AI Industry Development & Policy Dynamics

According to Ark Invest Tracker, Nvidia CEO Jensen Huang stated today (Oct 24) that AI token costs decline 30 to 40 times annually, significantly reducing AI usage expenses. As costs drop, customers' returns from AI scale exponentially, Huang

According to Fortune, SpaceX and Chinese tech firms are competing to deploy space-based AI computing infrastructure to address growing power demands. At a Fortune China 500 Summit held in Chengdu on October 22, space computing emerged as a focal point for investment and development. SpaceX plans to

Stablecoin issuer Tether announced on October 24 the launch of its open-source Mining Development Kit (MDK), partnering with 15 global Bitcoin mining companies. The framework aims to improve interoperability across Bitcoin mining infrastructure and enable AI-based mining automation. Tether stated th

Nvidia's role as a financier for AI infrastructure buildouts faces its first major stress test this month as credit markets tighten for key partners. The chipmaker has extended $36 billion in off-balance-sheet AI cloud agreements and guaranteed up to $105 billion for an OpenAI data center project in

According to CNBC, Nvidia's role as a financial backer for AI infrastructure companies is facing its first major stress test this month as credit markets tighten. CoreWeave, Nebius, and IREN—key neoclouds in Nvidia's portfolio—have all seen shares decline at least 30% in October, with CoreWeave's cr

According to a recent interview with YouTuber Matthew Berman, OpenAI CEO Sam Altman warned that unchecked AI development could exacerbate social inequality and called for global mechanisms to limit the speed of recursive self-improvement (RSI). Altman emphasized that AI's goal should be enabling

According to PANews, ByteDance has consolidated its AI short-drama creation platform Xiaoynque into Jimeng, an intelligent content creation platform, on Oct 24. The product maintains a separate entry point following the

According to Axios on October 24, U.S. lawmakers from both parties, congressional staff, and Washington officials criticized AI executives for becoming a political burden and transmitting conflicting messages to the public. The officials expressed concern that AI leaders focus excessively on

According to Bloomberg, Nvidia and TCV joined a Series B extension financing for Higgsfield, an AI video startup based in San Francisco, during Nvidia CEO Jensen Huang's visit to Kazakhstan. The extension size and valuation were not disclosed. In August, Higgsfield announced a US$400 million Series

According to CNBC, TikTok's US operating entity (USDS Joint Venture) is seeking multiple gigawatts of computing capacity over the coming years, targeting 4 gigawatts by the early 2030s, according to two people familiar with the discussions. The company runs mainly on Oracle Cloud Infrastructure

According to Moonshot AI, on October 24, the China-based company launched Kimi K3.1 Pro, an open-weight model with 3 trillion parameters and a 1 million-token context window, alongside Kimi K3.1 Pro Highspeed. Both models are now available on Kimi.ai and the company plans to release full model weigh

According to Reuters, Anthropic plans to complete its initial public offering in mid-November at a $2 trillion valuation, potentially surpassing SpaceX's $1.77 trillion IPO earlier this year. However, the timing faces pressure as recent AI sector declines and Securities and Exchange Commission

According to Bloomberg, OpenAI announced on October 23 that it had resumed training its most advanced AI models after a pause last month, but would keep those models offline during training until safeguards are deemed adequate. The move follows two significant security incidents. In late

According to Guru Club, Tencent's WorkBuddy announced the integration of the Kimi K3.1-Pro model on October 24. The platform simultaneously rolled out the Kimi K3.1 Pro Highspeed variant, offering the same intelligence level with output speed approximately 2-3 times faster. The feature is

According to Goldman Sachs, the latest forecasts show up to 9% of workers—approximately 15 million—could be displaced by AI, up from previous estimates of 6-7%. The bank's chief economist Jan Hatzius noted that AI adoption remains in its early stages, with fewer than 25% of companies currently

According to BlockBeats, on October 24, Rich Privorotsky, head of Europe fixed income and equities trading at Goldman Sachs, said the bank's prime brokerage business experienced record-scale tech stock selloffs

According to Bloomberg, SpaceX's proposed $40 billion AI funding round has faced obstacles recently as credit market volatility makes financing negotiations more complex. The company was planning to use a special purpose vehicle (SPV) structure to move related debt off its balance sheet and

According to Amazon CEO Andy Jassy, AWS's core AI computing capacity for 2026 is fully sold out, with 2027 nearly sold out and most of 2028 capacity already pre-booked. Jassy made the remarks in a podcast with Andreessen Horowitz published this week. AWS will continue to reserve computing

According to Benzinga, Micron announced at a recent technical conference a roadmap for next-generation vertical high-bandwidth memory (vHBM) with a 2TB/s bandwidth target and plans to begin mass production in 2028. The architecture directly stacks DRAM on top of processors to shorten data

According to Beating, DeepSeek recently disclosed that its inference framework DeepInfer processes 1.55 quadrillion tokens daily. Based on DeepSeek API off-peak pricing, the company said a single GPU hour generates approximately $38 in theoretical billing revenue. If operated continuously