows existing firms to apply for transitional provisions, enabling them to continue operating while their authorisation is reviewed. Existing registrations will not automatically convert. The rules cover a broad range of activities including stablecoin issuance, custody, trading platform operation, arranging deals, staking, and lending. Authorised firms must meet new prudential capital requirements, with permanent minimum capital ranging from £750,000 for proprietary trading to £75,000 for arranging deals. They must also hold a basic liquid asset buffer for operational resilience. Market integrity rules now apply to trading platforms and intermediaries, covering areas like insider trading and requiring best execution checks against at least three UK-authorised venues. Custodians face enhanced client asset protection rules (CASS 17), focusing on ownership, reconciliation, and private key management. For stablecoins, issuers must provide full backing from issuance and allow redemption at par value within a T+1 timeframe. Reserve assets are categorised into core (e.g., demand deposits) and expanded types, with specific liquidity requirements (ODDR and CBAR) to ensure redemption capacity. Systemically important stablecoin issuers may be jointly regulated by the FCA and the Bank of England. The FCA cautions that the rules do not eliminate the high-risk, speculative nature of most crypto assets but establish a more comprehensive financial regulatory framework for the industry.

Foresight News07/02 09:16

UK Crypto Regulation Enters Countdown: What Changes Are in the FCA's New Rules?

Foresight News07/02 09:16

From Pump.fun to Collector Crypt: Has Solana's Revenue Throne Changed Hands?

From Pump.fun to Collector Crypt: Is Solana’s Revenue Throne Changing Hands? Solana’s on-chain narrative is expanding from meme coins to tokenized trading cards. Previously, Pump.fun dominated consumer revenue on Solana. However, starting in Q2 2026, Pump.fun's quarterly revenue growth slowed, while Collector Crypt shows a stronger recent growth trajectory. Data indicates Pump.fun's Q1 revenue was $108.3M, dropping 36.1% to $69.2M in Q2 so far. Conversely, Collector Crypt's revenue grew 108.8%, from $12.3M in Q1 to $25.8M in Q2 to date, with recent weekly revenue hitting $5.1M. While Pump.fun remains larger in scale, Collector Crypt demonstrates stronger short-term momentum. Collector Crypt's model involves custodying graded physical trading cards, minting corresponding NFTs on Solana, and selling randomized packs. Revenue comes from pack sales, secondary market fees, and royalties. Its profitability, estimated around 4-5%, stems from bulk card purchases at a discount and a buyback system for users. However, a full-scale reversal in rankings hasn't occurred. Year-to-date, Pump.fun's revenue is ~$177.5M (with its ecosystem at ~$466.5M), vastly exceeding Collector Crypt's ~$38.1M. The shift is more about narrative and recent growth dynamics than total historical volume. Pump.fun relies on speculative token launches, while Collector Crypt focuses on collectibility, scarcity, and real-world asset backing. This highlights a broadening of Solana's consumer revenue sources beyond meme coins. The sustainability of Collector Crypt's growth depends on maintaining demand for randomized packs, expanding beyond Pokémon cards (currently dominant) into sports cards, and navigating potential regulatory scrutiny on "loot box" mechanics. In conclusion, Pump.fun remains the larger revenue engine, but Collector Crypt's rise signifies a diversification of Solana's consumer application revenue into tangible, collection-based scenarios alongside speculative meme coin activity.

链捕手07/02 09:15

From Pump.fun to Collector Crypt: Has Solana's Revenue Throne Changed Hands?

链捕手07/02 09:15

EF: Ethereum Is Becoming the Most Needed Neutral Infrastructure for Governments and Institutions

Ethereum Foundation's Global Policy Strategy team released a report titled "Ethereum for Governments and Institutions," positioning Ethereum as critical, neutral public infrastructure. It argues that current digital systems, controlled by centralized intermediaries, create single points of failure and trust risks. In contrast, Ethereum’s decentralized architecture offers a credible, neutral alternative for applications like digital identity, public records, and asset tokenization. The report highlights Ethereum's resilience, noting it has never experienced downtime since 2015. It is secured by approximately $76 billion in staked ETH, features a geographically diverse validator network, and maintains multiple independent client implementations. This design eliminates counterparty risk, as no single entity can alter rules or shut down the network. Compared to other blockchains assessed in a referenced OpenZeppelin report, Ethereum demonstrates superior economic security, lower validator participation barriers promoting decentralization, and a mature ecosystem with over 11,000 developers. Real-world implementations, such as decentralized identity projects in Bhutan and Buenos Aires and land registry systems in India, are cited as evidence of its practical utility for public sectors. The guide serves as a non-technical primer for policymakers, emphasizing the need for infrastructure that preserves sovereignty and enables coordination without reliance on centralized intermediaries.

Odaily星球日报07/02 09:06

EF: Ethereum Is Becoming the Most Needed Neutral Infrastructure for Governments and Institutions

Odaily星球日报07/02 09:06

After Snagging a Nobel Laureate, Anthropic Poaches Berkeley CS Department Head, Recruiting Four Top Talents in Two Weeks

In a stunning move, Anthropic has recruited Jelani Nelson, the chair of UC Berkeley's prestigious EECS Computer Science Division and a leading theoretical computer scientist, on a leave of absence. This follows a two-week hiring spree where Anthropic also secured Nobel laureate John Jumper and two key Gemini researchers from Google. Nelson's expertise in streaming algorithms, dimensionality reduction, and randomized algorithms—fundamentally about processing vast data with minimal resources—directly addresses core challenges in large language models: training efficiency, data compression, and computational complexity. His work on the Johnson-Lindenstrauss lemma underpins modern vector search and embedding compression. Anthropic's recruitment signals a strategic shift in the AI race from merely scaling models to optimizing foundational algorithms for efficiency. This "leave of absence" model, exemplified by figures like Fei-Fei Li, is becoming a mainstream talent pipeline, allowing scholars to retain academic positions while gaining industry access to unprecedented compute and real-world problems. The recent talent war has escalated from poaching between AI firms to raiding top university departments, with Berkeley being a prime target. As OpenAI and Anthropic near potential IPOs, offering pre-IPO equity, they are effectively becoming parallel research institutions. The competition's focus is now descending to the theoretical bedrock of algorithms.

marsbit07/02 09:03

After Snagging a Nobel Laureate, Anthropic Poaches Berkeley CS Department Head, Recruiting Four Top Talents in Two Weeks

marsbit07/02 09:03

A Year Consumes a Solid-State Drive: Codex Log Bug Slammed as 'Slopware'

OpenAI's flagship AI coding tool, Codex, was found to have a critical bug causing its feedback logging system to silently and rapidly wear out users' SSDs. A developer reported that Codex was writing approximately 640 TB of data per year to a local SQLite database (`logs_2.sqlite`) through a constant cycle of inserting and immediately deleting log entries, primarily at the verbose TRACE level. While the database file itself remained around 1 GB, the underlying write-amplification from SQLite's WAL mechanism meant the physical SSD endured the full write load. This was enough to exceed the typical 600 TBW endurance rating of a consumer SSD within a year. The root cause was a hardcoded default logging level (`Level::TRACE`) in the configuration, which overrode any user attempts to reduce logging via environment variables. Analysis showed that over 96% of the logged data—including noisy WebSocket packet dumps and repeated system file events—was useless debug information. The issue, which had at least nine related bug reports in the Codex repository, remained latent because it didn't visibly consume disk space, only silently accumulated write cycles. After the report gained traction on Hacker News, OpenAI merged fixes estimated to reduce writes by about 85%. However, even post-fix, the tool would still write an estimated 96 TB annually. The incident sparked broader criticism of "slopware" in AI-assisted development tools, highlighting a lack of resource budgeting for disk, CPU, and memory in always-on agent software, and a reliance on modern hardware to mask inefficient code. Competing tools like Claude Code were noted to have similar issues.

marsbit07/02 08:50

A Year Consumes a Solid-State Drive: Codex Log Bug Slammed as 'Slopware'

marsbit07/02 08:50

活动图片