Flash News

TREE NEWS update: Nvidia CEO Jensen Huang said AI safety is critical but does not require new legislation from the US Congress. Speaking on AI safety, Huang said companies should set their own pace until they are confident their products can win market acceptance. He did not call for new federal rules governing artificial intelligence.

TREE NEWS update: Google released Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, which it called its best conversational AI models to date. The company also announced partnerships with Salesforce, Genspark and Lumeris around the launch. No pricing, availability dates or further technical details were disclosed in the announcement.

TREE NEWS update: OpenAI is considering building a data center in Canada and expressed support for Prime Minister Mark Carney’s plan, the company said. The announcement signals potential expansion of OpenAI’s computing infrastructure into Canada. No site, capacity, investment figure or timeline was disclosed.

TREE NEWS reports: Listed bitcoin miners’ weighted average pre-tax cash cost of production was about $75,500 per coin in Q2 2026, above the roughly $58,400 BTC price at quarter-end, pushing the sector below cash break-even. Hash price fell to a record low of $27.7 per PH/s per day in June. All-in costs varied widely, from about $76,529 at ABTC to $353,168 at CIFR, and several miners have shut down hashrate or shifted to AI and high-performance computing.

TREE NEWS reports: Publicly listed Bitcoin miners produced bitcoin at an average ex-tax cash cost of roughly $75,500 in the second quarter, CoinShares said, a level that makes a price recovery unlikely to draw AI-focused miners back to BTC mining. The figure reflects the all-in cost of production for listed miners after taxes.

TREE NEWS update: Meta has launched Meta One, a new subscription service. The offering marks the company’s latest move into paid consumer subscriptions. No pricing, launch markets or included features were disclosed in the announcement.

TREE NEWS update: US House Speaker Mike Johnson said he will not allow the United States to lose its competitive edge in artificial intelligence and that AI development cannot be paused. Johnson said AI company executives are expected to visit the White House within the next week.

TREE NEWS update: Anthropic has released a beta version of a Claude plugin for Salesforce, extending its AI assistant into the customer-relationship-management platform. The beta rollout gives Salesforce users access to Claude inside their existing workflows, the company said. No pricing, general-availability date or user figures were disclosed.

TREE NEWS update: US House Speaker Mike Johnson said he discussed AI risks with President Donald Trump on the evening of the previous day. Johnson gave no further details on the content of the conversation or any policy measures discussed. The remarks were made without specifying what specific risks were raised or whether any action was agreed.

TREE NEWS update: Huawei supervisory board chairman Guo Ping told new employees the company will pursue a differentiated AI strategy by business scenario, building customer-facing compute and platform capabilities in computing and ICT with the goal of becoming Nvidia, while advancing large-model and agent capabilities in terminals, intelligent driving and cloud. Huawei’s large-model strategy stays tightly tied to its own businesses, he said.

TREE NEWS update: Cisco and Nvidia have expanded their Splunk agentic AI agreement. The two companies broadened the arrangement covering agentic AI capabilities tied to Cisco’s Splunk platform, extending a partnership that pairs Nvidia’s AI technology with Splunk’s data and security operations. No financial terms or timeline were disclosed.

TREE NEWS update: Huawei supervisory board chairman Guo Ping said the company’s future core strategy is “focus,” deepening its work in connectivity and computing, with no plans for business expansion. He made the remarks during a recent discussion with new employees. Guo currently serves as chairman of Huawei’s supervisory board.

TREE NEWS update: OpenAI said it will work with Anthropic and Google on AI safety issues. The three companies are set to coordinate on the topic. No further details on the scope, structure or timeline of the cooperation were disclosed.

TREE NEWS reports: DeepSeek machine learning systems (MLSys) engineer Liu Sheng published a long essay titled “I Had to Bury My Talent in Yesterday,” writing that rapid advances in AI capability mean he may not lose his job but will likely have to change profession. The essay reflects the career anxiety AI progress is creating even among top engineers at AI companies.

TREE NEWS update: Meta Platforms plans to begin deploying a new in-house AI chip, MTIA 450 (codename Arke), in its data centers in the first half of next year to cut costs and energy use in running AI models. The company is currently testing the third-generation chip, while the next version, codename Astrid, will finish design in about a month and reach data centers by late 2027. Engineering VP Yee Jiun Song said Meta expects the chips to beat Nvidia processors on performance per watt and per dollar.

TREE NEWS update: Huawei unveiled the world’s first 3D data center at the 2026 AIDC Industry Development Conference in Wuhu, Anhui. Rotating chairman Wang Tao said 100,000-card-plus Ascend super-node clusters will become the baseline configuration, as 10-trillion-parameter models and surging token demand raise AI infrastructure requirements. The 3D data center uses layered decoupling and vertical stacking to support what he called a supercomputing factory for Ascend super-node clusters.

TREE NEWS update: OpenAI has endorsed a House proposal to establish safety evaluations, the company said. The measure would set up assessment requirements for AI models. No further details on the bill’s scope, timing or the specific provisions OpenAI supports were provided.

TREE NEWS reports: Micron Technology expects its 512GB RDIMM memory modules to enter mass production in the second half of 2027, the company said. The timeline points to availability of the high-capacity server memory part roughly two years out, with no pricing or customer details disclosed.

TREE NEWS reports: Atlantic Council experts say corporate AI safety pledges require enforceable standards, warning that U.S.-China distrust is limiting the prospects for an international AI agreement. The analysis argues that an AI slowdown could unravel under commercial and geopolitical pressure.

TREE NEWS update: Huawei supervisory board chairman Guo Ping told new employees the company’s goal in ICT and computing is to become the Nvidia of China,000-word internal Q&A published on Huawei’s Xinsheng Community. Guo said Huawei focuses on connectivity and computing, and that its large-model strategy is tightly tied to its own business, with investment maintained but deployed differently across business scenarios.