Former head of Huawei’s Pangu team Wang Yunhe has secured fresh funding for his startup, with plans to train his own large language model next.

Bybit’s institutional-grade real-world asset (RWA) yield platform RWA Earn (branded as Bybit RWA Wealth Management) has added a new nOPAL product, further expanding its institutional business and product portfolio. nOPAL is built on Brazilian credit card receivables, managed by BlackOpal Finance and deployed on open finance platform Plume. Its yield stems from the spread generated when Brazilian merchants sell future credit card receivables at a discount; after acquiring institutions settle via Visa and Mastercard networks, the discounted portion is converted into investor returns. The product hedges Brazilian real (BRL) vs. U.S. dollar (USD) exchange rate volatility using institutional-grade non-deliverable forward (NDF) contracts, with yields paid in USD, and maintains a dedicated liquidity reserve to support redemptions. As of August 2026, nOPAL’s 30-day rolling yield stands at approximately 12%, with total value locked (TVL) exceeding $70 million. Since November 2025, the strategy has recorded zero defaults and zero credit losses. The product has obtained an investment-grade risk rating from Cicada Partners and secured over $300 million in institutional committed capital. nOPAL subscriptions and redemptions on Bybit RWA Earn are priced in USDC, with a minimum investment of 500 USDC and no subscription or redemption fees. To celebrate the product’s launch, eligible users will receive an additional promotional annualized rate of up to 5% on top of the underlying product’s returns.

Following NVIDIA’s earnings release, Wall Street institutions have unanimously raised their price targets for the company: JPMorgan Chase lifted its target from $280 to $320; Mizuho adjusted its target from $300 to $315; Melius raised it from $400 to $420; Goldman Sachs increased its target from $285 to $300; Bernstein hiked its target from $315 to $400; and Citigroup adjusted its target from $300 to $315.

The Sandbox has released a post-mortem analysis report on its August 22 vulnerability incident. The report shows that attackers exploited a flaw in cross-chain configuration-related contracts on Base and BNB Smart Chain (BSC) to steal 14,742,341.84 SAND tokens from its Ethereum vault, accounting for approximately 0.5% of the total maximum supply. The estimated economic impact is around $1.4968 million, with roughly $987,000 actually retained by the attackers. The Ethereum mainnet and Polygon network were not affected. Until further notice, do not purchase or send SAND on Base or BNB Smart Chain. Contracts deployed on Base and BNB Smart Chain have been permanently disabled and will not be reopened. The Sandbox stated that its team has reported the attacker’s wallet addresses to blockchain analytics firms TRM Labs and Chainalysis, and has directly communicated with relevant exchanges. The project also announced a compensation plan: wallets that lawfully held cross-chain SAND on Base or BSC prior to the incident will be compensated at a 1:1 ratio in Ethereum-based SAND, with funds coming from The Sandbox’s vault (no new tokens will be minted). The claim process will open in the next two weeks and remain accessible for two weeks.

According to HTX market data, Bitcoin has once again surpassed the $80,000 mark, currently priced at $80,175.72, with a 2.84% increase in the past 24 hours. Additionally, the crypto whale account "First Set 10 Big Goals" posted a message stating "The bull is back, hurry back", hinting that it has gone long on Bitcoin.

HashKey Holdings Limited today announced its unaudited consolidated interim results for the six months ended June 30, 2026 (the "Report Period"). During the period, the group achieved total consolidated revenue of HK$343 million, up 20.6% year-over-year; gross profit reached HK$208 million, rising 12.5% YoY; gross margin improved sequentially to 60.6% from the second half of 2025. As the group’s core growth engine, transaction facilitation service revenue hit HK$268 million, surging 38.6% YoY. Total platform trading volume rose against the trend to HK$282.2 billion, up 31.8% YoY. Institutional client trading volume continued to grow, jumping 58.8% YoY to HK$231.5 billion, accounting for 82.0% of total trading volume, reflecting institutional investors’ high confidence in compliant platforms. Beyond the trading segment, the group’s on-chain services and asset management businesses also made phased progress. During the period, HashKey launched Hong Kong’s first real estate RWA and first regulated silver RWA token, further consolidating its industry-leading position in on-chain financial innovation. Total on-chain RWA value reached HK$2.68 billion, up 167.8% YoY. Additionally, the group achieved notable growth in its asset management and investment segments. Its assets under management (AUM) stood at HK$5.94 billion, with revenue of HK$38.84 million. In the first half of this year, HashKey strategically invested in Vietnam’s crypto asset exchange Shengxingwang (CAEX), while a fund under HashKey Capital Investment led SignalPlus’ Series B+ round with a US$40 million investment.

Beating AI Express: IBM has open-sourced its new-generation Granite 4.2 language model, available in three variants: 3B, 8B, and 30B. All three models feature native reasoning, capable of switching between full reasoning, low-intensity reasoning, and no reasoning, as well as direct tool calling. The 8B and 30B versions additionally come with dedicated agent reinforcement learning, enabling them to execute tasks in real code repositories, terminals, and web search environments. Training rewards are directly tied to task outcomes, such as whether code passes tests or terminal tasks are completed. All three models are released under the Apache 2.0 license. Independent evaluations show the small models performed surprisingly well: Artificial Analysis awarded the 3B version 14 points, ranking it 2nd among 46 open-weight models of the same parameter size; the 8B version scored 20 points, significantly higher than the median of 9 for its peer group. In IBM’s internal tests, the 30B version achieved 57% on SWE-Bench Verified and 89.17% on AIME25. There is an architectural shift: Granite 4.0 featured a Mamba-2/Transformer hybrid architecture, which IBM claimed reduced memory requirements by over 70% in long-context and multi-concurrency scenarios; Granite 4.2 has switched back to a pure dense Transformer architecture. Compared to the previous generation’s focus on inference efficiency, this generation clearly prioritizes inference and agent capabilities.