Tech Depth and Strategy

No significant updates today.

AI Inference Pricing Shifts

• 🔥 DeepSeek V4-Pro Permanent Cut: DeepSeek locks in its 75% price reduction on V4-Pro permanently, lowering costs...

DeepSeek made its 75% V4-Pro price cut permanent after the May 31 promo ends, citing efficiency gains rather than temporary discounts.

• New rates:...

Engineering infrastructure, not model intelligence, blocks high-value AI work in complex domains. With proper knowledge bases, full data access...

Two independent warnings converge on the same problem: enterprises are shipping agents faster than controls can form.

• Architectural amnesia is...

The move from software-only AI to embodied systems is accelerating through unified simulation tools and sensor data exploitation.

• PhysX-Omni...

India's $931M DeepTech funding surge is overwhelmingly AI-driven, with 84% of startups and over 90% of sector funding linked to AI-enabled innovation....

• AutoResearch survey maps AI shift to workflow-level discovery: literature, hypotheses, experiments, validation
• Systems now span human-steered...

Modern breaches increasingly exploit third-party trust rather than direct intrusions.

• DBIR shows vulnerability exploitation now leads initial...

Torq's acquisition of Jit adds an AI context graph that maps real-time relationships between users, assets, privileges and alerts, enabling smarter...

Three releases highlight production-ready building blocks for AI agents:

• Hermes Agent delivers self-improving skills via markdown workflows,...

AI Research Papers

• KVServe: KVServe introduces a service-aware KV cache compression framework for disaggregated LLM serving, achieving up to...

Zoom's $51M 2023 investment in Anthropic has ballooned to $1.27B, fueled by the startup's rapid ascent.

• Timeline acceleration: Now seeking $30B at...

• Valuation mismatch: DeepSeek's $45B valuation is just 5% of OpenAI's $852B, yet its V4 model matches GPT-5.5 and Claude Opus 4.7 on benchmarks.
  -...

The push to slash LLM serving costs is accelerating with targeted innovations in attention and KV cache handling.

• Full-to-sparse transfer: RTPurbo...

NeuroWeave demonstrates a practical autonomous research agent using Perception → Decision → Action → Memory layers with pure Python and no agent...

Google's rapid Gemini integration is flipping the AI race narrative.

• User growth: Gemini active users doubled to 900 million in one year, matching...

Cyera's $50 million acquisition of Genie Security just five months after founding highlights urgent enterprise demand for endpoint tools that block...

AI agents promise faster incident response through triage, investigation, and bounded fixes under governance, yet they introduce untracked chaos by...

Kawasaki opens a San Jose hub to accelerate physical AI deployment in robotics, partnering with Nvidia, Microsoft, Analog Devices and Fujitsu. The...