Introduction
During a recent training session on AI Fundamentals, an audience member asked me a question that sparked deep reflection:
“How much energy does a single chatbot query consume?”
As a researcher, I strive to base my responses on data. While I didn’t have immediate numbers, a quick on-the-spot search led me to an Euronews article that provided eye-opening statistics – and a gateway to a much bigger conversation on AI sustainability.
The Hidden Energy Cost of AI Chatbots
According to the Electric Power Research Institute, each ChatGPT request uses about 2.9 watt-hours, compared to roughly 0.3 watt-hours for a typical Google search. That means a single chatbot query consumes nearly ten times more energy than a search query.
With billions of chatbot interactions daily, this could amount to an estimated 10 terawatt-hours of electricity per year – a staggering figure that highlights the environmental footprint of AI.
It’s a reminder that while AI promises efficiency and innovation, its hidden costs must be part of the sustainability discussion.
AI Symposium 2025: A Focus on Sustainability
The very next day, I attended the AI Symposium 2025, where, by coincidence, the opening keynote addressed exactly this issue. The session, titled “AI for Tomorrow: Strategic Funding, Smart Policy, Sustainable Growth”, explored Europe’s evolving approach to responsible AI development.
Here are some of the key insights that stood out:
Hungary’s AI Strategy: From Technical Vision to Practical Application
Hungary launched its Artificial Intelligence Strategy in May 2020, modeled after Finland’s initiative. Initially, the plan focused on technical infrastructure and research. But by 2025, Hungary began shifting toward practical, real-world applications.
One highlight is the creation of an AI agent trained on data from 2008, designed to support HR, logistics, and business strategy. This system will be showcased at the Hungarian Chamber of Commerce, demonstrating how AI can integrate seamlessly into daily operations – bridging theory and practice.
AI Factories: Europe’s Push for Sustainable AI
Europe is taking sustainability seriously. The AI Factory Antenna Program and AI Factories are new initiatives that provide small and medium-sized enterprises (SMEs) with AI compute resources, allowing them to fine-tune and validate models efficiently.
These AI Factories act as hubs built around AI-optimized supercomputers, offering computational support and training opportunities across the EU. So far, 13 new AI Factories have been established, each offering free, customized assistance to startups and research teams.
Hungary’s Role in the AI Factory Network
Hungary is collaborating with Munich to create a high-performance computing hub as part of Europe’s broader AI ecosystem.
The EuroHPC initiative selected Hungary to host LEVENTE, a petascale supercomputer located in Szeged, which will become operational by the end of 2025 — a huge leap forward in regional AI capability.
A Risk-Based Approach to Responsible AI
The EU’s AI Act introduces a risk-based classification system to regulate AI technologies. Higher-risk applications face stricter oversight, ensuring accountability without stifling innovation.
Hungary plans to take this one step further with a national regulatory sandbox, launching in September 2025, where companies can stress-test AI models before releasing them to the public. This proactive approach ensures both compliance and innovation can coexist.
Challenges Ahead: Skills and Sustainability
Sustainability isn’t just about energy efficiency – it’s about human capacity too.
The EU Symposium highlighted a critical issue: the shortage of applied machine learning engineers and AI specialists. Without more trained professionals, maintaining innovation momentum will be difficult.
The solution? Invest in AI literacy, education, and collaboration between governments, academia, and industry to build a workforce capable of leading the next wave of sustainable AI growth.
Reflections on a Sustainable AI Future
As I left the symposium, one thought lingered:
AI is advancing at lightning speed – but the real challenge lies in balancing innovation with responsibility.
The world is moving toward a future where AI isn’t just more intelligent, but also more ethical, efficient, and sustainable. From data centers to classrooms, the way we design, use, and power AI systems will define our progress in the coming decade.
This is just the beginning – one spark of AI at a time.
