Friday, March 20, 2026
Top story
AI startups are eating the venture industry and the returns, so far, are good
AI startups now dominate venture funding with 41% market share, signaling a fundamental shift in where smart money is betting and what returns early investors are seeing.
The briefs
Energy infrastructure has become the hidden constraint limiting AI scaling, creating massive investment opportunities beyond traditional AI companies. Smart money is pivoting to energy tech as the picks-and-shovels play of the AI boom.
As AI becomes critical defense infrastructure, regulators and military are stress-testing whether companies can be forced to weaponize or disable models mid-conflict. Anthropic's technical rebuttal reveals the real fault lines in AI governance and national security policy.
Blue Origin's orbital datacenter constellation represents a fundamentally new approach to AI compute distribution, bypassing traditional terrestrial infrastructure constraints and potentially disrupting how enterprises access GPU capacity at scale.
Sashiko demonstrates AI-assisted code review moving from enterprise tools into critical infrastructure. This matters because Linux kernel security directly affects billions of devices—if AI can reduce human review burden while catching edge cases, it reshapes how we maintain the world's most important open-source project.
Regulatory and operational reality check on AI deployment. A major law enforcement agency is deprioritizing facial recognition technology due to documented racial bias, signaling that safety governance failures have real institutional consequences—and raising questions about similar deployments elsewhere.
Salesforce is consolidating specialized talent from consumer productivity into its enterprise AI agents platform. The Clockwise team's expertise in scheduling intelligence and automation signals Agentforce's next capability frontier.
Palantir's developer conference signals a strategic shift toward AI-driven defense and battlefield applications, positioning the company as the dominant player in military-grade AI infrastructure while competitors focus on consumer and enterprise use cases.
A unified federal AI regulatory framework could reshape compliance costs and competitive dynamics for AI builders, but state-level resistance suggests a prolonged governance battle ahead.
Trivy is widely used by AI infrastructure teams and enterprises building LLM systems to scan for vulnerabilities. A live compromise of this dependency scanning tool represents a material security governance risk for AI labs and raises questions about the fragility of open-source tooling that underpins AI development pipelines.