PDF

We're pitching Gratheon to Tehnopol's AI Accelerator program with the following structure:

1. 1min for elevator pitch + team introduction
2. 1min for business plan
3. 1min for 6-month plan, detailing what we aim to achieve during the AI Accelerator program (month by month)

AI Accelerator Plan​

Our primary goal is to make Gratheon production- and GTM ready (securing first paying customers!) by focusing on targeted product development.

We've divided our 6-month plan into three key product areas, allocating 2 months to each:

IoT Sensors Device (Months 1-2)​

• Implement configurable alerting for anomalies (e.g., sudden weight loss)
  • Add SMS alerting channel via Twilio integration
• Integrate Prophet AI model to predict future hive weight based on:
  • Historical weight data
  • Weather data
  • Entrance observer bee counts
  • Potential future application: bee colony demography forecasting

Why: This AI-driven approach differentiates us from existing IoT solutions in the beekeeping market and adds significant value for beekeepers.

Web Application (Months 3-4)​

• Enhance queen detection from frame photos
  • Current limitation: Clarifai-trained model with small dataset and suboptimal precision
• Resolve AI model stability issues (resource constraints)
  • Evaluate Clarifai cloud hosting
  • Test models on our own infrastructure
  • Optimize for CPU usage
• Fix queen toggling functionality
• Improve basic statistics display and data visualization

Why: These improvements will increase the value proposition for customers who manage apiaries through frame photos, providing meaningful statistics and insights that drive initial customer acquisition.

Entrance Observer (Months 5-6)​

• Build a working pipeline for bee counting on the edge (NVIDIA Orin)
  • Reference: Issue #2
  • Reference: Issue #6
• Develop telemetry API integration with web application
• Implement real-time count display in web interface
• Validate real-time inferencing capabilities
  • Reference: Issue #8

Why: This creates a new product offering (entrance camera + GPU) with minimal development effort, while establishing a foundation for future iterations on more cost-effective and energy-efficient devices (ESP32 cam, Raspberry Pi).

Expectations from AI Accelerator​

• Refine product value proposition and messaging
• Improve business model (pricing strategy, GTM approach, unit economics)
• Technical guidance for AI implementation (addressing the challenges outlined above)
• Intellectual property support (logo design, trademark registration)
• Preparation for investor due diligence
• Community engagement and networking opportunities

Visited event organized by Tehnopol. We've submitted application to be in the next batch. Had networking with local AI community

Had a meeting with VIRUMAA DIGI- JA ROHETEHNOLOOGIATE INNOVATSIOONIKESKUS representatives (engineers, researchers and managers) on the topic of research and development collaboration with Gratheon. Seems like the most suitable topic is the area of IoT devices.

We talked about the project and its market, past experience of the team, funding sources, some technical details.

Specifically

• local team used nrf9160 chip (instead of esp32) to be more energy efficient
• MQTT and GSM were brought up as possible communication means for the device
• https://tradex.ee/ was used as a partner to make devices more prod ready (electronics)
• nokia dynanet - as a construction example of expandable frames and nest was brought up for robotic beehive construction
• we should have a small research paper abstract published

Led by Reet Allik and Lauri Vilibert, this webinar explains how robotics is being taught in local Estonian schools from the first grades and higher. https://www.facebook.com/events/1226346678472001

In short, schools use Lego Spike Prime. More advanced VEX modules were recommended too.

We took part in Founder Institute online 1 min pitch practicing and getting feedback from the mentors. Regine Harr gave a good advice on elevator pitch structure.

We now have shorter and more emotional elevator pitch structure. We tested it on other mentors and got positive feedback.

We got questioned by Dr. Raashid Saiyed on why we are open source, 📊 Market analysis and

We're excited to announce that we join NVIDIA Inception Program!

As we are relying on edge devices for video monitoring of bees and we are using NVidia Jetson Nano and Orin, it is natural for us to seek more knowledge and deeper research, gain pricing benefits on devices and have better reach to NVIDIA associated communities of engineers, startups and potential investors. This will especially crucial for robotic beehive.

According to Program Terms and Conditions, we may benefit from:

• credits to NVIDIA Deep Learning Institute (DLI) offerings
• preferred pricing on a wide range of products and platforms
• listing of your application in the applicable public NVIDIA catalog for visibility of your application
• NVIDIA facilitating introductions to third parties such as Investors/VCs
• brand exposure opportunities

We are excited to participate in the "Data Storm" hackathon, which focuses on utilizing open government data from Estonia, particularly in areas such as environment, weather, hydrology, and alerts. Our project addresses challenges related to storm winds collapsing beehives and proposes a solution in the form of weather alerts.

Our focus areas​

• 🗃️ Storm wind collapsing beehives
• 📡 Weather alerts

Team members​

• Artjom Kurapov
• Alonso Solis
• Adrian Ala
• Roop Runjan Khan

Participants and mentors collective foto by Tiit Tamme / Garage 48:

First pitch slides​

Meetup - https://www.meetup.com/openuk/events/301997990/?eventOrigin=home_page_upcoming_events$all

Tim Telford​

• https://github.com/devtank-ltd Use cases:

Gerrit Niezen​

https://gerritniezen.com/openuk-slides

Jenny Molloy​

• https://forum.openhardware.science/
• https://www.openplant.org/openmta/
• wetware
  • DNA
  • tests

https://storymaps.arcgis.com/stories/3e34000da265450bb7e4db54b03dde7c

Attended JAMK workshop that is looking for precise pollination and food-related startups.

AniCare was one of the example startups that was featured

https://www.jamk.fi/en/jamk/focus-areas/bioeconomy/bioboosters https://www.jamk.fi/en/project/agriventure-finland-2025

From startups that BioBoosters has supported, I liked:

• https://antrobotics.de/
• https://geopard.tech/
• https://lensor.io/lensor-application/
• https://soilscout.com/
• https://www.trabotyx.com/
• https://weedbot.eu/

https://www.meetup.com/startup-investor-accelerator-estonia/events/303277579/

Visited local k-space (hacker space) co-working space with various equipment opportunities. Was mainly experimenting with sensor telemetry as a product

Got some follow-up comms later

Gave an remote internship for QA engineers to practice finding bugs, organized by Tallinn Learning school

Said not yet, maybe potential seed round

Huge event, maybe too early for the project

Networking​

• with Jabok C. But he is pitching Austria as a place for startups + https://viennaup.com/
• with Janne A. He is pitching intellectual property protection
• networking with Kristjan. Engineer.

Pitches​

• angel investor Perttu I. Feedback was its a bit early for him, but interesting project
• vc fund byFounders with Magnus H.

Deck sent to​

• Nikko P., Aruma Ventures
• Serena K., Captium Group
• BDPartners.cz
  • feedback - no, they are not experts in this field
• Maria B. STRT Holding
• Iryna S. Memory Bank
• Anssi U. Nordic science investment

Streams​

DAY1: https://youtube.com/live/SHQgMd1SZVY https://www.youtube.com/live/jlp0uX0R_jE https://youtube.com/live/ljVjzLFLTWk

DAY 2: https://www.youtube.com/live/3vKvO7kIR74 https://www.youtube.com/live/50ciOuGC3ig https://www.youtube.com/live/o0R3yNvUGXk