▼AI Inference Shifts to Consumer Devices

Score history

Daily conviction score, 0 to 100. Higher means the thesis is more strongly corroborated.

Now 98 · -2 since Oct 9 · ranged 98 to 100

Showing the last few days. Unlock full score history.

Why the conviction moved

• Oct 9
  Strengthened +4

  Nvidia and Microsoft opened preorders for RTX Spark laptops with up to 128 GB of unified memory, marketed for running agents locally. Consumer hardware is now sized to keep agent workloads off the cloud.

• Oct 8
  Strengthened +4

  Nvidia and Microsoft opened pre-orders for Arm-based RTX Spark Windows laptops. The Surface Laptop Ultra starts at $2,599 and ships October 16, and its 128 GB unified-memory configuration can hold mid-sized open-weight models locally, moving serious inference onto mainstream consumer Windows hardware.

• Oct 8
  Strengthened +2

  Google's EmbeddingGemma 2 has 740 million parameters, is released under Apache 2.0, and covers text, code, images and audio. Google reports the text-only configuration runs in about 191 MB of active memory on a quantized Pixel 11 Pro, so retrieval embeddings can run on phones without a cloud call.

Showing the last 2 days. Unlock the full record.

Source trail

• Supporting · October 9, 2026

  Nvidia and Microsoft Open Preorders for RTX Spark Laptops With Up to 128 GB of Unified Memory for Local Agents

  Nvidia and Microsoft opened preorders for RTX Spark laptops with up to 128 GB of unified memory, marketed for running agents locally. Consumer hardware is now sized to keep agent workloads off the cloud.

  NVIDIA Blog

• Supporting · October 8, 2026

  Google releases EmbeddingGemma 2, a 740-million-parameter open embedder for text, code, images and audio under Apache 2.0

  Google's EmbeddingGemma 2 has 740 million parameters, is released under Apache 2.0, and covers text, code, images and audio. Google reports the text-only configuration runs in about 191 MB of active memory on a quantized Pixel 11 Pro, so retrieval embeddings can run on phones without a cloud call.

  Google DeepMind