OpenAI Executives Exit Before IPO
OpenAI, which recently closed its latest investment round at a post-money valuation of $852 billion and is targeting a record $1 trillion IPO, has seen yet another executive departure.
Following last week’s announcement from Brad Lightcap, an 8-year OpenAI veteran who shared on X that he would be leaving, Denise Dresser, the former Slack CEO who had been with the company for less than a year, has also departed. In a jargon-heavy LinkedIn post (🤢), Dresser, like Lightcap, cited “new opportunities” as her reason for leaving. With Dresser’s exit, the number of executives who have left OpenAI in 2026 now stands at 12.
The rapid turnover in leadership ahead of the IPO is far from reassuring at a time when OpenAI needs to present investors with a picture of an organized, market-ready company. Add to that OpenAI’s decision last week to buy back $7 billion worth of shares from employee shareholders at the $852 billion valuation, and these departures look anything but normal. Whether the goal of hitting a $1 trillion valuation before the IPO is financially realistic is debatable; the real question is how Sam Altman plans to manage this chaos while preparing the company to go public.
Zuckerberg’s AI Manifesto
Meta continues its (in our view) long-overdue AI moves: Mark Zuckerberg recently published a long, very long AI Manifesto. For those who haven’t read it, or simply don’t have the time, here’s the summary: Zuckerberg argues that AI technology should be made available to everyone and states that Meta will continue to support open-source models. (This brings to mind Zuckerberg offering “free” Facebook access to users in countries without internet connectivity. 😶🌫️) He also discusses AI agents that will allow users to benefit from AI capabilities while preserving data privacy. Of course, after all these promises to democratize AI, he can’t help but note the need for the US government not to slow AI companies’ progress too much through regulation, the importance of maintaining America’s technological edge over China, and the need for greater collaboration between leading AI companies and the government.
In short: AI for everyone, but most of all for American tech companies. 🙃 So why now? Nobody knows, but better late than never.
Data Still Pays
When a company went bankrupt, its physical assets would be sold off; now emails are being added to that list. Google purchased the entire data of bankrupt airline Spirit Airlines, including emails, flight schedules, and Teams messages, for $10 million at auction. It is widely reported that Google plans to use this data, which contains the company’s entire digital footprint, for AI training.
Bonus: Turkey’s AI Action Plan Published in the Official Gazette
Areas of focus include public sector use, AI education, and GPU support. While initiatives like this tend to fall short as technological advances outpace bureaucracy, the plan to educate citizens on AI is a promising sign.
Quick Hits ⚡️
• Tesla’s purpose-built robotaxi, the Cybercab, is on its way. Reuters →
• YouTube’s view count rules have changed: a view now counts as soon as a video starts playing. TechCrunch →
• Relay, an AI workflow automation startup, has shut down. The team is joining Google’s Chrome team. TechCrunch →
• OpenAI launches TeenGPT, amid ongoing controversy over its psychological effects on younger users and numerous lawsuits. OpenAI →
• Anthropic grew its annualized revenue 7x in a single year, reaching $65 billion. CNBC →
• France’s law banning social media for under-15s was struck down by the Constitutional Council. Le Monde →
📩 If you enjoyed this newsletter, feel free to forward it to a friend. Though it’d be even better if they subscribed themselves.
