AIDB daily briefs

The Stockholm-based AI-powered preventive health scanning company secured one of the largest health-tech funding rounds in European history, targeting New York City as its first American market.

President Lee Jae-myung unveiled a decade-long, 1,350 trillion won public-private plan spanning semiconductor fabs, AI data centers, and humanoid robotics, positioning South Korea as a full-stack artificial intelligence hardware power.

An independent expert panel finds that no major artificial intelligence developer earns better than a C+ on safety, transparency, and governance, with three labs receiving failing grades.

The Beijing-based lab backed by Alibaba launched Kimi K3—a 2.8-trillion-parameter Mixture-of-Experts large language model—days before the 2026 World Artificial Intelligence Conference in Shanghai, claiming the largest open-weight release in AI history.

Wang's departure, with several OpenAI colleagues expected to follow, underscores a broader talent migration from general-purpose frontier labs toward specialized, domain-specific artificial intelligence applications.

An independent expert panel found no frontier AI developer scoring above a C+ across six safety domains, with multiple labs quietly rolling back prior risk commitments amid intensifying competition.

President Lee Jae-myung unveiled a decade-long, 1,350 trillion won plan on June 29, 2026, anchored by Samsung and SK Hynix, to build four new memory fabs and 8.4 gigawatts of AI datacenter capacity.

The San Francisco AI drug-discovery startup closed its third financing round in under a year, backed by Index Ventures, Sequoia Capital, Kleiner Perkins, and returning investor OpenAI, as commercial deployments at Eli Lilly, Pfizer, and Novartis validate its generative molecular-design platform.

The Stockholm-based full-body scanning company, co-founded by Spotify's Daniel Ek, closed its largest funding round to date and will open its first US clinic in New York City later in 2026.

Taiwan Semiconductor Manufacturing Company's second-quarter 2026 revenue surged 36% year-over-year to an all-time high of $39.6 billion, with its 3-nanometer process node and advanced packaging capacity fully booked through year-end as artificial intelligence infrastructure spending overrides traditional semiconductor demand cycles.

Google's Agent2Agent (A2A) open protocol, now backed by Microsoft, Salesforce, and more than 150 organizations under Linux Foundation governance, is emerging as a parallel standard to Anthropic's Model Context Protocol (MCP) for enterprise AI agent interoperability.

Meta's first model from its newly formed Superintelligence Labs arrives as a proprietary, natively multimodal reasoning system with built-in multi-agent orchestration, marking a deliberate pivot toward the agentic AI workloads where token consumption is growing fastest.

Facing a $1.25 billion monthly compute contract and escalating inference costs, Anthropic is in early-stage discussions with Samsung to manufacture a purpose-built AI accelerator chip—marking the last major frontier lab's entry into custom silicon.

OpenAI's three-tier GPT-5.6 model family entered a government-coordinated limited preview before broader rollout, marking a new pattern for frontier Large Language Model releases under a June 2026 White House executive order.

A structured government-access framework for Anthropic's restricted frontier cybersecurity model is taking shape across allied nations, following diplomatic negotiations between US Treasury and Japanese officials.

Intel's largest-ever European foundry commitment upgrades Ireland's Leixlip campus to scale output of Xeon server chips as AI-driven demand for advanced processors outpaces supply.

The oversubscribed round, led by BMW i Ventures, backs a behavioral-intelligence platform that targets the critical window between a missed payment and a write-off across telecom, auto lending, and financial services.

Anthropic has pushed back the free-tier deadline for its flagship large language model a third time, now to July 19, 2026, as the company works to stabilize compute capacity following a disruptive government export-control episode — and as rival OpenAI rolls out GPT-5.6.

Following a 13-day limited rollout coordinated with U.S. federal officials, OpenAI has opened its GPT-5.6 model family—comprising flagship Sol, balanced Terra, and fast Luna—to ChatGPT subscribers, API developers, and Codex users worldwide.

Mountain View-based Commure secured $70 million led by General Catalyst at a $7 billion post-money valuation, underscoring intensifying investor conviction in enterprise-grade AI for clinical and revenue-cycle workflows.

The San Francisco-based AI cloud company closed one of the largest open-source infrastructure rounds on record, with Aramco Ventures leading a syndicate that includes Nvidia, General Catalyst, and Vista Equity Partners.

OpenAI's head of safety systems is departing after the company merges its safety and research organizations under a single leader, marking the sixth senior safety exit in two years.

SpaceXAI released Grok 4.5 on July 8, 2026, pitching the large language model as a faster and more affordable alternative to Anthropic's Claude Opus line for coding and agentic workloads.

OpenAI's three-tier GPT-5.6 model family became the first frontier AI release to undergo a formal pre-deployment government safety review under President Trump's voluntary AI security executive order before reaching general availability on July 9, 2026.

Google unveiled the Gemini Omni multimodal model and a 24/7 cloud agent called Spark at I/O 2026, as the Gemini app reached 900 million monthly users.

Sequoia-co-led financing values the Claude maker above OpenAI as the company projects $10.9 billion in Q2 2026 revenue and its first quarterly operating profit of $559 million.

The Anthropic Labs tool, powered by the Claude Opus 4.7 vision model, lets subscribers generate prototypes, slide decks, and design systems through conversational prompts.

The ChatGPT maker submitted draft registration paperwork to the SEC with Goldman Sachs and Morgan Stanley leading a Q4 2026 listing, days after a federal court dismissed Elon Musk's lawsuit.

Anthropic's latest flagship Large Language Model (LLM) ships at the same $5/$25 per-million-token price as Opus 4.7, with measurable gains on coding benchmarks and a new parallel-subagent workflow tool.

The federal law's 48-hour takedown mandate for nonconsensual intimate imagery, including AI-generated deepfakes, entered full enforcement on May 19, 2026, marking the first federal U.S. statute directly regulating harmful uses of artificial intelligence.