Elon Musk and Sam Altman are officially in their courtroom era. Musk is testifying in his lawsuit against OpenAI this week, and the Tech Tea is piping hot.

There’s Elon sparring with defense attorneys: “Your questions are not simple. They’re designed to trick me.”

The angry text to Greg Brockman pre-trial: “By the end of this week, you and Sam will be the most hated men in America. If you insist, so be it.”

The the long-running Twitter feud:

Musk even got emotional and told the court last week he was “a fool” for funding Sam in the first place 💔

The drama is fun. The sour grapes are real. But there is a serious point underneath all of this.

Musk invested in OpenAI back when it was a nonprofit, founded around the mission of building AI “for the benefit of humanity.” Then, OpenAI created a for-profit arm, raised billions, partnered deeply with Microsoft, and became one of the most valuable companies in the world.

Musk’s argument is basically: this is not what I signed up for.

It’s a fair question to ask:

Can a company start as a nonprofit “for humanity,” raise money and trust on that premise, then evolve into a profit-driven AI giant without consequence?

Now, Musks’s motives deserve scrutiny too. He is clearly bitter about losing control. And yes, he is conveniently building a direct competitor with xAI. And, I mean, he’s Elon.

Tune in to my Tea Take below for the full scoop.

1. The Pentagon signs AI deals with 7 tech companies to integrate AI into classified military operations. 

My Tea: Notably missing from the list: Anthropic.

2. The Elon Musk vs. OpenAI lawsuit kicks off in court.

My Tea: The courtroom drama has been spicy, but this lawsuit is about something much bigger: who controls the future of AI. See First Pour, above.

3. Meta buys Assured Robot Intelligence to push deeper into humanoid robots.

My Tea: One thing about Zuck: he’s always reinventing himself. slow clap.

4. Tech layoffs surge again, with ~45,000 jobs eliminated in April alone.

My Tea: The numbers are getting hard to ignore. Nearly 100,000 tech jobs have been cut so far this year.

5. China’s DeepSeek launches a new AI model that rivals top US models at a lower price.

My Tea: China is doing what China often does: making expensive things cheaper and faster… this time with serious geopolitical implications.

Excited to share that I was interviewed by Microsoft for their 2026 Work Trend Index Annual Report on the future of AI and work, alongside a thoughtful group of leaders and researchers including Karim Lakhani (my former Harvard Business School professor!), LinkedIn’s Chief Economic Opportunity Officer, and others. Sneak peek below, full report out on my socials later today.

It was the first Monday of May last night so obviously I went to…an OpenAI dinner with Codex leadership (what, you thought the Meta gala?!). It was an incredible evening learning about OpenAI’s push into agentic workflows. More Tea on that coming soon.

Jensen Huang said something weird recently.

On a podcast, the NVIDIA CEO pushed back on the idea that NVIDIA should operate from a place of restraint, saying, “we’re not losers.”

Very alpha CEO energy. Very leather jacket meets global chips strategy.

But TBH, it also sounded a little defensive.

The truth is twofold: yes, NVIDIA is still the undisputed king of AI chips. But also, yes: the kingdom is under siege.

Google’s TPUs are becoming a real alternative for frontier AI companies. Anthropic deep Google partnership is one of the clearest examples of a major lab building on that ecosystem.

And then there’s China.

Jensen’s argument is that if the U.S. cuts China off from American chips, China will just accelerate its own domestic alternatives.

He may be right.

But we’ve seen this movie before, with Tesla and others: Western company enters China, China learns the playbook, local competitors rise, and suddenly the market you were chasing becomes the market that replaces you. And with AI, there’s real national security risks attached to that.

So yes, NVIDIA is still winning.

But the next phase of AI growth is getting a lot less comfy for Jensen.

Full Tea below.

Apparently, ChatGPT’s coding agent really loves goblins.

No, seriously. Codex kept sneaking goblins and gremlins into responses so often that OpenAI engineers added an explicit prompt instruction: cool it with the creature references.

(which is honestly an incredible sentence)

The reason? Training incentives.

The “Nerdy” personality setting had rewarded playful, creature-y language a little too heavily, and the model did what models do: it responded to the reward signal.

A tiny, ridiculous reminder that AI behavior is not magic. It is shaped by what we optimize for.

And sometimes, apparently, what we optimize for is goblins.

in Tech Tea we trust,
Meghana