AI Box
AI Box | Daily
01 · Business

Anthropic gives investors 48 hours to commit to $50B round at $900B valuation

Anthropic is asking investors to commit to its next round within 48 hours, with the roughly $50B raise expected to close inside two weeks at a valuation of $900B or more, according to TechCrunch reporter Marina Temkin. The compressed timeline reflects what sources described as soaring demand, and the final number may exceed the $900B target. At that price, Anthropic would more than double the $380B valuation it set in February.

The details:
  • Anthropic is targeting a $900B+ valuation on a roughly $50B round expected to close within two weeks.
  • Investors have 48 hours to submit allocations, per sources cited by TechCrunch.
  • The company's last round priced at $380B in February 2026; the new mark would more than double that.

Why it matters: Still, the trajectory speaks for itself. Anthropic has gone from a $380B valuation in February to a target of $900B in roughly nine months, with run-rate revenue climbing from low single-digit billions to something between $30B and $40B over the same span.

Read the full story →

02 · Security

Shivon Zilis emerges as Musk's covert OpenAI liaison in week one of trial

Shivon Zilis, the longtime Musk lieutenant and mother of four of his children, surfaced this week as the central back channel between Elon Musk and OpenAI during the years he publicly broke from the company. Court exhibits in Musk v. Altman show Zilis advising Musk from inside OpenAI's board room while simultaneously coaching Sam Altman on how to manage Musk's temper.

The details:
  • Shivon Zilis joined OpenAI as an advisor in 2016 and served on its nonprofit board from 2020 to 2023.
  • Zilis is the mother of four of Elon Musk's children, with the first two born in 2021.
  • Musk donated about $38 million to OpenAI but imposed no conditions blocking a future restructuring.

Why it matters: Musk used what is likely his last turn on the stand to drive home a single line, repeating a version of 'you just can't steal a charity' at least five times. He testified that his concerns over OpenAI's drift only crystallized around 2023, telling the court, 'Only recently has it been obvious that the charity had been stolen.' That timing is the case's weakest joint, given how long Musk waited to sue and how publicly…

Read the full story →

03 · Business

Legora hits $5.6B valuation as Nvidia joins legal AI fight against Harvey

Legora has raised a $50M Series D extension at a $5.6B post-money valuation, with Nvidia's corporate venture arm NVentures leading the round in what is reportedly its first legal AI investment. The Swedish startup crossed $100M in annual recurring revenue in the month between this extension and its $550M Series D, a pace that is narrowing the gap with U.S. rival Harvey, which hit an $11B valuation last month.

The details:
  • Legora closed a $50M Series D extension at a $5.6B post-money valuation, one month after its $550M Series D.
  • Nvidia's corporate VC arm NVentures led the round in what is reportedly its first legal AI investment, alongside Atlassian.
  • Legora crossed $100M in annual recurring revenue, with 1,000+ law firms and in-house legal teams across 50 markets.

Why it matters: For the broader AI market, the read-through is that the application layer is finally producing real revenue at real scale. Legora went from product launch to $100M ARR in 18 months. Harvey is at $11B. That is faster monetization than most of the SaaS comparables a generation earlier, and it is happening on top of compute the startups themselves do not own.

Read the full story →

With Love from AI Box Try It!
AI Box AI Box
Every AI model. One chat.
The latest models from ChatGPT, Claude, Gemini, ElevenLabs — 80+ models in a single chat. Compare answers side by side. Pick the best one every time.
  ChatGPT, Claude, Gemini, Grok, DeepSeek — in one chat
  Generate images & video with Sora, Veo, Ideogram
  Compare any two models side by side
  From $8.99/mo · 80+ models, all included
Try AI Box  → aibox.ai
Trusted by 3,000+ teams · Powered by OpenAI, Anthropic, Google, Meta, Perplexity
04 · Business

AWS Launches OpenAI Models on Bedrock as Microsoft Exclusivity Ends

Amazon Web Services has put OpenAI's frontier models, its Codex coding platform and its managed agents on Bedrock, ending the cloud exclusivity that has tethered OpenAI to Microsoft Azure since 2023. The launch dropped one day after the Microsoft–OpenAI exclusivity deal officially expired, a sequence that reads less like coincidence than choreography.

The details:
  • AWS launched OpenAI's GPT-5.5 and GPT-5.4 models, plus the Codex platform and managed agents, on Amazon Bedrock.
  • The Bedrock launch landed one day after Microsoft Azure's exclusivity deal with OpenAI officially expired.
  • Microsoft's original $10 billion investment in OpenAI had locked the models to Azure since the launch of ChatGPT.

Why it matters: The bigger shift is distribution. OpenAI now reaches the cloud where most large enterprises already run their workloads, and AWS gets a headline model family to pair with its existing Anthropic alliance. If usage follows availability, the next quarter of model-revenue league tables could look very different from the last one.

Read the full story →

05 · Security

Musk testifies xAI used OpenAI models to train Grok

Elon Musk testified Thursday in a California federal court that xAI used OpenAI's models to help train Grok, answering 'Partly' when OpenAI lawyer William Savitt asked directly whether xAI had distilled its rival's systems. Pressed further, Musk said it is 'standard practice to use other AIs to validate your AI' and that 'generally all the AI companies' do the same.

The details:
  • Under cross-examination Thursday, Musk answered 'Partly' when asked if xAI used distillation on OpenAI's models to train Grok.
  • Musk said it is 'standard practice to use other AIs to validate your AI' and that 'generally all the AI companies' do it.
  • OpenAI told a House committee in a February 2026 memo it has 'taken steps to protect and harden our models against distillation.'

Why it matters: For xAI specifically, the admission is awkward but not fatal. Grok's commercial position depends on its integration with X and on Musk's willingness to keep funding compute, not on a clean training provenance. For OpenAI, the bigger problem is that its own legal team just got Musk to confirm, on the record, the thing it has spent a year telling regulators only adversaries do. That line is now a lot harder to draw.

Read the full story →

Everything else in AI today

How was today's issue?

One tap helps me dial in tomorrow's send.

🔥
Nailed it
5 / 5
👍
Solid
3 / 5
😬
Missed
1 / 5

Share more feedback →

That's it for today.

Same stories, fuller context: catch the AI Chat podcast wherever you listen. Reply with what we missed — I read every one.

Browse all today's stories · forward to a friend.

Keep Reading