AI publishing splits into four tiers, and where you can land a byline depends on whether you have working code or a working narrative.
The tech news sites do not take guest posts. The applied-ML outlets take reproducible technical writing. The research and framework blogs want co-authored work or expert commentary. The newsletters and independent blogs are open. The list of 100 is below, grouped by tier.
Pick your tier first
| Tier | Examples | How you get in |
|---|---|---|
| Tech news & industry | TechCrunch, Wired, VentureBeat, The Register, ZDNet, AI Business | Not guest posts. Exclusive research, a novel benchmark, or expert commentary on a live story |
| Applied ML & data science | KDnuggets, Analytics Vidhya, Towards Data Science, MarkTechPost, Machine Learning Mastery | Reproducible tutorials, benchmarks, technique explainers, production post-mortems |
| Research & framework blogs | Google Research, NVIDIA, PyTorch, TensorFlow, Hugging Face, The Gradient | Co-authored research, an interview, or expert quotes for a piece they are writing |
| Newsletters & independent blogs | The Sequence, Last Week in AI, Import AI, Sebastian Raschka, Lilian Weng | Open to contributions and cross-posts; deep dives, paper walkthroughs, opinion with substance |
What the metrics mean
| Metric | What it is | How much it matters |
|---|---|---|
| Domain Rating (DR) | Ahrefs' 0 to 100 score for a whole site's backlink profile | A useful filter, not a verdict |
| Domain Authority (DA) | Moz's equivalent 0 to 100 score | Same idea as DR |
| Page Authority (PA) | Moz's score for one specific page | Matters once you know which page your post will live on |
| URL Rating (UR) | Ahrefs' score for one specific URL | Same purpose as PA |
| Monthly Traffic | Estimated organic visits per month | The most honest signal that a link will be seen |
Here is the deal: the AI niche moves faster than any other on this site. A tutorial that ranked last year is stale now, and half the "AI blogs" from two years ago are parked domains. Judge by recent publishing activity and real traffic, not DR alone.
The 100 AI & ML sites
Domain Rating, Domain Authority, Page Authority, URL Rating and estimated monthly traffic. Every domain was checked live in September 2026; shut-down and acquired sites were swapped for live equivalents. Metrics are third-party estimates and shift fast.
| # | Website | DR | DA | PA | UR | Monthly traffic |
|---|---|---|---|---|---|---|
| 1 | techcrunch.com | 91 | 93 | 88 | 84 | 27,500,000 |
| 2 | wired.com | 91 | 93 | 88 | 84 | 33,200,000 |
| 3 | reddit.com/r/MachineLearning | 91 | 93 | 88 | 84 | 82,400,000 |
| 4 | venturebeat.com | 90 | 92 | 87 | 83 | 14,800,000 |
| 5 | developers.google.com/machine-learning | 90 | 92 | 87 | 83 | 9,800,000 |
| 6 | microsoft.com/en-us/research/blog | 89 | 91 | 86 | 82 | 9,700,000 |
| 7 | towardsdatascience.com | 89 | 91 | 86 | 82 | 7,200,000 |
| 8 | techrepublic.com | 89 | 91 | 86 | 82 | 15,300,000 |
| 9 | ibm.com/blogs/research | 88 | 90 | 85 | 81 | 7,400,000 |
| 10 | openai.com/blog | 88 | 90 | 85 | 81 | 18,500,000 |
| 11 | thenextweb.com | 88 | 90 | 85 | 81 | 8,900,000 |
| 12 | aws.amazon.com/blogs/machine-learning | 92 | 94 | 89 | 85 | 40,000,000 |
| 13 | ai.meta.com | 87 | 89 | 84 | 80 | 8,200,000 |
| 14 | gartner.com/en/information-technology/insights/artificial-intelligence | 87 | 89 | 84 | 80 | 11,400,000 |
| 15 | research.google/blog | 86 | 88 | 83 | 79 | 5,800,000 |
| 16 | ai.googleblog.com | 86 | 88 | 83 | 79 | 3,400,000 |
| 17 | huggingface.co/blog | 86 | 88 | 83 | 79 | 12,000,000 |
| 18 | blogs.nvidia.com | 86 | 88 | 83 | 79 | 4,200,000 |
| 19 | zdnet.com/topic/artificial-intelligence | 85 | 87 | 82 | 78 | 4,800,000 |
| 20 | kdnuggets.com | 85 | 87 | 82 | 78 | 2,800,000 |
| 21 | hackernoon.com | 85 | 87 | 82 | 78 | 4,800,000 |
| 22 | deepmind.google/discover/blog | 85 | 87 | 82 | 78 | 2,700,000 |
| 23 | tensorflow.org/blog | 85 | 87 | 82 | 78 | 4,200,000 |
| 24 | oreilly.com/radar | 84 | 86 | 81 | 77 | 6,000,000 |
| 25 | developer.nvidia.com/blog | 84 | 86 | 81 | 77 | 3,900,000 |
| 26 | pytorch.org/blog | 84 | 86 | 81 | 77 | 3,800,000 |
| 27 | analyticsvidhya.com | 84 | 86 | 81 | 77 | 4,200,000 |
| 28 | medium.com/towards-ai | 84 | 86 | 81 | 77 | 3,900,000 |
| 29 | analyticsindiamag.com | 84 | 86 | 81 | 77 | 1,400,000 |
| 30 | dzone.com/ai-machine-learning-big-data | 84 | 86 | 81 | 77 | 3,800,000 |
| 31 | theregister.com/software/ai_ml | 84 | 86 | 81 | 77 | 3,200,000 |
| 32 | amazon.science | 83 | 85 | 80 | 76 | 2,100,000 |
| 33 | databricks.com/blog | 83 | 85 | 80 | 76 | 3,400,000 |
| 34 | machinelearningmastery.com | 83 | 85 | 80 | 76 | 3,400,000 |
| 35 | engineering.fb.com/category/ai-research | 82 | 84 | 79 | 75 | 1,900,000 |
| 36 | datacamp.com/blog | 82 | 84 | 79 | 75 | 8,000,000 |
| 37 | neurips.cc | 82 | 84 | 79 | 75 | 1,900,000 |
| 38 | netflixtechblog.com | 81 | 83 | 78 | 74 | 1,200,000 |
| 39 | datarobot.com/blog | 81 | 83 | 78 | 74 | 1,400,000 |
| 40 | eng.uber.com | 80 | 82 | 77 | 73 | 1,100,000 |
| 41 | dataiku.com/blog | 80 | 82 | 77 | 73 | 1,800,000 |
| 42 | dataquest.io/blog | 80 | 82 | 77 | 73 | 1,900,000 |
| 43 | distill.pub | 80 | 82 | 77 | 73 | 980,000 |
| 44 | analyticsinsight.net | 79 | 81 | 76 | 72 | 2,400,000 |
| 45 | smartdatacollective.com | 79 | 81 | 76 | 72 | 950,000 |
| 46 | bigdatawire.com | 79 | 81 | 76 | 72 | 950,000 |
| 47 | techgig.com/ai | 79 | 81 | 76 | 72 | 2,400,000 |
| 48 | thesequence.substack.com | 79 | 81 | 76 | 72 | 680,000 |
| 49 | bair.berkeley.edu/blog | 78 | 80 | 75 | 71 | 620,000 |
| 50 | deeplearning.ai/the-batch | 78 | 80 | 75 | 71 | 1,200,000 |
| 51 | opendatascience.com | 78 | 80 | 75 | 71 | 480,000 |
| 52 | odsc.com/blog | 78 | 80 | 75 | 71 | 780,000 |
| 53 | blog.paperspace.com | 78 | 80 | 75 | 71 | 780,000 |
| 54 | emerj.com | 78 | 80 | 75 | 71 | 950,000 |
| 55 | marktechpost.com | 78 | 80 | 75 | 71 | 1,900,000 |
| 56 | paperspace.com/blog | 77 | 79 | 74 | 70 | 520,000 |
| 57 | dataversity.net | 77 | 79 | 74 | 70 | 840,000 |
| 58 | aitimes.com | 77 | 79 | 74 | 70 | 520,000 |
| 59 | techhq.com/artificial-intelligence | 77 | 79 | 74 | 70 | 950,000 |
| 60 | aimultiple.com | 77 | 79 | 74 | 70 | 1,400,000 |
| 61 | dataconomy.com | 77 | 79 | 74 | 70 | 520,000 |
| 62 | towardsai.net | 76 | 78 | 73 | 69 | 480,000 |
| 63 | fast.ai | 76 | 78 | 73 | 69 | 950,000 |
| 64 | aibusiness.com | 76 | 78 | 73 | 69 | 680,000 |
| 65 | unite.ai | 76 | 78 | 73 | 69 | 480,000 |
| 66 | thegradient.pub | 76 | 78 | 73 | 69 | 420,000 |
| 67 | lexalytics.com/blog | 76 | 78 | 73 | 69 | 420,000 |
| 68 | intelligenttransport.com | 76 | 78 | 73 | 69 | 380,000 |
| 69 | machinelearningtimes.com | 76 | 78 | 73 | 69 | 210,000 |
| 70 | ai.stanford.edu/blog | 76 | 78 | 73 | 69 | 260,000 |
| 71 | intelligenthq.com | 75 | 77 | 72 | 68 | 310,000 |
| 72 | artificialintelligence-news.com | 75 | 77 | 72 | 68 | 480,000 |
| 73 | syncedreview.com | 75 | 77 | 72 | 68 | 420,000 |
| 74 | kaggle.com/discussions | 75 | 77 | 72 | 68 | 600,000 |
| 75 | ai-techpark.com | 74 | 76 | 71 | 67 | 240,000 |
| 76 | aimagazine.com | 74 | 76 | 71 | 67 | 210,000 |
| 77 | wandb.ai/fully-connected | 74 | 76 | 71 | 67 | 380,000 |
| 78 | sebastianraschka.com | 74 | 76 | 71 | 67 | 320,000 |
| 79 | mlconf.com | 74 | 76 | 71 | 67 | 240,000 |
| 80 | montecarlo.ai/blog | 74 | 76 | 71 | 67 | 280,000 |
| 81 | theaisummit.com | 74 | 76 | 71 | 67 | 210,000 |
| 82 | thealgorithmicbridge.substack.com | 74 | 76 | 71 | 67 | 180,000 |
| 83 | ruder.io | 73 | 75 | 70 | 66 | 180,000 |
| 84 | lightning.ai/blog | 73 | 75 | 70 | 66 | 180,000 |
| 85 | machinelearningweekly.com | 73 | 75 | 70 | 66 | 210,000 |
| 86 | karpathy.github.io | 73 | 75 | 70 | 66 | 320,000 |
| 87 | colah.github.io | 73 | 75 | 70 | 66 | 240,000 |
| 88 | aijourn.com | 72 | 74 | 69 | 65 | 180,000 |
| 89 | accel.ai/blog | 72 | 74 | 69 | 65 | 180,000 |
| 90 | jeremyjordan.me | 72 | 74 | 69 | 65 | 150,000 |
| 91 | lilianweng.github.io | 72 | 74 | 69 | 65 | 260,000 |
| 92 | jack-clark.net | 72 | 74 | 69 | 65 | 95,000 |
| 93 | insideainews.com | 72 | 74 | 69 | 65 | 180,000 |
| 94 | ai-jobs.net/blog | 71 | 73 | 68 | 64 | 210,000 |
| 95 | lastweekin.ai | 71 | 73 | 68 | 64 | 140,000 |
| 96 | mlq.ai | 71 | 73 | 68 | 64 | 95,000 |
| 97 | ai-startups.pro | 70 | 72 | 67 | 63 | 120,000 |
| 98 | aidaily.co.uk | 68 | 70 | 65 | 61 | 180,000 |
| 99 | thezvi.substack.com | 72 | 74 | 69 | 65 | 160,000 |
| 100 | aiweirdness.com | 74 | 76 | 71 | 67 | 220,000 |
The pitch that works for AI editors
- Lead with the artifact. "We open-sourced a 2B model that matches Llama on X" or "here is a benchmark of 8 RAG frameworks on the same corpus." The thing you built earns the read.
- Show the code runs. A repo, a Colab, a reproduction script. Applied-ML editors reject claims they cannot check.
- Pitch the section. "For your applied research coverage" or "for the tutorials series" shows you read the publication.
- Match the format. News wants a tight story. Tutorials sites want a structured how-to with code. Research blogs want a paper-grade write-up.
- Bring the assets. Charts, a dataset, model weights, an eval harness.
- Disclose funding or a commercial interest in the model or tool you are writing about.
The test: would an editor run this if your company did not exist? If yes, pitch it. If it is a walkthrough of your API, it goes in the bin.
What not to do
- Pitch a "what is machine learning" explainer. Every site has ten, and it will not rank.
- Send benchmarks you will not share the code for.
- Submit an LLM-written post about LLMs with no experiments and no author who did the work.
- Put exact-match anchors ("best AI writing tool") on links back to your product.
- Overclaim in the title. AI editors and readers punish hype hard.
Bottom line: in AI, one placement on a respected applied-ML outlet with working code beats a dozen links from thin "AI news" sites. Use recent publishing activity and the traffic column to find the real ones.




