In short: On May 22, 2026, Airtable cofounder and CEO Howie Liu (@howietl) launched Hyperagent, Airtable's agent platform, together with the Founding 500, $10M in inference credits for 500 agent-first founders. We produced the launch film and ran the launch-day distribution at LaunchVideo.com. The post reached 14.37M views on X, with 5.4K likes, 629 reposts, 500 replies and 6.9K bookmarks, and we rated it 4.6 out of 5. It's the biggest launch we've produced, and it worked because the first three seconds told people who gets what.
Every number in this breakdown comes from the launch's page in our directory: launchvideo.com/directory/airtable-hyperagent.
Airtable Hyperagent launch at a glance
| Field | Value |
|---|---|
| Date | Friday, May 22, 2026 |
| Account | Howie Liu (@howietl), Cofounder and CEO of Airtable, about 362K followers |
| Offer | The Founding 500: $10M in inference credits for 500 founders (about $20K each) |
| Product | Hyperagent, Airtable's agent platform |
| Views | 14,374,722 |
| Likes | 5.4K |
| Reposts | 629 |
| Replies | 500 |
| Bookmarks | 6.9K |
| LaunchVideo.com rating | 4.6 out of 5 |
| Produced and distributed by | LaunchVideo.com (film plus launch-day distribution) |

Airtable: Hyperagent Founding 500 (about 14.4M views)
What did Airtable launch with Hyperagent?
Hyperagent is Airtable's platform for multi-step work you hand to AI agents instead of supervising them turn by turn. Each session runs in its own isolated environment with a browser, shell, filesystem, image and video generation, warehouse access and hundreds of enterprise integrations. Airtable named insurance brokerages, real estate firms and venture funds as early users.
The launch didn't open with features, though. It opened with an offer:
We're giving away $10,000,000 to founders building agent-first businesses. … We're backing 500 founders building them. The Founding 500.
So the post did two jobs at once. It made the case for a new kind of company run by agents, and it gave 500 founders a concrete reason to apply. That pairing became the spine of the script.
Why does a big CEO account still need a launch plan?
I built LaunchVideo.com around one idea: distribution is the product. A company can ship something brilliant and watch it die at 2,000 views because the right people never saw it.
Howie already had about 362K followers, so Airtable didn't have a reach problem in the usual sense. But an enterprise agent platform still has to earn attention in one of the noisiest feeds on X, and a big account on its own doesn't do that. This post reached roughly 40 times Howie's follower count. Followers don't explain that number. The offer, the film and the launch-day plan do.
How was the Hyperagent launch video structured?
The first three seconds of a launch video have one job, which is to stop the scroll. The Founding 500 film does it with a number and a who. Here's how the story ran:
The stake. "$10,000,000" and "agent-first businesses" up front, a category claim and a concrete offer in one line.
The thesis. Autonomous, proactive agents will run tomorrow's companies, and intelligence is no longer the bottleneck.
The program. 500 founders, inference credits and a fixed application window, which gave people a reason to act now.
The product proof. Hyperagent came out of Howie "burning through billions of tokens a week as a builder", and it gives you a fleet of specialist agents working on your own tools and data.
The ask. Apply through the links in the post.
Why did putting Howie on camera work for an enterprise product?
My standing advice to every founder is to open with a clear three-second hook and a human face. For an enterprise product, the face is part of how buyers decide whether to trust it.
- Authority. People evaluating an agent platform are buying judgment as much as software. The cofounder and CEO talking straight to camera carries more weight than a logo animation ever will.
- The numbers back it. In our directory, founder-led launches reach 1M views about 25% of the time, close to cinematic films at 28%. This one went more than 14 times past that mark.
- Personal proof. Howie's "billions of tokens a week" line makes Hyperagent feel like something he actually uses, not something he's pitching.
The Founding 500 would have been news as a screen recording too. The film is what turned Howie into the distribution point, and the post went out from @howietl, not from a brand account.
Does launch day matter? Why a Friday launch still hit 14M views
X shows every new post to a small test audience in its first 30 to 60 minutes, and how fast it moves in that window still decides how far it goes, even on a 362K-follower account.
Hyperagent went out on a Friday. Across our directory, Tuesday and Wednesday have the highest median views and weekends are the weakest, so Friday isn't a day I'd normally pick. What I take from it:
- A strong offer, the CEO on camera and a real first-hour plan can beat a weaker day.
- Friday isn't the new default. It's proof that when the story is sharp, the film and the amplification can outweigh the day of the week.
We handled the launch-day timing and the creator amplification alongside the film.
How did people engage with the Hyperagent launch?
I'm not going to invent reply quotes. What the numbers do show is the shape of the engagement:
| Signal | Hyperagent launch | Directory median | What it suggests |
|---|---|---|---|
| Replies per 10K views | 0.3 | 3.5 | Few replies per view at this scale |
| Bookmarks per 10K views | 4.8 | 14.4 | Lower rate, but a huge absolute number |
| Bookmarks vs likes | 6.9K bookmarks vs 5.4K likes | n/a | More saves than likes, so people meant to come back |
| Total bookmarks | 6.9K | n/a | A big save pile for an enterprise offer |
| Total replies | 500 | n/a | Still hundreds of conversations |
At 14M views, engagement rates always shrink. Our Launch Video Index found the same pattern across every 1M+ launch. What stands out here is that bookmarks beat likes. That's what founders do with a credit program they plan to apply to, not a post they simply enjoyed. Replies per view came in under the median, which fits a post asking people to apply rather than to argue.
How did we rate the Hyperagent launch video?
| Dimension | Score |
|---|---|
| Hook | 5 / 5 |
| Editing and creativity | 4 / 5 |
| Copy | 5 / 5 |
| Sentiment | 5 / 5 |
| Distribution strategy | 4 / 5 |
| Overall | 4.6 / 5 |
Hook and copy got full marks, which matches the Founding 500 open and the thesis behind it. Distribution scored a 4 and still produced one of the biggest launches in our directory, which tells you how high we set the bar for a 5.
What we learned producing the Airtable launch
Working on the Hyperagent film with Airtable, the biggest lesson was that the open matters more than the calendar. A Friday post reached the top of our directory because the first line was an offer people could act on, not a feature tour.
The second lesson was about who speaks. Agent platforms are dense and hard to explain in a few seconds. Putting Howie on camera gave people a human way into the product before the details, and it made the offer feel like it came from a builder, not a marketing team.
What can other enterprise teams copy from the Airtable launch?
- Lead with an offer your buyer can act on, like credits, a cohort or design-partner spots, instead of a feature list.
- Put the CEO or technical founder on camera when trust is what's holding buyers back.
- Write the three-second line before you write the shot list.
- Track bookmarks and clicks on your call to action, not reply counts. Hyperagent's bookmark numbers were the real signal.
- Plan launch-day distribution together with the film, not after it.
- Don't copy the Friday. Copy how clear the offer was.
This is also why we guarantee views. On a launch we commit to 1M, 3M or 5M views on the launch post, with a pro-rated refund if we miss. The Airtable launch shows what the top end of that creative and distribution work looks like on a CEO account.
Sources
Airtable Hyperagent launch page in the LaunchVideo.com directory, for views, rating and engagement
What Predicts a 1M-View Launch on X?, for format and launch-day data
The LaunchVideo.com Index 2026, for engagement at 1M+ views







