FAQ

1. What is Exabits, and what do you do?

Exabits is a leading provider of foundational AI compute infrastructure, specializing in the optimization, management, and deployment of high-performance GPU clusters. We act as the "refinery of compute," transforming raw GPU hardware into fully optimized, reliable, and scalable compute services. Our platform enables businesses and innovators to access and use the GPU resources they need for AI projects, without the complexities of building and maintaining their own infrastructure.

2. What are the core use cases of Exabits?

Exabits is designed for GPU marketplaces, AI enterprises, research labs, blockchain networks, and gaming companies that need scalable GPU compute for:

• AI Model Training and Serving

• Rendering & 3D Graphics

• Blockchain Validation & Web3 dApps

• High-Performance Computing (HPC)

• Cloud Gaming & Metaverse Applications

3. Who are your customers?

Positioned uniquely at the base layer of the AI compute ecosystem, we bridge the colossal gap between chip manufacturers and the end users by proprietary technology and tokenized business model.

• GPU Marketplaces

• AI Companies

• Gaming Companies

• Blockchain Companies

• Industrial enterprises

• And More

4. What GPU types does Exabits offer?

Exabits offers a variety of GPUs suitable for different AI workloads. These include H200, H100, A100, 4090, L40, Ada6000, A6000, A5000, A4000, L4, T4, P4, V100, and upcoming 5090 and B200 GPUs. Whether you need the latest high-performance models or more cost-effective options, we provide the best fit for your requirements.

5. How fast can I get access to Exabits’ GPU clusters?

Exabits offers an on-demand platform and API, allowing users to begin utilizing our GPU services immediately after registration. For long-term clients, we provide personalized, one-on-one service to meet custom requirements and ensure tailored deployments.

Additionally, for larger-scale clusters, such as those with 1,000/2,000+ H100s, Exabits delivers industry-leading deployment speeds up to 1-2 months, ensuring that even the most extensive GPU setups are up and running quickly—far faster than the typical 4-8 month timeframe common in the industry.

6. Where are Exabits’ data centers located?

Exabits partners with 27 data centers globally, including in North America, Europe, and Asia. We have strategic partnerships with leading data center providers such as Racks Central, Equinix, Centersquare, and Digital Realty.

Additionally, we have a global delivery team that can deploy GPUs to client-specified locations based on their unique needs. This flexibility ensures that we can meet specific regional requirements or integrate seamlessly into existing infrastructure, giving our clients maximum control and customization over their compute resources.

7. How does Exabits ensure high-speed transmission between GPUs in the network?

Exabits provides high-speed networking solutions tailored to the needs of AI workloads. For high-performance GPUs like the H100/H200, we typically provide 1.6TB/3.2TB InfiniBand interconnection, ensuring fast data transfer across the cluster. For specific user needs, we also offer 400GB/800GB RoCE connections.

For cost-effective GPUs like the RTX 4090, we use 25GB-100GB Ethernet. For internet access, we standardly provide 1-10GB bandwidth shared bandwidth, which can be upgraded to 200GB dedicated lines based on requirements. We also offer custom-designed Data Center Interconnectivity (DCI) solutions for long-distance, high-bandwidth needs.

8. What is Exabits’ approach to reliability and uptime?

At Exabits, reliability is critical. Our GPU machines undergo rigorous burn-in testing with real-world workloads to ensure stability before they are deployed. Our monitoring systems continuously track computing, storage, networking, and security to ensure systems remain stable. We offer 95%+ to 99+% uptime for our clusters based on the customer's needs. This ensures that your AI workloads are not interrupted by system failures.

9. How does Exabits optimize GPU performance?

Exabits uses a comprehensive approach to optimize GPU performance, which includes hardware, firmware, and network optimizations. Our proprietary AIOps agents continuously monitor and adjust workloads in real-time, ensuring GPUs operate at peak efficiency. In addition, we fine-tune hardware settings, optimize firmware configurations, and adjust network parameters to maximize throughput and minimize latency.

By leveraging these advanced optimizations, we boost the performance of GPU clusters ensuring faster, more cost-effective AI model training while maintaining reliability and stability throughout the entire system.

10. Why is Exabits' compute platform different from others?

Exabits stands apart from other compute providers because we operate at the upstream of the AI compute industry. As a supplier to GPU marketplaces, we don’t just provide raw GPU hardware; we refine and optimize GPUs from manufacturers like Nvidia and AMD, turning them into fully managed, enterprise-grade compute resources that are highly reliable and scalable.

Our platform offers end-to-end solutions that include optimized hardware, fine-tuned firmware, and high-performance network configurations, ensuring that every GPU is utilized to its full potential. We simplify the complexity of GPU compute, allowing GPU marketplaces and end users to access top-tier compute power without the need to manage or maintain the underlying infrastructure.

Our commitment to optimizing every layer of the infrastructure—from hardware to network—ensures that our clients experience maximum performance, reliability, and cost-efficiency.

11. How can I get started with Exabits?

For developers who need quick access to GPU resources, Exabits offers an on-demand platform and API, allowing users to start utilizing our GPU services immediately after registration. This ensures instant access to high-performance compute without long setup times.

For long-term clients or those with custom requirements, we provide personalized, one-on-one service to tailor solutions to specific needs. Our team will assess your compute demands, recommend the ideal GPU configuration, and provide a seamless integration process to maximize efficiency.

Regardless of your compute needs, Exabits ensures a smooth onboarding experience, offering both immediate access and customized solutions to help you scale AI workloads efficiently.

12. How does the $EXA token drive the ecosystem?

$EXA is the ecosystem’s native token, driving utility, incentive, governance, and liquidity.

• Serves as Management fees for hosting providers to pay Exabits

• Governance Mechanism: $EXA holders vote on platform upgrades, staking rewards, and ecosystem fees, ensuring decentralized decision-making.

• Incentive & Rewards: Users stake $EXA for additional yield and receive fee discounts on GPU infrastructure.

• Tradability & Liquidity: $EXA is actively tradable on exchanges, ensuring flexibility for market participants.

13. How does Exabits tokenize compute?

We have developed a robust mechanism for transforming physical GPU infrastructure into liquid financial assets. We issue $XBIT tokens representing quantifiable compute power by tokenizing GPU compute capacity. Investors can purchase and trade these tokens on a transparent, auditable market, monetizing the underlying GPU assets through enterprise-grade compute services that generate recurring revenue.

$XBIT is not just a compute credit but a financial asset representing fractional ownership of the Exabits compute infrastructure.

• Tokenized Compute Ownership: Holding $XBIT represents a stake in Exabits’ AI compute economy, allowing investors to capture the long-term value of GPU infrastructure.

• Compute Payments & Discounts: Compute consumers use $EXA for GPU access, receiving cost benefits over fiat or other payment methods.: $XBIT can be redeemed for GPU access and model services.

• Earning Rewards: $XBIT holders stake their tokens to earn $EXA, receiving a yield based on actual compute operations.

• Value Preservation: $XBIT’s token’s price floor is maintained as compute resources scale and improve, ensuring long-term stability rather than speculative volatility.

By staking $XBIT, participants gain exposure to the growing AI compute economy while benefiting from consistent $EXA rewards.

14. Why is Exabits More Cost-Effective Than Other Compute Providers?

Exabits doesn’t race to the bottom on price, we provide the best performance-to-cost ratio in the industry. Unlike services that sacrifice quality for a low sticker price, we optimize every part of the AI compute experience, ensuring that businesses get faster speeds, higher reliability, and enterprise-grade service, all at a competitive cost.

15. Breaking the Myth: "All GPUs Are the Same"

A common misconception is that all GPU services are equal as long as they use the same hardware (e.g., an A100 GPU). In reality, GPU performance varies widely based on several factors:

• Server Configuration – Poorly optimized setups slow down processing, even with high-end GPUs.

• Software & Firmware Optimization – We enhance performance 2-7x through proprietary optimizations.

• Network & Storage Infrastructure – High-speed interconnects and advanced data management eliminate bottlenecks.

• Scalability & Reliability – Our enterprise-ready clusters ensure consistent, high-speed performance.

• 24/7 Technical Support – Mission-critical AI applications can’t afford unreliable infrastructure.

16. Why Quality Matters More Than Price

Think of AI compute like air travel:

You could book the cheapest airline ticket, but if the flight is constantly delayed, has no legroom, and loses your luggage, is it really worth it?

On the other hand, a premium airline might cost slightly more, but it ensures on-time arrival, comfortable seating, and top-tier service.

Exabits is the premium airline of AI compute offering reliability, speed, and security at a competitive price.

17. How Exabits Stays Competitive Without Subsidies

Many decentralized compute providers artificially lower costs by subsidizing GPU rentals with token incentives. This model is not economically sustainable, leading to unstable pricing and service quality. In contrast, Exabits operates on real demand-driven economics:

We generate revenue from real-world AI enterprises, not token speculation.

We reinvest in high-performance infrastructure, ensuring long-term stability.

Our efficiency and optimizations allow us to provide superior service without cutting corners.

18. Why Customers Choose Exabits

Our customers, ranging from Web2 enterprises to Web3 AI startups—aren’t just looking for the cheapest GPU time; they need:

• Reliability – AI models require 24/7 uptime and guaranteed speed.

• Scalability – Seamless expansion as compute needs grow.

• Security – Enterprise-grade protection for sensitive data.

• Support – Instant troubleshooting and system monitoring.

With Exabits, customers get the best balance of performance, stability, and cost efficiency, not just the cheapest price tag.