A case study in strategic GPU infrastructure partnerships and the power of transparent platform economics
Executive Summary
In an industry where 74% of companies report dissatisfaction with their current job scheduling tools and face resource allocation constraints regularly¹, DCENT, a sustainable edge datacenter operator, achieved remarkable results by partnering with Aethir's decentralized GPU network. Within months, DCENT reduced idle time by over 50%, increased revenue per GPU node by 30%, and positioned themselves for confident expansion—demonstrating how the right platform partnership can transform infrastructure economics.
The Industry Challenge
The AI infrastructure market faces a utilization crisis. Recent analysis shows that between 70-85% of current AI initiatives fail to meet their expected outcomes, with infrastructure issues being a primary contributing factor². More alarmingly, S&P Global Market Intelligence found that 42% of businesses scrapped most of their AI initiatives in 2024, up from just 17% the previous year³.
The GPU utilization statistics paint an equally concerning picture:
- 15% of organizations report less than 50% of their available and purchased GPUs are in use⁴
- Only 7% of companies believe their GPU infrastructure achieves more than 85% utilization during peak periods¹
- Nearly a third of organizations experience GPU utilization under 15%⁵
The DCENT Challenge
DCENT is a sustainable edge datacenter operator headquartered in the Netherlands, focused on providing high-performance infrastructure for AI and Web3 workloads. Operating facilities in Amsterdam, Frankfurt, and London, with expansion plans for New York and Paris, DCENT serves global clients including decentralized storage networks, AI model developers, and GPU compute platforms.
Like many infrastructure providers, DCENT faced significant utilization challenges. Their high-performance GPU systems frequently operated at suboptimal capacity.
"We dealt with underutilized GPU capacity and unpredictable demand," explains Hidde Hoogland, DCENT's Founder and CEO. "Hardware often sat idle despite strong global compute needs, especially during off-peak client cycles or between contracts."
The Strategic Decision
Traditional platforms presented significant limitations for DCENT's operations. "Hyperscalers impose high costs, slow deployment cycles, and lack the flexibility needed for fractional GPU sharing or short-term high-performance workloads. They don't cater well to niche or bursty AI/Web3 demands, making them inefficient for our operation," Hoogland notes.
Rather than accepting these constraints, DCENT's leadership prioritized finding a partner that could offer:
- Operational transparency
- Technical integration quality
- Long-term ecosystem alignment
- Sustainable economic models
The Aethir Partnership Solution
DCENT selected Aethir's decentralized GPU network for its unique approach to infrastructure economics and operational excellence.
"The team was one of the best and most responsive we've worked with—debugging is always a pleasure," Hoogland explains. "The offering stood out for its simplicity, transparency, and fair economics. Aethir, with its base fee, already reduces our OpEx by covering idle time. When workloads are active, the model delivers sustainable margins for us and our investors."
Hardware Deployment
Through Aethir, DCENT has deployed:
- Large batches of T4 GPUs tailored to custom gaming workloads
- Numerous RTX 3090 and 4090 systems
- AMD EPYC CPUs in high-speed environments (25–100 Gbps uplinks)
- Infrastructure housed in Tier 3 conditions with global routing capacity
Key Performance Improvements:
- 50%+ reduction in idle time
- 30% increase in revenue per GPU node
- Significantly reduced hardware ROI timeline
- GPU utilization consistently above industry averages
"Aethir has dramatically improved our efficiency, helping us maintain GPU usage well above industry averages," Hoogland confirms. "The platform's automated workload scheduling reduces the need for manual intervention and enables faster revenue generation from new deployments."
The Transparency Advantage
Unlike traditional cloud providers with complex fee structures, Aethir's model provided DCENT with clear visibility into all cost components. This transparency proved crucial when compared to providers like AWS, which charges $0.09 per GB for the first 10 terabytes of data egress⁶, with costs scaling significantly for data-intensive AI workloads.
Oracle's analysis demonstrates that hyperscalers apply very high margins on egress costs despite decreasing internet transit prices over time⁷, creating substantial hidden expenses that only emerge after contracts are signed.
Ecosystem Benefits
The partnership opened new revenue streams and market opportunities for DCENT beyond the expected efficiency gains.
"Aethir introduced us to new client verticals and ecosystem partners, provided global exposure, and simplified many parts of our daily operations," Hoogland explains. "Being part of a shared compute network also strengthens our positioning in Web3 and decentralized AI."
Diversified Workload Portfolio:
DCENT's infrastructure now supports:
- AI inference workloads
- Generative model testing
- Gaming servers
- Stable diffusion applications
- Decentralized compute platforms including blockchain validators and render networks
Strategic Alignment and Future Growth
The Aethir partnership aligns perfectly with DCENT's strategic vision. "Aethir's model is directly aligned with our goals: distributed, low-latency compute that scales elastically with demand. It supports diverse workloads from AI to blockchain compute while helping us maintain high hardware efficiency and power sustainability targets," Hoogland states.
This alignment has enabled confident expansion:
- Expanded cluster sizes
- Added new GPU SKUs
- Preparing new deployment zones beyond Europe
- Currently onboarding 100 RTX 5090 GPUs optimized for inference customers
"The demand through Aethir allows us to confidently grow our infrastructure footprint," Hoogland confirms.
Market Impact and Competitive Advantage
Partnership with Aethir has delivered benefits beyond operational metrics, enhancing DCENT's market position.
"Being part of Aethir adds credibility to our capabilities and demonstrates that we're integrated into a forward-thinking, scalable compute ecosystem. This helps in both customer conversations and investor discussions," Hoogland notes.
Key Takeaways for GPU Infrastructure Providers
The DCENT case study illustrates several critical lessons:
1. Total Cost of Ownership Matters
With some providers offering 25% lower bandwidth costs and 100x more free data transfer than competitors⁷, transparent, all-inclusive pricing models deliver superior economic outcomes despite potentially higher base rates.
2. Operational Excellence Drives Success
Technical responsiveness, automated scheduling, and reliable support create value that exceeds marginal pricing differences.
3. Ecosystem Access Creates Strategic Value
Access to diverse workloads and new market segments generates opportunities beyond traditional infrastructure monetization.
4. Confidence Enables Growth
Predictable demand and transparent economics enable data-driven expansion rather than speculative scaling.
The Bottom Line
"Aethir helps us turn GPU infrastructure into a reliable, high-yield asset. Their responsive team, streamlined platform, and demand-side access make it easy to grow profitably in the AI and decentralized compute space," states Hidde Hoogland, Founder & CEO of DCENT.
With 74% of respondents seeing value in unified AI/ML platforms but only 19% actually having effective scheduling tools⁸, the market clearly demands integrated solutions. The DCENT-Aethir partnership demonstrates how providers can successfully address these gaps.
"It's rare to find a compute marketplace that's this efficient and collaborative. Aethir has become a key strategic partner, not just a platform we plug into. We're excited to keep growing with them," Hoogland concludes.