[DigitalToday reporter Chi-gyu Hwang (황치규)] Anthropic was hit with U.S. export controls over insufficient safety measures as soon as it released its new Mythos-based AI model, Claude Fable 5. The measures cut off overseas access to Mythos 5 and Fable 5 for anyone other than U.S. nationals, fuelling a strong backlash. For Anthropic, which had sought to accelerate steps toward an initial public offering with Fable 5, prolonged controls could become a major negative factor.

• U.S. government bans foreigners from accessing Anthropic AI 'Fable 5·Mythos 5' • "After an Amazon CEO tip-off, U.S. government blocks overseas access to Anthropic's latest model" • How the White House and Anthropic clashed over export controls on Mythos·Fable

The export controls have also increased interest in sovereign AI and open-source AI models in various countries. Debate over AI and geopolitics is also becoming more active.

• China may have been able to access 'Mythos' • Move to block access to Anthropic's latest model boosts sovereign AI debate • U.S. government restricts access to Anthropic's Mythos... Ministry of Science and ICT: "Checking the facts" • U.S.-China AI war intensifies... both countries label the other's AI a 'security threat'

The export controls apply only to Anthropic's Mythos and Fable 5. OpenAI plans to release a new AI model soon that surpasses its current flagship model GPT-5.5, and attention is on how the U.S. government will respond. Anthropic is reported to be meeting White House officials directly to resolve the export-control issue early.

• OpenAI to unveil new AI model in June, codename '5.6'... countering Anthropic Fable 5 • Anthropic engineers to meet the White House directly... all-out effort to lift export controls on Fable·Mythos

Nvidia CEO Jensen Huang’s visit to South Korea also drew attention. During the trip, multiple partnerships between Nvidia and South Korean conglomerates and startups were announced.

• Jensen Huang wraps up South Korea visit... highlights physical AI cooperation and startup support • LG to strengthen cooperation with Nvidia... expand physical AI, AI infrastructure and mobility

It also compiled moves by companies in South Korea and abroad around AI.

Apple unveiled a next-generation AI voice assistant, Siri, developed in cooperation with Google at its annual Worldwide Developers Conference (WWDC).

• Apple unveils AI Siri after 2 years... bets on conversational AI to take on ChatGPT and Claude

Microsoft unveiled 'Web IQ', a search engine dedicated to AI agents. Web IQ is a search engine for AI agents, not humans. While existing search engines rank results in a format suited for humans, Web IQ provides results in a compressed form that AI agents can quickly grasp while using as few tokens as possible. Database company Tiger Data launched 'Ghost', a managed PostgreSQL service dedicated to AI agents.

• Microsoft unveils AI agent-only search engine 'Web IQ' • Tiger Data launches AI agent-only managed PostgreSQL service 'Ghost'

OpenAI is acquiring startup Ona to strengthen its AI coding tool Codex. Ona provides a preconfigured secure cloud environment that allows AI agents to access tools, systems and work context.

• OpenAI acquires secure cloud startup 'Ona'... expands use of coding AI

Google cut the monthly subscription fee for its AI subscription service 'AI Plus' to $4.99 from $7.99 and doubled storage to 400GB from 200GB.

• Google cuts AI Plus monthly fee to $4.99... AI subscription price war also in the U.S.

Amazon Web Services (AWS) unveiled an AI tool called 'AWS FinOps Agent' to help reduce corporate cloud costs. AWS FinOps Agent is currently available as a public preview.

• AWS unveils 'AWS FinOps Agent' to help optimise cloud costs

Xiaomi open-sourced a terminal-based AI coding tool, 'MiMo Code V0.1.0'. Xiaomi said MiMo Code showed capabilities surpassing Anthropic's Claude Code in a key agentic coding benchmark.

• Xiaomi open-sources terminal-based AI coding tool 'MiMo Code'... claims it beats Claude Code

Industrial AI startup Prometheus, led by Amazon founder Jeff Bezos and former Google executive Vic Bajaj, raised $12 billion in a Series B round at a $41 billion valuation. Physical AI company RLWRLD unveiled its in-house robotics foundation model (RFM) 'RLDX-1' and is stepping up its push into the South Korean market. Trillion Labs is developing 'Industrial World Models' for AI factories based on Nvidia Omniverse libraries and Nvidia Nemotron open models.

• Industrial AI startup Prometheus raises $12 billion... valued at $41 billion • "Automating the last mile of labor with AI... implementing robots with human-level dexterity" • Trillion Labs to develop 'world model' based on Nvidia Omniverse·Nemotron... targeting AI factories

Hancom reorganised BGF Group’s internal structured and unstructured data into a form AI can read and process, based on its AI data processing solution 'Hancom Data Loader' and knowledge search solution 'Hancom Pedia'. KSTEC launched a 'Smart IDP (Intelligent Document Processing) solution' that shortens and automates corporate document processing. LG CNS signed a contract with global AI company Anthropic to adopt 'Claude Enterprise'. Upstage acquired Timely, a generative AI multi-platform company.

• Hancom carries out BGF Group AI knowledge search system project • KSTEC launches AI-based 'Smart IDP solution'... supports document-processing automation • LG CNS steps up Claude-based AX business after OpenAI • Upstage acquires AI agent platform Timely

As more companies worry that AI costs are rising to burdensome levels regardless of size, price has emerged as a heavyweight variable in the AI industry beyond performance. With demand rising for relatively cheaper AI models, leading AI firms such as OpenAI and Anthropic, which aim to pursue IPOs as early as this year, appear to be under pressure to cut prices.

• AI price war begins... OpenAI·Anthropic dilemma • Shaken by cost burdens... will an AI landscape centered on large models change? • As AI costs jump, AI cost-optimisation startups emerge

With hundreds of billions of dollars being poured into AI infrastructure investment, significant changes are also emerging in financial markets. Efforts to financialise capital flowing into AI infrastructure are becoming active. Beyond borrowing against Nvidia GPUs, products that allow trading over computing power costs, including futures contracts linked to GPU rental prices, are expected to emerge soon.

• AI infrastructure financialisation gains traction... major investment banks also eye it

It was confirmed that an AI utilisation report published by consulting firm KPMG included many false cases based on AI hallucinations. KPMG removed the report from its website.

• KPMG AI use-case report includes a batch of false information... due to AI hallucinations