Scaling a hardware startup with zero VC intros | Kostas Koukoravas

Funding stage: Pre-seed. Qalzy announced a pre-seed round backed by Jenson Ventures in spring 2026, following a six-figure Kickstarter that functioned as the demand proof. No priced Seed round is public and the round size is undisclosed.

Kostas Koukoravas came out of lockdown about ten kilos heavier, having taken up baking like half the planet. So he did what a former Microsoft product manager does. He started logging his food. MyFitnessPal first, which meant weighing everything and then typing it in, which meant quitting. Then the AI photo apps arrived and made logging fast, which was great until he found the research: photo estimation nails what the food is and badly misses how much of it there is. Qalzy cites error rates reaching 65%.

Kostas is the co-founder and CEO of Qalzy, aiming to become what he calls the Oura of nutrition: one device in every kitchen. Before this he led product at the BBC and Microsoft, where he worked on Skype for Windows 8 and Surface Hub, and founded the AI fashion styling platform Intelistyle.

Here is the part that turns a nuisance into a business. Those misses do not cancel out. Two hundred calories on breakfast, two hundred on lunch, two hundred at dinner, and by bedtime you have invisibly eaten an entire extra meal. That is the difference between losing weight and wondering why you are not.

So Kostas and his co-founder Michael Michelis, a nutritionist who runs one of the biggest calorie-counting communities online, built a kitchen scale with an AI camera bolted to it. Put food on the glass, press one button, and it identifies the ingredient, weighs it to the gram, and logs the calories and macros against a database of 2.9 million verified foods. No phone. Which matters more than it sounds, because the moment you are actually cooking is the moment your hands are covered in raw chicken and your phone is the last thing you want to touch. Ryan, a serious home cook who has been tracking macros since the spreadsheet era, immediately gets it: with a scale, logging happens while you cook, not from a plated photo that has no idea how much butter went in the pan.

The receipts are early but loud. Kostas says users with the scale stick with tracking roughly 60% longer than app-only users. Returns have been close to nonexistent since launch. And Ryan's own story is the perfect control group: 35 pounds down between September and April, then six flat months of what he cheerfully calls gaslighting himself about portions.

Then the conversation turns into a full hardware masterclass, and this is the part every founder should sit still for. Two software guys who were used to pressing a button and shipping code spent a year on prototyping alone. Kostas's line of the episode: plastic is alive. Prototype resins behave nothing like production ABS, tolerances drift with humidity, and a high gloss finish means every stray thumb during assembly is a scratch you just paid for. Their factory team in China ground through iteration after iteration to get it right.

They funded it the way more hardware founders should. Kickstarter, where a six-figure pre-order haul proved demand a VC would have asked them to prove first, covered the molds and the minimum order quantities, and doubled as a product research panel. The multi-user feature, now nine profiles per device, only exists because backers kept asking whether their spouse could use it. Glucose response tracking in the app exists because type 1 diabetics asked for it. Stretch goals became a feature vote.

Now the pull is coming from directions they did not plan. Hospitals. The NHS, which cares about prevention over cure. Corporate canteens that want to show employees what lunch actually costs them. A restaurant owner who wants accurate nutrition facts on her menu. Ryan pushes on the commercial angle and the food-cost math behind it, and Kostas does not take the bait. Consumer first. The goal, in his words, is to become the Oura of nutrition, one device in every kitchen.

Also covered: whether GLP-1 users abandon calorie counting (Qalzy ran an ad assuming yes and got a polite beating from customers who do both), why carbs at night are not the villain, how his Greek mother's cooking changed once she learned what a spoonful of olive oil costs, and why the win condition is logging every day rather than logging perfectly.

Qalzy is about $199 with a launch discount running, a 30-day money-back guarantee, and a one-year warranty. Kostas and Michael personally answer every email that comes in.

If you're building an AI workflow that depends on a human doing something every single day, the lesson here is worth more than the scale.