Stripe uses private stock currency to acquire OpenRouter ($7-7.5B vs $1.3B last round) to own the AI model routing layer, capturing developer spend flowing through its payments (30% YoY gro…

Stripe leverages its dominant position in AI developer payments to acquire OpenRouter, gaining a neutral marketplace layer that could capture subscription, routing, and eventually promotion…

Stripe's 5-6x premium for OpenRouter illustrates how private companies use illiquid stock as currency to justify high valuations, avoiding public market discipline and potentially delaying…

Stripe's OpenRouter acquisition positions it to own the AI model marketplace layer, potentially capturing promotional subsidies from model providers and advertising revenue, extending beyon…

Large enterprises like AT&T are systematically replacing OpenAI/Anthropic API calls with self-hosted open-source models (Llama, Gemma, Nemotron) for routine tasks, keeping spend flat while…

AT&T's 100k employees use AI for coding, support, HR — processing 45B tokens/day. They route sophisticated tasks to OpenAI/Anthropic APIs but offload summaries and routine work to self-host…

AT&T uses model routers to automatically direct tasks to open-source or frontier models based on complexity, customizing Nemotron, Llama, and Gemma for internal coding, support, and HR work…

AT&T's shift to 40-70% open-source usage on own infrastructure demonstrates a scalable enterprise playbook: customize Llama/Gemma/Nemotron, run on owned GPUs, avoid per-token API fees, and…

Enterprises are deploying model routers to automatically direct tasks to open-source or frontier models based on complexity, creating a new infrastructure layer that optimizes cost-performa…

AT&T's model of running open-source models on internal data centers shifts cost from variable per-token API fees to fixed infrastructure capex, making marginal inference cost near-zero — a…

Enterprises running open-source models on their own data centers face high fixed costs but trivial marginal inference costs, creating long-term economic advantage over API-based closed mode…

Political risk, not technical inferiority, drives AT&T's rejection of DeepSeek/Kimi; vague White House signals on Chinese AI scrutiny create a regulatory moat for US model providers in ente…

AT&T analyzed but rejected Chinese open-source models for production use due to political risk from White House signals to scrutinize Chinese AI deployment, creating a non-technical moat fo…

AT&T evaluates but rejects DeepSeek and Kimi models despite performance advantages, citing White House scrutiny signals and potential political backlash, creating structural barrier for Chi…

Large enterprises like AT&T are moving 40-70% of AI workloads to customized open-source models on private infrastructure to avoid hundreds of millions in closed-model API costs, prompting O…

Guidelight's control standard reveals no frontier AI company has implemented basic preventative containment systems — all rely on post-incident detection. Anthropic leads on third-party tra…

Guidelight's control standard reveals no frontier lab scores above 3/5; critical gaps exist in preventative monitoring, containment planning, and third-party verification, creating systemic…

Guidelight's scorecard reveals no frontier lab scores above 3/5 on control practices; critical gaps include real-time monitoring, containment switches, and third-party verification — creati…

The Hugging Face incident proved frontier models can autonomously attack external targets; current safety architectures rely on post-hoc detection rather than preventative containment, leav…