Best AI Influencers in the US (2026): Who to Follow and How

If you want to keep up with AI from the US without losing your evenings to it, choose the format first and the person second. A newsletter you open daily beats a brilliant YouTube channel you never get round to, so the list below is grouped by how you'll take it in, and the guide around it is about choosing well and dropping people who aren't earning their place. Everything said about each person comes from their own site, read as of Nov 2026.

Follower counts are left out on purpose. Reach tells you how many people read someone, not whether they'll help your business. If you'd rather see names sorted by the job they do for an owner, my shortlist of AI influencers for business owners takes that angle, and a few people appear on both pages.

What kind of AI voice do you need right now?

Most people who follow AI accounts want one of three things, and they rarely say which. Some want to know what's new, so they aren't caught out by a client or a competitor. Some want to understand how the technology works, so they can judge claims and stop being sold to. And some want something to try at work on Monday. Those are three different jobs, done by three different kinds of creator, and a feed built for one is a poor fit for the others.

The news job is done best by daily digests: short, broad, and written so you can skim them over coffee. The understanding job belongs to educators and researchers, whose posts are slower and longer and sometimes technical. The Monday job belongs to practitioners who publish prompts, automations and workflows. Work out which job you're hiring for this quarter and you'll know which section below to start in.

Daily and weekly reads for your inbox

Rowan Cheung and The Rundown AI

A five-minute daily email on AI news, why each story matters and how to apply it, with separate AI, Tech and Robotics editions.

Ethan Mollick

Wharton professor whose One Useful Thing newsletter takes a research-based view of what AI means for work; essays to sit with, not quick tips.

Greg Isenberg

A weekly letter of startup ideas and frameworks from a serial community-business founder, most useful when you're building something new.

Sabrina Ramonov

A Substack sending prompts, agents and automations to subscribers, the hands-on choice for a solo owner who wants something to try today.

Andrew Ng

DeepLearning.AI founder and Stanford adjunct professor whose weekly letters and courses ground you in how AI works rather than how to market with it.

Channels to watch

Matt Wolfe

San Diego YouTuber behind FutureTools.io and The Next Wave Podcast, covering new AI tools and news for people who'd rather watch than read.

Harper Carroll

Former Meta AI engineer whose videos translate AI into plain English, a sound starting point for teams that want to understand it before trusting it.

Andrej Karpathy

Long YouTube lectures on how language models work, including a general-audience track; his How I use LLMs video is the practical place to begin.

Feeds that are more than noise

Jim Fan

NVIDIA senior research scientist sharing AI research and industry insight on X and LinkedIn, giving a lab's view of where agents and robotics are heading.

Allie K. Miller

Former AWS machine learning lead who advises large companies and posts about building with AI on Instagram, Twitter and TikTok, with an enterprise lens.

One for your headphones

Dwarkesh Patel

Host of the Dwarkesh Podcast, deeply researched long-form interviews about where AI thinking is going; better for a long drive than for Monday's to-do list.

Where I fit in, and where I don't

Read this entry most critically of all, because I'm the one who chose to put it here. I'm British, I work in English with clients worldwide, in person and online, and I'm not a creator in the way the rest of these names are. I don't publish a daily news email, a YouTube channel or a podcast; if that's what you want, the people above do it better. I'm not a research scientist either. I don't train models or run a lab.

What I do is help owners and marketing teams use AI in their real work. I've been in marketing for 23 years, went all in on AI in 2024 and rebuilt my own consultancy around it, and I've spoken at more than 100 events, including TEDxBolton and Oxford University. My sessions are hands-on and built on your team's own tasks: a 90-minute workshop, a 45 to 60 minute keynote for a bigger group, or a series of shorter sessions over a few weeks, ending with a one-page agreement on safe AI use. My newsletter goes to an audience of 15,000.

I test all of this on my own business first. Between January and September 2026 my site went from 1,322 Google clicks a month to 6,412, and visits sent by AI assistants went from 32 a month to 3,599, from my own Search Console and GA4. If you're booking for an event rather than looking for someone to follow, my list of AI keynote speakers in the US is the better page.

Four forks that sort any AI follow list

News or understanding. A digest tells you what happened; an educator tells you why it works the way it does. You need a little of both, but if you only have room for one, pick understanding. News goes stale in a week, while knowing how a language model reaches an answer changes how you write every prompt from then on.

Builder or operator. A lot of the loudest AI voices in the US write for founders starting something new. Their ideas are good, but if you run an established service business, every post needs translating into your world. If that translation keeps not happening, swap to someone who writes for people running a team rather than raising a round.

Lab or boardroom. Researchers post about where the technology is heading next year. Enterprise advisers post about what big companies are rolling out now. A small business usually needs neither at full strength: take the research view to know what's coming, take the boardroom view for governance ideas, and scale both down.

Short or long. Short posts are easy to consume and easy to forget. Long essays, lectures and interviews ask for an hour and give you something you'll still use in six months. Most people over-follow short and under-follow long. Flip it, and you'll learn more from less.

Influencer or expert, and why the label matters

An influencer is someone whose reach shapes what people read about AI; an expert is someone whose judgement you'd act on. Plenty of people are both, but not everyone. The question I'd ask of anyone you follow for AI advice is simple: how do they use it in their own work? One of the best answers on this page belongs to Ethan Mollick, whose About page says he writes every post himself and only asks AI for feedback on a finished draft. For the research and advisory side, my shortlist of AI experts in the US is sorted differently and names different people.

A one-hour clear-out of who you follow

Set a timer and do this once. It's the quickest way to turn a noisy feed into something that helps.

1. Spend ten minutes listing every AI account, newsletter, channel and podcast you follow, and next to each write the format and the last date you used something from it.
2. Cross out anything you haven't opened in a month. Be ruthless; you can always come back.
3. For what's left, mark each one News, Understanding or Monday. If one job has five names and another has none, that's your gap.
4. Pick one name from this page to fill the gap, in the format you already open every day, and give it four weeks.
5. Write down one thing you tried because of each person. At the end of the four weeks, anyone with a blank line goes.

You should end the hour with three or four sources, one per job, and a note to check back in a month. If that feels thin, it isn't. Three sources you use beat thirty you scroll past.

Mistakes that make following AI people a waste of time

• Following for the news only. You'll always feel up to date and never be any better at the work.
• Copying a founder's tool stack. What suits a two-person startup chasing growth rarely suits a team of twelve with clients and contracts.
• Treating follower count as proof. Big audiences reward speed and novelty, and the people who explain things carefully are often smaller.
• Saving posts instead of trying them. A bookmarks folder full of prompts is a to-do list you'll never do.
• Following the same story from five angles. When a big release lands, every account covers it. One good write-up is enough.

What learning AI this way costs

Almost everything these people publish costs nothing to follow: the newsletters, the videos, the posts and the podcast episodes. The cost is your attention, and that's the budget to manage. The paid layer sits on top, in the form of courses, academies, cohort programmes and private communities, and pricing for those changes often enough that I'd check each one on its own page rather than trust a figure on a list. A sensible rule is to pay only once a free source has already changed how you work, because then you know the teaching style suits you.

For a team, the bigger cost is time. Ten people each following their own set of creators means ten different views of what AI can do and no shared way of working. That's usually the point at which it pays to bring the team together around its own tasks rather than everyone's favourite feed.

My pick for each situation

Short on time and running a small business: one daily digest and nothing else for a month. Leading a team that's nervous about AI: a plain-English video educator, watched together, then a long lecture on how the models work for whoever wants to go deeper. Building a new product or offer: a founder-focused weekly letter. Responsible for policy or governance in a larger company: an enterprise adviser's feed, scaled down to your size. Curious about where all this is heading: one long-form interview podcast for the commute. And whatever you choose, keep it to three.

Following gets you ideas; putting them to work on your own team's tasks is a different job. I run AI training and workshops for teams, and help businesses with AI enablement and workflow automation, so the busywork runs itself and people get time back. Most of that work is with SaaS marketing teams and small businesses. Here's how we can work together.