← Back to newsletters

AI Daily Brief — Tue Jul 14

2026-07-14

By Vadym · Generated with AI, curated by me


Listen to this issue

Today’s stories are less about new models and more about who controls the plumbing underneath them — the compute, the capital, the supply chains, and the borders states are starting to draw around which models can cross them.


Headlines & News
Workforce

Anthropic Pulls Monzo Co-Founder Tom Blomfield Onto Its Compute Team

Tom Blomfield, who co-founded and ran UK digital bank Monzo before becoming a Y Combinator general partner, is taking a leave from YC to join Anthropic’s compute team as a member of technical staff, working alongside Anthropic’s Tom Brown. It follows Andrej Karpathy’s earlier 2026 move to Anthropic to lead pretraining.

Source: Sifted

Funding

SambaNova Raises $1B at $11B Valuation, Five Months After Its Last Mega-Round

AI chip maker SambaNova closed the first tranche of a $1 billion Series F on July 8, led by General Atlantic with backing from BlackRock, Intel Capital, Qatar Investment Authority, and T. Rowe Price, pushing its valuation to $11 billion — roughly 5x its February 2026 Series E mark. JPMorgan Chase signed on as an inference infrastructure customer alongside the round.

Source: TechCrunch

Robotics

Agility Robotics to Go Public in the Largest Capital Raise in Humanoid-Robot History

Agility Robotics, maker of the Digit humanoid robot already deployed at Schaeffler, GXO, Toyota, and Mercado Libre, is merging with Michael Klein-backed SPAC Churchill Capital Corp XI in a deal valuing the combined company at $2.5 billion. The transaction brings in more than $620 million in gross proceeds, including a $200 million PIPE led by Foxconn, and Agility says it already holds over $300 million in multi-year orders for its next-gen Digit v5.

Source: Yahoo Finance

Policy

Beijing Weighs Restricting Overseas Access to China’s Top AI Models

Chinese officials led by the Ministry of Commerce met with Alibaba, ByteDance, and Z.ai to discuss limiting overseas access to the country’s most advanced AI models — including closed and open-weight releases not yet public — and floated treating leaks of proprietary AI technology as a national security offense. No formal rule exists yet, and the scope remains undecided.

Source: Taipei Times

Research

Anthropic Finds a Hidden Layer of Words Models Use to Think but Never Say

Anthropic researchers identified what they call “J-space” — a region inside large language models filled with words and concepts that never surface in the model’s visible output but that appear to shape how it works through a problem. The finding is part of Anthropic’s ongoing interpretability research into how models reason internally, distinct from the text they actually generate.

Source: MIT Technology Review


Analysis

Takeaway

Strip away the headlines and today is about control over the layers beneath the model. Talent is flowing toward compute logistics, not just research. Chip capital is chasing signed enterprise contracts, not projections. Public markets are pricing humanoid robots on deployed orders. And a government is starting to treat model access itself as a resource worth gating at the border. Even Anthropic’s own research lands on the same theme from a different angle: before you decide who controls a system, you need to know what’s actually happening inside it — and right now, nobody fully does.

— Boba


Curated by Vadym