COGNOSCERE Daily Tech Review — Issue T192 · Wednesday, July 29, 2026

Or visit Intelligence Overview for deeper analysis.

PREPAREMicrosoft Launches Agentic Security Platform Designed to Combat AI-Based Attacks

Microsoft unveiled Project Perception, an agentic security system coordinating red, blue, and green AI agent teams for continuous attack simulation, threat investigation, and automated remediation, alongside MAI-Cyber-1-Flash, its first purpose-built cybersecurity model. The model scores 96% on the CyberGym benchmark — 12 points above Anthropic's Mythos — at roughly half the cost; the platform enters public preview August 3.

Cybersecurity Dive · Cybersecurity · Relevance: 0.9 · Source →

agentic security, AI cybersecurity, vulnerability management, enterprise security, autonomous AI, cloud security

PREPAREClaude Shared Chats Indexed by Search Engines Raise Privacy Concerns

Publicly shared Claude conversations — including those containing resumes, business discussions, code, and sensitive information — appeared in Google and Bing search results, surfacing an enterprise privacy risk where users treat shared-link chats as private. The incident was not an Anthropic system breach but highlights a governance gap: organizations should treat AI share-link URLs as public webpages and establish policies accordingly.

The Cyber Sec Guru · Cybersecurity / AI Governance · Relevance: 0.8 · Source →

AI privacy, data leakage, enterprise governance, AI policy, search indexing, information security

WATCHOrigin Energy Data Breach Affects ~900,000 Customers

Australia's largest energy retailer Origin Energy has confirmed a data breach affecting approximately 900,000 current and former customers, with a hacker reportedly gaining access using a former employee's credentials and threatening to leak data on up to 4.8 million customers unless a ransom was paid. Exposed data includes names, dates of birth, addresses, phone numbers, account details, and partial payment card digits; Australian Federal Police, the Australian Cyber Security Centre, and the Information Commissioner are investigating.

SecurityWeek · Cybersecurity / Data Breach · Relevance: 0.8 · Source →

data breach, energy sector, critical infrastructure, ransomware, credential theft, Australia, PII exposure

ACTReleasing the Model Weights and Technical Report of Kimi K3

Moonshot AI published the full weights of Kimi K3, a 2.8-trillion-parameter Mixture-of-Experts model with a 1-million-token context window, making it the largest open-weight model ever released. The release includes high-performance attention kernels and agent infrastructure, though running the model locally requires approximately 1.4 TB of fast memory.

Moonshot AI · Open-Weight Models · Relevance: 1.0 · Source →

open weights, Chinese AI, LLM, MoE, AI policy, self-hosting, data sovereignty

WATCHIlya Sutskever's Safe Superintelligence Inc. and NVIDIA Announce Long-Term Strategic Partnership

Nvidia announced a long-term strategic partnership with Safe Superintelligence (SSI), Ilya Sutskever's AI safety-focused lab, including an investment reported at approximately $5 billion and priority access to Nvidia's next-generation Vera Rubin GPU platform, which is expected to increase SSI's compute by an order of magnitude. Nvidia entered the deal after obtaining rare access to SSI's closely guarded research, which the company describes as having reached significant milestones ready for large-scale compute.

NVIDIA Newsroom · AI Safety / Venture Funding · Relevance: 0.8 · Source →

AI safety, alignment, superintelligence, GPU compute, venture investment, AI infrastructure

PREPARETech Sector Pours $1T Into AI and Sends Customers the Bill

Technology companies are projected to spend roughly $1 trillion on AI and cloud infrastructure in 2026, with the costs increasingly passed to enterprise customers through higher subscription fees, bundled AI add-ons, and usage-based pricing. CIOs are reporting significant budget pressure as vendors restructure pricing around AI capabilities, often bundling features customers have not yet adopted.

The Register · Risk Assessment / IT Economics · Relevance: 0.9 · Source →

IT spending, cloud pricing, enterprise budgets, AI ROI, vendor lock-in, SaaS pricing, CIO

PREPAREWhere Should Your Company's AI Brain Live?

As enterprises embed institutional knowledge and decision-making workflows into AI systems, they risk deep vendor lock-in before consciously choosing where that infrastructure should reside. The article compares model providers, incumbent SaaS platforms, closed AI startups, and open-source alternatives, arguing that organizations should retain control over AI processes central to competitive advantage while using managed tools for more generic workloads.

TLDR IT · Decision Frameworks / AI Strategy · Relevance: 0.8 · Source →

AI strategy, vendor lock-in, enterprise architecture, build vs buy, data governance, AI infrastructure, decision framework

WATCHGartner Forecasts Worldwide IT Spending to Grow 14.2% in 2026, Totaling $6.37 Trillion

Gartner raised its 2026 global IT spending forecast to $6.37 trillion — a 14.2% year-over-year increase — driven primarily by a 62.5% surge in data center systems spending to $822 billion and 29.3% growth in IaaS. The update reflects accelerating AI infrastructure investment by hyperscalers, though hardware supply constraints and rising costs are pressuring enterprise budgets as vendors bundle AI into subscriptions.

Gartner · Strategic Planning / Market Intelligence · Relevance: 0.9 · Source →

IT spending, cloud computing, data center, AI infrastructure, market forecast, enterprise budgets, hyperscale