AI News Weekly – Can you guess how much cash Open AI was losing with Sora? – 26 March 2026

by ai-intensify
0 comments
AI News Weekly – Can you guess how much cash Open AI was losing with Sora? - 26 March 2026

Welcome

Who can say what AI can and cannot do?

The center of gravity of the week was accountability – who can say what AI can and cannot do. A federal judge ruled the Pentagon’s blacklisting of Anthropic was wrong. "Attempt to maim." Two separate juries found Meta liable for designing an addictive platform. And OpenAI quietly killed Sora, admitting that it was spending $15 million a day on a product no one was ready for.

ai weekly

sponsor

Become an AI Consultant and provide ‘ROI with AI’ to your clients

AI is changing every workplace – but executives are afraid of becoming one of those companies that “gets no ROI on AI.”

That’s where you come in, and how you can Build a 6-Figure Consultancy Innovating with proven methods of AI to deliver faster ROI on AI projects.

Click here to request access to the AI ​​Consultancy Project →

innovatingwithai.com

in the news

Weekly AI Briefing:

March 20-26, 2026

Arm ships silicon, OpenAI kills Sora, a judge interrogates the Pentagon, and a jury decides Addiction by Design is now a legal liability.



🎬Watch and listen first

Jensen Huang: "I think we have achieved AGI" · March 22 · Lex Friedman Podcast #494
This year, four words were most debated in AI. Huang redefines AGI "An AI that creates a billion dollar company" -Conveniently requires Nvidia chips at every step.


court week

Judge finds Pentagon’s anthropogenic blacklisting wrong "attempt to maim" · March 24 · cnbc
The strongest judicial protest ever against an admin’s AI action. If the injunction goes into effect, ethical red lines become legally defensible — and the Pentagon’s social-media-post-a-policy approach could be dead on arrival.

Meta and YouTube found liable in historic social media addiction trial · March 25 · npr
Two juries in two days – $6 million in Los Angeles and $375 million in New Mexico. The legal strategy of targeting design choices rather than content opened only 2,000 pending cases. This is Big Tech’s big tobacco moment.


silicon shift

Arm ships its first ever chip – ends 35 years of pure licensing · March 24 · cnbc
The AGI CPU, co-developed with Meta and already ordered by OpenAI and Cloudflare, claims 2x performance-per-rack over x86. Arm expects $15B annual revenue from chips by 2031. The stock jumped 16%.

Reflection AI eyes raising $2.5B at Nvidia-backed $25B valuation · March 25 · reuters
A DeepMind-spinoff less than two years old, with minimal revenue, has seen its valuation triple to $25B. "sovereign ai" Thesis – an open-source model for allied countries – is now a $2.5B stake.


OpenAI’s spring cleaning

OpenAI kills Sora, Disney’s $1B deal collapses · March 24 · hollywood reporter
Sora spent ~$15M/day in estimated costs against lifetime revenue of $2.1M. 30 minutes into a work meeting, Disney was told the product was finished. OpenAI’s Fidji Simo said this "side search" – Now the company is moving rapidly towards venture and Q4 IPO.


Robots walk the red carpet

Melania Trump reached the White House with a humanoid robot · March 25 · cnn
Image 03, World leaders greeting in 11 languages ​​at an education summit, made for a surreal photo op. But the real story is that the first lady put humanoid tutors in every home — an Overton Window shift to robotics policy that happened before most people noticed.


AI tax on workers

Oracle plans to lay off up to 30,000 to fund AI data center buildout · March 5 · bloomberg
The clearest example yet of an AI capex squeeze: Cut 18% of your workforce to finance data centers that won’t return cash by 2030. Oracle joins the growing list of companies hit the block (cut 4,000) where humans finance the machines that replace them.


The supply chain is the attack surface

LiteLLM backdoor via compromised Trivi scanner – 3.4M daily downloads exposed · March 24 · snyk
A security scanner was hacked, which compromised the LLM gateway used by 36% of cloud environments. The credential-stealing malware remained live for three hours. Threat group TeamPCP is now openly partnering with LAPSUS$, calling for open-source security status "A joke."


In 2026, there is no question "Can AI do this?" Its "Who decides what he does – and who pays when it goes wrong?"

ai weekly


This RSS feed is published at https://aiweekly.co/. You can also subscribe via email.

Related Articles

Leave a Comment