AILANTA
← All research
GitHub · Hugging Face · npm

Local AI is becoming a deployment stack

First-half data shows local model formats expanding, desktop runtimes growing, and browser package usage rising sharply. This is no longer a narrow reaction to cloud cost.

16.07.20264,403 GitHub2,373 Hugging FaceProduct Hunt excluded
Main finding

The market is moving at two speeds

The local-oriented share of a broad GitHub AI cohort rose from 10.1% to 13.2%. New supply is concentrating around GGUF and llama.cpp, while ONNX and WebGPU produce fewer new repositories but their established browser packages are seeing much higher usage.

01

Local share within AI tooling

The same broad GitHub selection method in both quarters.

Local share10.1% → 13.2%

+3.1 pp

Targeted runtime projects321380

18% Q/Q

llama.cpp projects107 → 167

+56.1%

Desktop solutions153 → 207

+35.3%

02

New GitHub runtimes by month

Targeted searches deduplicated by repository ID.

079158JanFebMarAprMayJun

April was the peak month for new local runtime projects. June's decline is not proof of market contraction: newer repositories had less time to reach the five-star inclusion threshold.

03

Formats are diverging

A project may match more than one technical format.

FormatQ1Q2Change
GGUF6883+22.1%
MLX4628-39.1%
ONNX2815-46.4%
WebGPU1413-7.1%
Quantized21210%
llama.cpp107167+56.1%
04

Hugging Face model supply is accelerating

Top-by-downloads cohorts within each filter; categories overlap.

0394787JanFebMarAprMayJun
FamilyQ1Q2Q2 median downloads
GGUF11839228.5K
MLX1594662.1K
ONNX96791.4K
WebGPU571551
Quantized1255572.9K
mobile18319137

GGUF, MLX, and quantized models show the clearest supply expansion. WebGPU model count increased, but median downloads remain extremely low—early supply rather than validated mass adoption.

05

Browser inference is already being used

Average daily downloads for a fixed benchmark of npm packages.

PackageQ1/dayQ2/dayChange
@huggingface/transformers53.3K143.8K+169.6%
onnxruntime-web133.3K276.9K+107.8%
@mlc-ai/web-llm4.6K7.5K+61.8%
node-llama-cpp111.6K79.2K-29.1%
@mediapipe/tasks-genai3161.3K+323.6%
@xenova/transformers44.7K67.7K+51.5%
06

What this means for the market

01

Local AI is no longer one category

Distinct layers are forming around models, runtimes, desktop shells, browser inference, and mobile deployment.

02

ONNX is shifting from creation to usage

New ONNX repositories declined, while onnxruntime-web average daily downloads rose 107.8%. This signals maturing infrastructure, not disappearing demand.

03

Desktop currently leads mobile

New supply is concentrating in desktop runtimes and execution cores. Mobile models are present, but the product layer is developing more slowly.

Methodology and limitations

GitHub combines a broad control cohort with targeted searches across ten runtime themes from January–June 2026; repositories are deduplicated. Hugging Face takes up to 1,000 current top-by-downloads models per filter and then restricts them to H1 2026 creation dates. npm compares average daily downloads for a fixed package benchmark.

Hugging Face downloads and likes are current metrics, not historical snapshots. The Hugging Face sample does not estimate the entire Hub. GitHub stars introduce age and survivorship bias. The npm benchmark is not a full market-share estimate. Model size is excluded because the list API did not return sufficient usedStorage data.