The Omniscient Bulletin · 2026-06-03
The Omniscient Bulletin — June 3, 2026
Beijing pours state capital into DeepSeek as Anthropic puts AI cyber risk on the policy table
After yesterday's executive-order headline, today is the industrial build-out underneath. Beijing is putting state capital directly behind DeepSeek's first-ever raise, Anthropic is moving AI-enabled cyber threats into the policy literature, agent observability has its first unicorn-scale round, and a new modality is using waveforms in liquids to predict antibody clinical outcomes. None of these will top a cable news hour; all of them will shape the next 18 months.
Industry
DeepSeek nears a $7.4B first-ever raise as Tencent, CATL, and China's national AI fund anchor a $59B valuation
Chinese AI startup DeepSeek is finalizing roughly 50 billion yuan in its maiden funding round, with Tencent and battery maker CATL set to be the largest external investors and the state-backed national artificial intelligence fund also participating. Founder Liang Wenfeng is putting in 20 billion yuan personally, the single largest commitment. The post-money valuation lands between 350 and 400 billion yuan, or $52 to $59 billion. The signal is capital concentration, not just valuation: Beijing is putting state money directly behind the lab that V3 and R1 vaulted to national-champion status.
Policy
Anthropic publishes a year-long map of AI-enabled cyber threats, indexed to the MITRE ATT&CK framework
Anthropic's policy team released today a structured mapping of AI-enabled offensive cyber activity drawn from 832 accounts banned between March 2025 and March 2026, with each incident tagged against MITRE ATT&CK. The point is to give defenders, agencies, and regulators a shared taxonomy rather than anecdote. Key findings: medium-or-higher-risk actors rose from 33% to 56% of the dataset across the two six-month periods studied, and higher-risk actors increasingly concentrate AI use in post-compromise techniques, not initial access.
Agents
Coralogix raises $200M Series F at $1.6B to monitor the AI agents enterprises are pushing into production
Advent and CPPIB led the round, with Greenfield Partners and Brighton Park Capital joining, in a deal Coralogix is openly framing as a bet that observability for autonomous agents becomes its own enterprise category, separate from the existing application-monitoring stack. The thesis tracks Microsoft's Agent Control Specification yesterday: as agents move from pilot to production, governance and visibility are turning into a buyable layer, not a feature. The agent-infrastructure trade is now a venture category.
Research
Apoha exits stealth with $36M betting AI can learn material properties from waveforms in liquid
Singular led the latest round, with Draper Associates joining existing seed backers including Redalpine and Seedcamp. Apoha measures the waveforms substances generate when suspended in liquid and acted on by outside forces, treating those signatures as a new training modality for predicting reactivity, smell, taste, and biological behavior. In a multi-year Boehringer Ingelheim partnership the platform flagged high-risk antibody candidates at greater than 90 percent precision from 8 micrograms of material; on a 236-drug clinical-trial dataset it beat 12 standard industry tests.
Industry
Anthropic launches a Services Track and Partner Hub to formalize the systems integrators around Claude
The Claude Partner Network adds a formal Services Track covering systems integrators, advisory firms, and consultancies, plus a Partner Hub where enterprises can search partners by region, industry, and Claude capability. Coming days after Monday's confidential S-1 filing, the move is structural: Anthropic is building the indirect-sales motion that public-market underwriters expect to see at trillion-dollar valuations, and that OpenAI has been quietly assembling for the past year. Channel strategy is becoming a frontier-lab feature.