\n\n\n\n Liquor Stores, Drone Defense, and a Donation That Lost Three Zeroes - AgntBox Liquor Stores, Drone Defense, and a Donation That Lost Three Zeroes - AgntBox \n

Liquor Stores, Drone Defense, and a Donation That Lost Three Zeroes

📖 4 min read•749 words•Updated Oct 13, 2026

$220 million went to a single company last week. Capitolis, a Series E round led by Citi Ventures, pulled in more capital than every other deal on the 10/12/26 notable funding lists combined — and it did not come with a shiny AI demo video attached.

I spend most of my week testing AI tools and writing down which ones survive contact with real work. Funding reports are not my usual beat. But I read them the way a mechanic reads auction listings: not for the hype, for the signal about what buyers actually believe is worth money right now. The week of October 12 was instructive, mostly because of what is missing from the top of the list.

What actually got funded

Here is the week as reported, stripped of commentary:

  • Capitolis — $220M, Series E, led by Citi Ventures
  • AMP Robotics — $70M, debt, from Galvanize Climate Solutions
  • Fortem Technologies — $50M
  • Vitalize — $31M
  • Pro Shop — $24M, Series B, led by Arthur Blank Sports and Entertainment
  • Santé — $15M

Santé is the one I keep thinking about. New York-based, founded in 2023 by Antoine Balaresque and Darren Fike, and the product is inventory management, e-commerce, and industry-specific tooling for liquor stores. That is it. Not a general-purpose agent platform. Not a model. Software for people who sell bottles.

Narrow beats clever, again

The pattern in my own tool reviews is boringly consistent. The products that score well are the ones that picked a specific job and refused to expand until that job was handled. The products that score badly are the ones that could theoretically do anything and reliably do nothing.

Santé at $15M is a wager on the first kind. A liquor store has strange problems — state-by-state compliance, SKU counts that look absurd next to a normal retailer, a mix of walk-in and delivery demand. A horizontal tool handles none of that well. A vertical tool can encode the rules directly and stop asking the operator to configure them.

This is the same reason I keep recommending unglamorous, single-purpose AI tools over the all-in-one dashboards. Narrow scope lets you write real defaults. Broad scope forces you to push every decision back onto the user, then call it flexibility.

AMP Robotics raising $70M as debt rather than equity is its own kind of tell. Debt financing generally wants predictable cash flow behind it. You do not lend against a promise of future intelligence; you lend against machines that are already running and already being paid for. If you are evaluating automation vendors, the financing structure is a cheap proxy for how real the deployments are.

Fortem Technologies at $50M, in counter-drone systems, points at a category where the buyer has no tolerance for “mostly works.” That constraint tends to produce better engineering than the consumer side of the market, where a 70% success rate still gets a product launch.

A quick note on reading these reports carefully

One item on the week’s roundup deserves a flag, and I’d rather be the person who says it out loud. A donation to The Street Dog Coalition from The Veterinary Cooperative and Elanco Animal Health is being passed around as $14.088 million. The Business Wire release describes a combined $14,088 donation.

Those are not the same number. One is a mid-size philanthropic commitment, the other is a community group chipping in. A decimal point and a unit label got shuffled somewhere in the aggregation chain, and then the wrong version traveled.

I bring it up because this is exactly the failure mode of the AI research assistants people now use news. They ingest a feed, they normalize numbers into a familiar shape, and “$14,088” looks suspiciously like a malformed “$14.088M” to a system trained on funding announcements. The output reads clean and confident. It is also wrong.

If you are testing summarization tools, this is the kind of case to throw at them. Not “summarize this article” — anyone can pass that. Give them a page with an unusual magnitude and see whether they round it into something more comfortable.

What I’d take away

Three things. Capital is still flowing toward specific, verifiable work: financial infrastructure, recycling robots, drone defense, liquor store inventory. Debt rounds are worth more attention than equity rounds when you’re judging whether a product is deployed or just demoed. And the reporting layer on top of all this is noisy enough that a four-figure donation can get inflated by a thousand times before anyone blinks.

Test the tools you trust with these numbers. Then decide what they’re worth.

🕒 Published:

🧰
Written by Jake Chen

Software reviewer and AI tool expert. Independently tests and benchmarks AI products. No sponsored reviews — ever.

Learn more →
Browse Topics: AI & Automation | Comparisons | Dev Tools | Infrastructure | Security & Monitoring
Scroll to Top