Comment by davidhunter | Hacker News Reader

Comment by davidhunter | Hacker News Reader

We (Optimal Agriculture) have built an autonomous growing system for high-tech greenhouses and have achieved 'superhuman' performance in a recent tomato trial. We collect a bunch of sensory data about the crop and growing environment, run cloud-based optimisation, and then control the actuators of the greenhouse autonomously (lighting, screening, heating, cooling, irrigation, etc). We are turning $100M greenhouse facilities into IoT-connected-devices and then building an intelligence layer on top.

We are now building a commercial-scale facility in the US which will produce 50,000kg of fresh produce daily. And then we will scale from there.

Our goal is to accelerate the deployment of new greenhouses around the world to increase the availability of healthy locally-grown food and to secure our food system for the future.

Find out more about our mission: https://www.youtube.com/watch?v=LDJ_QdUaap4

If you want to find out more about us or generally want to geek out on tommies then give me (founder) a shout: david@optimal.ag

Just for the density of jargon, this is quite an amusing post. I just hope your project is actually working out, though I can't tell from what you wrote.

How many square meters will that facility be, and what is the cost to construct? At $100M I don’t see how it would be financially viable.

50k kg a day is approximately 600 acres of tomatoes with our yields. That much land in CA might sell for $6-30M depending on water rights and latitude.

We’ll be getting over 100kg/m2/year. Construction varies but typically $450/m2 for a lit greenhouse.

Definitely a good opportunity for selling the higher value tomatoes restaurants and some upscale markets desire. I like the emission benefits of shorter supply chains too. Best of luck, I doubt our tomatoes will be competing with each other anytime soon.

He is doing tomatoes eaten out of hand vs your processing tomatoes. They might as well be different vegetables.(or fruits as this is a tomato)

Agreed. That’s why I said I think there is room for us both for a long time.

You may be interested to know about Costas Group in Australia. They have gone the high tech route, cramming every bit of productive technology into their greenhouses. They are an established company listed on Australia's stock exchange (ASX:CGC), so can be taken to have proved the viability of such technology. I gather their share price and the size of their profit is a bit of a roller coaster ride, but they do make a profit each year. They mention yields of 70kg/m^2/year (=50000kg/day from 65 acres). I assume that's per square metre of greenhouse and not per square meter of growing bed.

https://thefarmermagazine.com.au/the-costa-group-is-a-runawa...

https://www.tradefarmmachinery.com.au/features/1408/costa-gr...

Edit: Article also mentions a 100kg/m^2/year target.

I have been somewhat obsessed with the idea of floating greenhouses. The economics are different than operating a farm on land but with automation it could be possible to deliver superior produce directly to coastal population centers at lower costs.

It will fail. Other than the novelty of floating greenhouses, how do you think this will work?

Where would you get water? Desalination is still terribly energy intensive.

How do you model the crop response to environmental factors? Did you have to run experiments?

Take everything we know from plant sciences as a starting point and then generate additional data through actually growing the crop and measuring impulse/response with sensors.

Are you optimizing for any experiential aspects of your tomatoes, like taste or texture? Personally, in my area tomatoes are cheap enough to not think about the price, but the products widely available from supermarkets taste like water

> Are you optimizing for any experiential aspects of your tomatoes, like taste or texture?

Yes - but this mainly comes down to the varieties that you grow. You can grow varieties that are very tasty and are usually sold in the premium tier. And you can grow high-yielding tomatoes with less taste.

>Personally, in my area tomatoes are cheap enough to not think about the price, but the products widely available from supermarkets taste like water

Where are you based? High-tech greenhouses already supply a lot of premium-tier produce in Europe but there is an undersupply in the US/Asia.

Interesting. I've always wondered why tomatoes in the US taste like shit. Why is there a trade off between high yield and taste? Is there not a high yield + great taste variety?

I think its largely about shelf life and bruising?

there is this kind of tomato I buy if there aren't any others. it always looks perfect on the outside...and seems to last forever sitting on the counter at home.

but if its old enough, the seeds will have sprouted when you cut into it.

pretty tasteless regardless

What do you mean by "cloud-based optimization?" Are you running reinforcement learning algorithm(s) on these plants?

Yes we use techniques from RL as one of our technologies - its my personal research background

cloud based optimization is superior to on organic compute optimization how?

Do you know Sfera Agricola in Italy?

Reply on news.ycombinator.com