Alibaba on Aug. 14 open-sourced Qwen3.8-27B, a 27-billion-parameter dense native multimodal model released under Apache 2.0, with no monthly active user or revenue threshold attached to its license. The company said the model supports a native 262K context window and can be extrapolated to 1 million tokens through YaRN. In Alibaba’s published model card, Qwen3.8-27B was compared against Claude Opus 4.6 Max across 14 benchmarks with side-by-side scores, winning 9 and losing 5. Its largest gains were in visually grounded and agent-style tasks such as MathVision, CharXiv and AndroidWorld, while every loss was concentrated in pure reasoning, code generation, terminal coding and related long-chain tasks. Alibaba also said a 4-bit quantized version can run in roughly 14 GB to 17 GB of VRAM, enough to fit model weights on a consumer GPU such as an RTX 4090, though that figure excludes KV cache, concurrent agent sessions, runtime buffers and multimodal components. All benchmark figures cited in the model card came from Alibaba’s own release, and no third-party independent replication was provided in the input.

Lenovo's CEO announced on Aug 13 that the company will partner with Nvidia (NVDA.O) to release AI PCs powered by RTX chips later this year. The news was reported by Jinshi and carried by BlockBeats.

The U.S. Securities and Exchange Commission is set to hold a public meeting on Aug. 14 to consider a proposed "Regulation Crypto" framework that would let some crypto projects raise capital without completing a full securities registration process. If advanced for public comment, it would become the SEC’s first formal, durable rulemaking effort aimed specifically at the crypto sector. The proposal would also outline a route for projects to exit SEC oversight once developers stop actively managing the network and the project becomes decentralized. SEC Chair Paul Atkins has previously said the exemption period could last as long as four years, though no fundraising cap was disclosed. On the litigation front, a federal judge in Connecticut ruled that Kalshi’s sports-event contracts are not swaps under the Commodity Exchange Act, meaning the Commodity Futures Trading Commission did not obtain exclusive jurisdiction on that basis. The decision lands as Kalshi remains under pressure in New York, where the CFTC said it used “emergency powers” to require the prediction-market operator to continue operating after the company sought assistance following a lawsuit from New York Attorney General Letitia James. Elsewhere, the market snapshot showed broad declines among major tokens over the past 24 hours, while project, funding, security and AI headlines spanned Bitwise layoffs, a reported COLDCARD Mk3 wallet flaw, Hyperliquid and Robinhood Chain user metrics, and a string of new financings across crypto, AI and financial infrastructure.

Bitcoin News shared on X a technical analysis by @KLoaec finding that some vulnerable Coldcard Mk3 wallets may derive from only ~4.5 million RNG starting states, searchable in about 3 seconds on a single RTX 4090. Even with extra per-wallet uncertainty, an attacker could finish the search on a high-end GPU in ~50 minutes. The flaw could cause different devices to generate identical mnemonics; assuming 30,000 Mk3 units, roughly 120 pairs may produce the same random stream.

Meta has reopened its open-weight AI strategy with the release of Muse Glimmer, a model whose underlying parameters can be downloaded and modified by developers. The company said it will also release the weights for the more capable Muse Spark 1.2 in the coming weeks, marking a clear shift back toward the approach it once used to distinguish itself from rivals. CEO Mark Zuckerberg backed the move in a post on Meta’s website, arguing that powerful and free AI should reach billions of people rather than remain concentrated in large institutions. He also defended model distillation and said people should retain the ability to learn from observable information, a position that cuts against recent complaints from OpenAI and Anthropic over how their closed models’ outputs are used. Muse Glimmer is Meta’s first open-weight model since Llama 4 and its first to ship under the Apache 2.0 license. The model supports text and image input, carries a 128K context window, and is aimed at local agent use on personal devices. Meta’s release also comes with a broader commercial and infrastructure angle, as the company plans to spend as much as $145 billion on AI infrastructure this year while building out related cloud services.

Nvidia’s GeForce RTX 50-series graphics cards have posted another round of retail price increases in the U.S., driven by sharply higher costs for key components, according to the source report. The biggest moves were concentrated in midrange Blackwell models, raising the cost of upgrades for PC DIY builders and creators. Retail tracking data cited in the report showed the GeForce RTX 5060 at a median retail price of $469.99, well above its $299.99 suggested retail price. The RTX 5060 Ti 8GB reached a median of $529.99 after an increase of about 13% to 15%, while the RTX 5060 Ti 16GB rose from $579.99 to $799.99, a gain of more than 39%. The GeForce RTX 5070 was listed in an $850 to $900 range after a 29% jump from the prior reading. Higher-tier models including the RTX 5070 Ti and RTX 5080 saw little change in this survey, while the entry-level RTX 5050 held near $300. AMD’s RDNA 4 lineup recorded smaller price changes, though supply conditions shifted across several products, including fewer available Radeon RX 9070 16GB variants and limited stock for the RX 9060 XT 8GB.

On August 5, Jensen Huang accidentally followed @TradeSager, a community member of the BSC-based stock meme coin RTX, on X. The token spiked 7-fold within minutes before quickly retracing, with its market cap falling to $1.51 million, according to GMGN. BlockBeats cautioned that the token's price could remain highly volatile due to community sentiment and potential follow-up actions by Huang. Investors are advised to exercise caution. The accidental follow, possibly triggered by a hashtag-related tweet or unfamiliarity with crypto community tagging rules, briefly fueled speculative trading. The episode highlights how social media actions by prominent figures can temporarily move low-cap meme tokens.

AMD has notified its add-in-board partners that it plans to raise GPU memory kit pricing by at least 10% in August, according to ABMedia, citing Wccftech and a screenshot obtained by ChannelGate on July 31, 2026. The report says the move will hit board makers’ component costs first and could later show up in retail graphics card prices, with the RX 9000 series expected to be among the most affected product lines. ABMedia said AMD had been considering a price increase about a month earlier but held back because NVIDIA still dominates the discrete GPU market in many regions. The company reportedly moved only after NVIDIA raised pricing on its RTX 50 series, a step that was followed by price increases in China. The article ties the change to sharply higher upstream DRAM costs this year, with some DRAM items nearing a doubling in price. Because AMD’s RX 9000 series and prior-generation products use GDDR6, the report says the increase may affect multiple generations. ABMedia also pointed to AMD’s recent Helios AI rack pricing as a sign that the company’s broader pricing strategy may be shifting beyond pure low-price competition.