- Bay Area Times
- Posts
- Anthropic said to cross $19B ARR amid Pentagon feud
Anthropic said to cross $19B ARR amid Pentagon feud
Bay Area Times
March 04, 2026
Brought to you by:
Top stories today:
- Anthropic said to cross $19B ARR amid Pentagon feud
- Anduril said to seek $4B raise at ~$60B valuation led by Thrive, a16z
- GPT-5.3 Instant debuts with improved web search for ChatGPT users
- Google launches Gemini 3.1 Flash-Lite, faster model
- U.S. Clarity Act urged by Trump as he attacks banks over GENIUS
- OpenAI considers NATO AI deployment on âunclassifiedâ networks
0. Data and calendar
Listen to our AI-generated podcast summarizing todayâs newsletter (beware of hallucinations):
1. Anthropic said to cross $19B ARR amid Pentagon feud
Strong adoption of Anthropicâs AI models and Claude Code has helped drive the growth.
Anthropic has higher retention than OpenAI, CEO Dario Amodei said at an investor conference on Tue., citing its âmissionâ to reinforce staff loyalty.
OpenAI reached $20B in ARR in 2025.
Relatedly, Lockheed Martin to follow federal Anthropic ban as contractors comply with DoD order, lawyers said.
2. Anduril said to seek $4B raise at ~$60B valuation led by Thrive, a16z
Anduril aims to modernize the military with AI and advanced manufacturing, founder Palmer Luckey said.
Feb.: Anduril was reportedly in talks to raise $8B, up from a reported $2B in January.
2025: Revenue expected to double to ~$2B, driven by U.S. and allied defense contracts.
Most teams know AI can help. The hard part is building workflows that actually run reliablyâconnecting agents, automations, and LLMs into systems people use every day. You need someone who can design for the messy middle:
AI agent design and deployment
No-code and LLM workflow integration
Process and dashboard automation
This is the kind of talent you get with Athyna Intelligence: vetted LATAM AI specialists working in U.S.-aligned time zones.
*Disclaimer: We have equity in Athyna.
4. OpenAI launches GPT-5.3 Instant with improved web search for all ChatGPT users
GPT-5.3 Instant performs between GPT-5.1 and 5.2 on safety tests, with some regressions in sexual and self-harm content:
