seek

According to Bloomberg, informed sources revealed that JPMorgan Chase has begun preliminary discussions with potential lenders to seek a $5 billion debt financing plan to support Volta Infra Holdings Ltd. in building artificial intelligence data centers.Volta AI completed a $300 million venture capital financing earlier this month, with a valuation of $2.4 billion, aiming to help more tech companies access the costly resources of artificial intelligence chips.

Zhipu released the lightweight flagship model GLM-5.3 Flash, aimed at competing with DeepSeek V4 Flash, and utilizes over 100,000 domestic chip clusters to provide inference services. The model has a parameter scale of approximately 300B, about 40% of GLM-5.3, and is roughly on par with DeepSeek V4 Flash; it adopts the MoE architecture, with actual activated parameters around 18B, possesses native multimodal capabilities, supports image and video input, and is the first new model with multimodal capabilities following Zhipu's focus on coding strategy.The comprehensive evaluation score exceeds that of the previous flagship GLM-5.2, is on par with Claude Opus 4.8, and is higher than the relevant scores of DeepSeek V4 Pro official version. The pricing is 0.8 yuan per million tokens for input, 2.8 yuan for output, and 0.23 yuan for cache hits, approximately one-tenth of GLM-5.3, with a half-price discount offered for the first two weeks after launch. The model is fully compatible with domestic computing power, online traffic is handled by domestic chips, and Zhipu claims that the cluster hardware efficiency and unit token cost have reached levels comparable to mainstream NVIDIA GPUs; this model was previously tested anonymously on overseas platforms under the code name Ox Alpha.

According to The Information, DeepSeek is seeking a second round of financing with a valuation of 500 billion yuan, planning to raise 50 billion yuan. The company's revenue from January to July 2026 reached 475 million yuan, approximately 10 times the total revenue for 2025; during the same period, the net loss narrowed to 715 million yuan, compared to a net loss of 935 million yuan for the entire year of 2025. The gross margin is 44.6%, with the API business gross margin reaching 82.9%, mainly due to reduced inference costs from optimized computing infrastructure.Recently, DeepSeek raised its API pricing, with the V4-Pro peak output price increasing from $0.87 to $3.96, but still far below Kimi K3's $15 and Claude Opus 5's $25. The company's investment in computing infrastructure surged from 1.2 billion yuan in 2025 to 11 billion yuan from January to July 2026, primarily for server leasing and AI chip procurement. DeepSeek completed its first round of external financing of 50 billion yuan in June, and the company has hired investment banks to prepare for an IPO in Shanghai next year.

According to Cointelegraph, the Securities and Exchange Commission (SEC) of Thailand has advanced its regulatory framework for locally listed spot Bitcoin and Ethereum ETFs from a principled proposal to the rule draft stage and is publicly soliciting opinions on this. The regulatory agency released two consultation documents on Monday, one containing the rule draft for Thailand's crypto ETFs, and the other proposing qualification principles for foreign digital asset custodians.In the initial phase, asset management companies can establish passive ETFs that track Bitcoin or Ethereum, which are the only qualified crypto assets. According to the proposed rules, Bitcoin and Ethereum ETFs will only trade on the Stock Exchange of Thailand (SET), with each ETF tracking a single crypto asset and required to maintain at least 80% net asset exposure to that asset within each accounting year. Mutual funds and private funds can also invest in Thailand's local crypto ETFs, as well as the foreign crypto ETFs they are permitted to invest in, but must comply with existing investment limits. However, in the initial phase, the regulatory agency does not allow alternative products linked to foreign crypto ETFs, including depositary receipts that track them.Regarding custody, the revised plan still primarily relies on domestic digital asset custodians as the main service providers in the initial phase, and the Thai SEC may allow the use of qualified foreign digital asset custodians when necessary. Foreign custodians must be supervised by a regulatory agency with legal authority and meet the regulatory and investor asset protection standards deemed sufficient by the Thai SEC. The deadline for public opinion collection on the two consultation documents is September 20.

According to Bloomberg, the Taiwanese research institution TeamT5 stated that the scale of cyberattacks linked to China has significantly increased after incorporating DeepSeek and other open-source AI models into their operations. Since they began delegating daily tasks to AI and using it to develop more advanced malware, the number of attacks carried out by the relevant organizations has more than doubled.Researchers noted that it is not always possible to accurately identify the specific AI models used by attackers, but products related to DeepSeek are popular among hacker groups in the country due to their high performance and ease of customization. The report indicates that experienced attackers are leveraging relatively limited AI tools to scale up their operations and achieve breakthroughs, highlighting the attackers' ability to use basic AI tools to target overseas objectives.

Przemysław Kral, the head of the Polish cryptocurrency exchange Zondacrypto, is reportedly facing criminal charges for involvement in a large-scale fraud and has begun cooperating with Polish prosecutors. According to Polish media Onet, Kral is seeking a reduced sentence in exchange for testimony, which may involve Zondacrypto's funding of right-wing political figures.Prosecutors estimate that Zondacrypto customers have lost at least 2.4 billion Polish złoty (approximately 650 million USD). Investigators claim that the exchange only used a portion of customer funds to purchase cryptocurrencies, while the remaining funds were transferred to private accounts controlled by Zondacrypto's management. Kral has remained silent on X since mid-April when he publicly disclosed that Zondacrypto could not access a cold wallet holding approximately 4,500 bitcoins. He denies misappropriating funds, stating that the private keys for the wallet were supposed to be handed over by Zonda's founder and former CEO Sylwester Suszek, who has been missing since 2022.Reports also state that Zondacrypto was a major sponsor of the Polish CPAC conference, which took place before the second round of the presidential election, ultimately won by Karol Nawrocki. The exchange spent 37 million złoty on advertising with the broadcasting station Telewizja Republika in the last year, while Kral's company also made payments to foundations associated with politicians Zbigniew Ziobro and Przemysław Wipler.

The B.AI platform has launched the new multimodal experimental flagship model DeepSeek-V4-Flash-Vision-Exp. The official API team has completed full integration and opened it up for free to all users.This model builds on the top-tier pure text and code Agent capabilities of V4-Flash, officially unlocking native image understanding capabilities. The comprehensive performance of the multimodal Agent is now close to the flagship level of Claude Opus-4.8.The new model supports mixed text and image input and 1M ultra-long context, easily handling hardcore scenarios such as multimodal PPT generation, complex UI reconstruction, and frontend effects. Image input is billed by token, with a single image costing a maximum of only 384 tokens. Now, all users can log in to the B.AI platform to access this latest multimodal model capability for free without any barriers, allowing top-tier computing power to accelerate creativity and productivity.

On August 18, 2026, the Financial Accounting Standards Board (FASB) issued a proposed Accounting Standards Update (ASU) aimed at clarifying how the current definition of "cash equivalents" applies to certain digital assets, such as stablecoins, and enhancing the transparency of disclosures regarding important components of cash equivalents. Stakeholders are encouraged to submit comments by November 19, 2026.In the 2025 FASB agenda consultation project and other feedback, stakeholders pointed out that there is uncertainty regarding whether certain digital assets, including stablecoins, meet the definition of cash equivalents under current Generally Accepted Accounting Principles (GAAP), leading to differences in practical treatment. The proposed ASU will provide illustrative examples to promote a more consistent application of this definition and enhance comparability among entities choosing to report qualifying digital assets as cash equivalents, but it will not change the current definition of "cash equivalents."At the same time, the proposed rules require all entities reporting assets as cash equivalents, regardless of whether they include digital assets, to enhance disclosures regarding the important components of cash equivalents and related amounts, so that investors and other financial statement users can obtain more transparent information. The relevant proposed ASU and methods for submitting comments can be found on the FASB website.

The B.AI platform's DeepSeek-V4-Flash zero-threshold free usage event is gaining momentum, with the total Token throughput across the platform exceeding 641.524 billion, including a single-day peak of 285.963 billion Tokens on August 20, nearly quadrupling since the first day of the event and continuously breaking historical records.Behind this set of data is not only a hardcore verification of the platform's high concurrency processing capability but also reflects the real explosion of AI universal demand. B.AI will continue to reduce the cost of AI usage, providing developers and Web3 users with a low-threshold, unlimited Agentic application building and experience environment, making top-tier computing power truly accessible.

DeepSeek announced the first batch of Agent scores for V4-Flash-Vision-Exp. In the evaluation of 4 multimodal Agents, it achieved 2 wins and 2 losses against Opus 4.8: Agents' Last Exam 27.3 vs 25.7, ZeroBench 35.0 vs 34.0; ApexBench 36.5 vs 39.4, Chartography 64.3 vs 65.0.Compared to the pure text version V4-Flash-0731, the visual version improved from 26.2 to 36.5 in ApexBench, and from 25.2 to 27.3 in Agents' Last Exam. The official note states that the pure text version ignores multimodal content, thus mainly reflecting the capabilities after adding visual input. After adding visual input, the text Agent's capabilities did not significantly decline, with 6 out of 7 text evaluations higher than V4-Flash-0731, such as DeepSWE rising from 54.4 to 59.3 (surpassing Opus 4.8's 58.0), and Toolathlon at 75.9 nearly matching Opus 4.8's 76.2.These results come from DeepSeek's official self-testing, not from a third-party independent ranking; the public Code Agent text task used DeepSeek Harness in minimalist mode, with the inference setting at max.

DeepSeek officially launches the new visual model deepseek-v4-flash-vision-exp. The official API documentation has listed it alongside V4 Flash and V4 Pro, allowing developers to directly input images through the DeepSeek API. The model supports a context of 1 million tokens, with a maximum output of 384,000 tokens, and also supports JSON Output, Tool Calls, Responses API, and Anthropic API.The pricing is directly aligned with V4 Flash. For every million tokens input, when there is a cache miss, the peak period is 3 yuan, and the idle period is 1.5 yuan; cache hits only cost 0.1 yuan and 0.05 yuan respectively. For every million tokens output, the peak period is 9 yuan, and the idle period is 4.5 yuan. Compared to V4 Pro, the same tier pricing is only one-third.Images are not charged separately per image. DeepSeek will convert images into tokens based on their size and bill them together with text tokens. The peak periods are from 9:00 to 12:00 and 14:00 to 18:00 Beijing time, with prices halved during other times.

According to X user MaxForAI, signs related to the DeepSeek visual model have emerged. At 3 PM today, the name deepseek-v4-flash-vision-exp appeared in the code related to DeepSeek Harness, and the latest version of Harness has begun to incorporate native image request adaptation. However, the model will not be included in the official model directory until the model endpoint is ready.Currently, it seems that DeepSeek is testing the visual version of V4 Flash, and the public API has not yet been officially opened. Some users have directly requested the DeepSeek API using the model ID, but it still returns errors, and the official documentation has not listed this model yet. The adaptation path for the client and Harness side has basically been laid out, and DeepSeek V4 Flash is expected to soon have visual capabilities.

On August 21, the B.AI platform announced that Tencent's next-generation foundational model Hy3 is officially launched and is fully open for free on both Web Chat and API. As a hardcore model that adopts a fusion architecture of fast and slow thinking, Hy3 has achieved breakthrough advancements in dimensions such as agent engineering, complex code generation, long-range intent understanding, and resistance to hallucinations, specifically designed for real productivity scenarios. Starting today, all users can log in to the B.AI platform with zero barriers to freely use this top-tier model capability from Tencent, experiencing the strong empowerment that cutting-edge AI technology brings to development and creation at no cost.

Ukraine is seeking to use drones equipped with the Starlink system to strike missile launch silos within 200 kilometers inside Russia due to a shortage of air defense interceptors. This move requires approval from Elon Musk, whose position is uncertain.Communication has faced challenges since the dismissal of former Defense Minister Mykhailo Fedorov. Military experts say Ukraine is struggling to obtain enough interceptors to maintain its air defense strategy.

OpenCode CEO Jay V stated that after the launch of DeepSeek V4 Flash on August 1, the daily Token consumption on OpenCode surged from about 3 trillion to 18 trillion, a sixfold increase within two weeks, approaching the daily processing volume of the entire OpenRouter platform. Its pricing is about 1/350 of Fable, 1/175 of GPT-5.6 Sol, and 1/70 of Claude Sonnet, and it is regarded by users as "AI so cheap it doesn't need to be measured."After raising peak hour prices by 5 times (and 2.5 times during off-peak hours) on August 16, the model's daily Token consumption fell by over 50% from its peak. Jay V believes that the price increase is a passive response to rising GPU costs. Currently, OpenCode is looking for suppliers who can handle traffic at original prices, requiring about 1,000 B300 units to support current demand. DeepSeek Flash has exposed the "impossible triangle" of the AI model market—cheap, powerful, and large-scale stable supply are difficult to achieve simultaneously. After the price increase, OpenCode Go reduced its quota from $60 to $15 and then restored it to $30, introducing a peak and off-peak rate mechanism.

Market news indicates that Broadcom (AVGO.O) is seeking to raise over $60 billion in the latest artificial intelligence debt transaction.

Qatar's Prime Minister and Foreign Minister Al Thani stated at a joint press conference with the Egyptian Foreign Minister that Qatar will coordinate and cooperate with Egypt to promote a final agreement on the Iran issue to end the war and military escalation.The Qatari Ministry of Foreign Affairs emphasized the need to ensure freedom and safety of navigation in the Strait of Hormuz and the Bab el-Mandeb Strait, condemning actions that hinder international shipping.

The B.AI platform DeepSeek-V4-Flash zero-threshold free open event has reported good news again, with the total token throughput across the platform exceeding 35 billion, and the massive high-concurrency practical calls continue to break records, demonstrating the platform's stability under extreme testing.This event is open to all users on the internet with zero threshold, unlimited access, and completely free. Whether it is heavy programming, long text processing, multi-agent collaboration, or complex workflows such as large-scale API calls, B.AI fully covers the computing costs to ensure that every developer and AI enthusiast can enjoy a stable and efficient top model experience. The event is still ongoing and very popular.

On August 19, 2026, the U.S. Commodity Futures Trading Commission (CFTC) issued a request for comments to better understand and regulate the compute derivatives market. CFTC Chairman Michael S. Selig stated that for the U.S. to win the AI race, a robust compute derivatives market is needed, just as the U.S. market once established standards for commodity trading required by the industrial economy, it will also establish corresponding markets for core commodities of the intelligent economy. This request for comments is the first step in establishing clear rules for the U.S. computing market.The request for comments focuses on the scale, liquidity, and other considerations of the compute spot market, as well as issues related to market regulation and manipulation risks, customer protection, perpetual compute futures, and welcomes opinions on all aspects of the compute market. The comment period is 60 days from the publication in the Federal Register.