Top stories today:

1. OpenAI launches $100/month ChatGPT Pro tier with 5x Codex usage
2. Anthropic said to weigh designing its own chips, no team yet
3. SpaceX posts ~$5B loss on $18.5B+ revenue in 2025, including xAI
4. Anthropic’s Mythos flagged as risk by Bessent, Powell in bank meeting
5. Amazon’s SpaceX-rival Leo to launch mid-2026 after delays

0. Data and calendar

Listen to our AI-generated podcast summarizing today’s newsletter (beware of hallucinations):

1. OpenAI launches $100/month ChatGPT Pro tier with 5x Codex usage vs. Plus

Almost identical to Claude’s pricing:

• Through May 31, OpenAI is offering up to 10x Codex usage on the new plan.

• The existing $200/month Pro remains available with 20x higher limits than Plus.

• “Both Pro plans include the same core capabilities,” OpenAI said.

• Codex has 3M+ weekly users, up 5x in three months with 70% MoM growth.

Separately, OpenAI is reportedly finalizing a new cybersecurity-focused model to rival Anthropic’s Mythos

• The model could be released to a small set of partners, mirroring Anthropic’s limited Mythos rollout.

• OpenAI also told investors its early compute push gives it an edge over Anthropic.

OpenAI is also said to see ad revenue hit $102B by 2030, with avg. ad revenue per user at $60

• Additionally, Florida’s AG has launched a probe into OpenAI, citing security risks and a mass shooter’s ChatGPT use.

3. Anthropic said to weigh designing its own chips, no team yet

• Designing an advanced AI chip can cost ~$500M, per industry estimates.

• Anthropic’s plans are in early stages and it may still opt to buy chips instead.

• Rivals OpenAI and Meta are also exploring designing their own AI chips.

4. SpaceX said to post ~$5B loss on $18.5B+ revenue in 2025, including xAI