Category intelligence
29 current items analyzed and ranked.
Executive synthesis
Record-breaking funding and infrastructure deals dominated this cycle. An unnamed reinforcement learning startup raised a record $1.1B seed round targeting superintelligence, while Meta signed a major chip deal with AWS and Lightelligence surged 400% on its Hong Kong IPO debut, signaling investor conviction in optical interconnect as AI's next bottleneck.
Military and enterprise AI saw major moves:
The Musk v. Altman trial opened with dramatic testimony, with potential implications for OpenAI's corporate structure. Meanwhile, the FIDO Alliance, Google, and Mastercard began building authentication standards for AI agent commerce.
Primary evidence
Continuing our coverage of David Silver's new venture from News yesterday, A reinforcement learning startup raised a record $1.1 billion seed round with the stated goal of achieving superintelligence. This is the largest seed funding round ever recorded.
By Sanya Mansoor and agencies
Google has signed a classified deal with the US Pentagon allowing the military to use its AI models for 'any lawful government purpose.' This puts Google alongside OpenAI and xAI as Pentagon AI suppliers.
NVIDIA released Nemotron 3 Nano Omni, a long-context multimodal model for enterprise AI agents that can process documents, audio, and video. It expands NVIDIA's software and model offerings beyond hardware.
By Unknown
NVIDIA published a detailed blog on Nemotron 3 Nano Omni, a multimodal model for long-context document, audio, and video understanding designed for enterprise AI agents.
By Dashveenjit Kaur
Lightelligence, a Chinese photonics chipmaker, saw its stock surge ~400% on its Hong Kong IPO debut, briefly hitting $10B market cap on just $15.5M revenue. Investors are betting optical interconnect will solve AI's next major infrastructure bottleneck.
Meta has signed a deal to use AWS chips for scaling its AI infrastructure, adding to a flurry of major compute procurement deals among tech giants.
By Dara Kerr in San Francisco and Nick Robins-Early
Continuing our coverage of the Musk v. OpenAI trial from News yesterday, The Musk v. Altman trial began with dramatic opening arguments. Musk's attorney accused Altman and Brockman of having 'stole a charity' when OpenAI pivoted to a for-profit structure.
By Asif Razzaq
OpenAI released Privacy Filter, a 1.5B-parameter open-source model for PII redaction under Apache 2.0 license. With only 50M active parameters, it's small enough to run in a browser while detecting 8 categories of sensitive data.
By Kyle Orland
First spotted on Reddit yesterday, now getting official mainstream coverage, GitHub is shifting Copilot to usage-based billing starting June 1, citing the need to align pricing with actual compute costs. The change reflects the wide variance in backend costs between simple queries and multi-hour autonomous coding sessions.
By Lily Hay Newman
The FIDO Alliance has partnered with Google and Mastercard to develop standards for AI agents making purchases on behalf of users. The effort aims to prevent fraud and ensure proper authentication for autonomous spending.
By Paresh Dave, Maxwell Zeff
Elon Musk testified in the Musk v. Altman trial, claiming he founded OpenAI to prevent a 'Terminator outcome.' The judge warned both Musk and Altman to stop using social media to inflame the dispute.
Latent.Space highlights the creative and productive potential of GPT-Image-2 combined with Codex, showing applications in education, pop culture, infographics, and integrated asset generation during coding workflows.
