AI Mornings with Andreas Vig
Your daily AI news briefing in under 10 minutes. New models, product launches, research breakthroughs, and industry shifts, explained clearly, no hype.
AI Mornings with Andreas Vig
Cerebras IPO Pops 108% & Musk-Altman Trial Goes to Jury
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Hey, welcome to AI Mornings with Andreas Vig. It's the 15th of May 2026. The first major tech IPO of the year just happened, and it was a blockbuster. Cerebra Systems, the AI chip company known for building dinner plate-sized processors purpose-built for AI workloads, raised $5.5 billion in its IPO yesterday. Shares were priced at $185 and opened at $385. That's a 108% pop on day one. The company now trades at a $66 billion valuation. What's remarkable is where Cerebras was a year ago. The IPO was shelved due to regulatory concerns about a large investment from Abu Dhabi-based Group 42. But the company doubled its revenue to $510 million in 2025, swung from a near half billion dollar loss to $238 million in profit, and diversified its customer base beyond a single client. Now, OpenAI, Amazon Web Services, and Saudi Arabia's AI University are all customers. It's a strong signal that investors still have enormous appetite for companies that can challenge Nvidia's dominance in AI compute. Meanwhile, in a California courtroom, nine jurors are now deliberating the future of Open AI. Elon Musk's lawsuit against Sam Altman, Greg Brockman, and Microsoft has reached the jury after weeks of testimony covering everything from the founders' breakup in 2018 to Altman's firing and rehiring in 2023. The jury will decide three main claims: breach of charitable trust, whether OpenAI violated agreements about how Musk's donations would be used, unjust enrichment, whether founders personally benefited from those donations through the for-profit arm, and aiding and abetting whether Microsoft knowingly helped violate that charitable mission. OpenAI's defenses hinge on Statute of Limitations arguments and what's called the unclean hands doctrine, essentially arguing Musk's own conduct undermines his claims. If Musk wins, it could potentially end OpenAI as a for-profit company. The judge will begin hearings next week on what consequences a verdict for the plaintiffs might mean. A provincial audit in Ontario, Canada just revealed some deeply concerning findings about AI in healthcare. The Auditor General evaluated 20 AI scribe systems approved for doctors and healthcare providers, and the results were stark. Twelve of the 20 systems inserted incorrect drug information into patient notes. Nine fabricated treatment suggestions that were never discussed. 17 missed key details about patients' mental health issues. Perhaps most damning, when these systems were evaluated for procurement, accuracy counted for only 4% of the total score, while having a domestic presence in Ontario counted for 30%. Over 5,000 physicians in Ontario are currently using these systems. And the health ministry says there have been no known reports of patient harm yet, but this audit raises serious questions about how these tools are being vetted for critical medical use. Alright, a few more things worth knowing about today. OpenAI is preparing possible legal action against Apple. Their two-year-old ChatGPT Siri partnership hasn't delivered the subscription revenue OpenAI expected, and Bloomberg reports the company has hired an outside law firm to explore options including a breach of contract notice. The 2024 deal integrated ChatGPT into Apple devices, but users apparently haven't been converting to paid ChatGPT accounts at the rate OpenAI anticipated. It's a reminder that even high-profile AI partnerships don't always pan out as planned. Speaking of OpenAI, the company just brought Codex to the ChatGPT mobile app. You can now control your AI coding agent from your phone reviewing outputs, approving commands, and monitoring progress on long-running tasks remotely. More than 4 million people are now using Codex weekly. The feature uses a secure relay to connect to your development environment without exposing it to the public internet. It's a smart move as AI agents take on longer, more complex work that requires occasional human oversight. SpaceX AI, the combined entity from Elon Musk's merger of XAI with SpaceX, is facing a talent drain. More than 50 researchers and engineers have departed since February, including members of the pre-training team that's critical to building new AI models. Reports cite Musk's culture of extreme work as a factor. It raises questions about whether the $1.25 trillion combined company remains committed to developing leading AI models, or if the merger has disrupted the research culture needed to compete with OpenAI and Anthropic. Finally, ArxSiv has announced a new policy: a one-year ban for authors who submit papers with AI hallucinated citations. The move comes as data shows hallucinated references have risen tenfold since 2023, now appearing in one of every 277 papers. It's a strong signal that academic institutions are starting to treat AI generated misinformation in scientific literature as a serious problem requiring real consequences. That's it for today. Thanks for listening.