newsroom
Enterprise AI Adoption · AI ROI is a throughput problem solved by investing in context infrastructure, not raw model access
now playing · Enterprise AI Adoption
Open Source AItailwindscore 9/10arvind jain
Open source models reach frontier parity for 90%+ of enterprise use cases, driving massive cost-driven adoption
Models like GLM 5.2 have closed the gap to within 3 months of frontier capabilities; enterprises are shifting to open source primarily for cost control (10x cheaper), with majority of workl…
Model layer commoditization and consumption pricing will compress frontier lab margins and break Microsoft bundling
Three-way frontier lab competition plus open source pricing pressure (order of magnitude cheaper) makes standalone model business less lucrative; consumption-based pricing lets enterprises…
AI ROI is a throughput problem solved by investing in context infrastructure, not raw model access
Enterprises waste tokens brute-forcing context assembly; the winning approach is building semantic context layers (like Glean) that feed agents the right information, making AI faster and c…
Contrarian view: AI will grow teams not shrink them, as 10x productivity demands 10x output
Arvind argues companies that cut headcount will lose to competitors who keep talent and use AI to build 10x better products; Glean plans to grow from 1,000 to 5,000 employees. Composite rol…
100% of code AI-generated but human review becomes bottleneck; triage agents handle 95% of issues at high inference cost
AI has shifted coding bottleneck from writing to review; Glean's triage agent automates 95% of production issue handling but costs $1M/month, showing inference economics must improve for fu…
Chinese open source models dominate usage rankings; US mobilizing Nvidia-backed open source response
Top 6 models on OpenRouter are Chinese; US lacks capital-efficient open source model development. Nvidia and other motivated parties are now investing heavily to build US open source altern…