ZilverNet Signals

What's being invented, and what it actually changes.

AI, open models, robotics and the economics underneath them — explained without the hype, with a source for every claim and an honest read on how far along each thing really is.

Agility Robotics' Digit humanoid robot in GXO colors

Image credit: Agility Robotics

RoboticsPilotFeatured

Events Oct 21 – Oct 26, 2026

Humanoid robots reached the production line this month. Nobody has made them pay yet.

Tesla started building Optimus in Fremont, and Agility signed up two more industrial customers for Digit. The logistics company running the most humanoid pilots says it hasn't earned its money back.

The short version: "We have not achieved ROI on humanoids yet," says GXO's CEO, after 45 pilots. That sentence is the state of the industry.

Oct 29, 20268 min read6 sources

Latest

Mistral's launch artwork: the words '#1 in Sovereign AI.' above orange and red stripes

Image credit: Mistral

Open-source AIShipping

Weights released Oct 29, 2026

Mistral releases the weights for its trillion-parameter model, pitched at cyber defense.

Mistral Large 4 is now downloadable, three weeks after launch. Its argument for open weights: "Access that can't be cut off by definition."

1T / 52B activeweights on Hugging Faceon-prem

Oct 29, 20265 min read

The Oracle and NVIDIA logos side by side

Image credit: Nvidia

Chips & computeShipping

Generally available Oct 27, 2026

Nvidia's next data-center platform, Vera Rubin, is now generally available on Oracle's cloud.

72-GPU Rubin racks and the new Vera CPU, rentable now. The "multi-gigawatt" expansion is a plan without a number.

NVL72Vera CPUOracle Cloud

Oct 29, 20264 min read

Image credit: Anthropic

AI economics

IPO filing made public Oct 26, 2026

Anthropic files to go public: $16.3 billion in half-year revenue, and founders who keep control.

An $80.7B net loss that is mostly a non-cash charge, a $693M operating loss underneath it, and co-founders holding 50.1% of the vote. No price yet.

Nasdaq: ANTH1 rev $16.3BS-1

Oct 29, 20266 min read

A laptop shown open at several angles, with its internals visible through a translucent shell

Image credit: Nvidia

Chips & computeShipping

On sale Oct 16, 2026

Nvidia's first Windows chip is built to run AI on your laptop, not in the cloud.

An Arm CPU and a Blackwell GPU sharing up to 128 GB — enough, Nvidia says, for a 120B model locally. From $2,599, or $2,399, depending on which Nvidia page you read.

Grace CPU + Blackwellup to 128 GBlaptops

Oct 29, 20265 min read

Image credit: Reflection

Open-source AIShipping

Weights released Oct 16, 2026

Reflection's Beam is out under Apache 2.0 — a US open model aimed squarely at the Chinese leaders.

A genuinely permissive license on a 501B model (23B active). Unverified so far: Reflection's own benchmarks putting it level with the Chinese leaders.

Apache 2.0501B / 23B activetext only

Oct 29, 20265 min read

False-color micrograph of the five-qubit silicon device, qubits Q1 to Q5 marked in a row

Image: Nature Physics, CC BY

Chips & computeScienceResearch

Announced Oct 16, 2026

Five silicon qubits, one control wire, 99.9% accuracy at once.

RIKEN researchers found the hidden error that appears when qubits run together — and calibrated it away. A step toward scaling up silicon-based quantum computers.

Nature PhysicsRIKEN

Oct 29, 20263 min read

Meta's launch graphic: 'Muse Spark Open Weights' over flowing blue lines

Image credit: Meta

Open-source AIShipping

Released Oct 15, 2026

Meta's new open model runs on two Mac Studios. Read the license before you build on it.

Muse Spark Open is free to use, modify and sell — until you offer it as a service and your group earns over $20M a year. Then Meta wants a separate agreement.

Muse Model License816B / 42B active1M context

Oct 29, 20265 min read

Five app icons arranged in front of a night-time view of Earth

Image credit: Nvidia

Open-source AIShipping

Released Oct 15, 2026

Nvidia's new open models are built to run agents on your own PC, not in the cloud.

Nemotron 3.5 comes in two tiers: a 120B model for enterprise agents, and a 74B "Bolt" model sized for 24 GB that runs on Nvidia's new AI PCs.

120B Super VLBolt 74B · 24 GBlicense not stated

Oct 29, 20264 min read

A black server rack branded AMD Helion, door open to show rows of compute nodes

Image credit: AMD

Chips & computePrototype

Announced Oct 12, 2026

AMD's answer to Nvidia's AI racks is an open blueprint built for agents.

Helion is a rack design, not a chip — mixing CPUs and GPUs for agent workloads, with the spec headed to the Open Compute Project.

open specEPYC + MI350PNov 2026

Oct 29, 20264 min read

A graph whose nodes are colored with three colors so that no edge joins two nodes of the same color

Illustration: ZilverNet

ScienceAIResearch

Announced Oct 6, 2026

OpenAI says an AI proved one of computer science's biggest open problems. No independent expert has reviewed it yet.

The unique games conjecture has stood since 2002. Weeks earlier, as rumors spread, an MIT team rushed out their own hand-written result — solved "with their minds," as one colleague put it.

Oct 29, 20266 min read

Schematic of a thin gel body with two curved fins, one lit from above

Illustration: ZilverNet

InventionsRoboticsResearch

Announced Sep 29, 2026

MIT built a robot the size of a stick of gum that swims using living muscle.

Light-controlled muscle cells on a half-millimeter gel film move it about four body lengths a minute — the first flat robot to travel on living tissue.

Adv. Functional MaterialsMIT

Oct 29, 20263 min read

The Artemis project banner: a thin crescent moon beside the word ARTEMIS

Image credit: Google

Open-source AIAIPilot

Open-sourced Aug 13, 2026

Google's Artemis lets AI coding assistants drive a real Android phone — and it's open source.

Plain-language instructions in, Android automation out — wired into Claude Code, Codex and Antigravity through MCP. Its 99% benchmark score is self-reported.

Apache 2.0AndroidMCP

Oct 29, 20265 min read