A new category of "NeoCloud" or "AI-native cloud" is rising, focusing specifically on AI training and inference. Unlike general-purpose clouds like AWS, these platforms are GPU-first, catering to massive AI workloads and addressing the GPU scarcity and different workload patterns found in hyperscalers.

Unlike compute-rich giants, AppLovin's bootstrapped culture enforces extreme efficiency in its AI infrastructure. Engineers don't have unlimited GPUs, forcing them to optimize code and models for cost and performance. This constraint-driven approach leads to significant cost savings and a lean operational model.

The advertised per-hour GPU cost is misleading. Because research workloads are spiky and unpredictable, labs over-provision compute. This rampant underutilization means the effective price paid is often 10 times higher than the marketed rate, creating massive deadweight loss.

AI Infrastructure (AI Infra) solves problems unique to AI/ML, such as managing compute-heavy, GPU-dependent workloads. This marks a shift from traditional infrastructure, which was often more focused on data input/output rather than intensive computation.

Unlike general-purpose cloud resources, AI training infrastructure with specialized networking (e.g., InfiniBand) and storage cannot be added fungibly. It requires significant pre-planning and deep integration, breaking the standard cloud deployment model of simply adding more commoditized compute or storage as needed.

In AI infrastructure, the capital cost of GPUs (~80%) dwarfs operational costs. Therefore, getting a multi-billion dollar cluster online a few months earlier generates far more value than optimizing for TCO, justifying seemingly wasteful spending on stopgaps like mobile chillers to bypass construction delays.

While AI inference can be decentralized, training the most powerful models demands extreme centralization of compute. The necessity for high-bandwidth, low-latency communication between GPUs means the best models are trained by concentrating hardware in the smallest possible physical space, a direct contradiction to decentralized ideals.

Beyond catastrophic hardware failures, a significant problem in large-scale AI training is the "soft failure" of a single GPU slowing down. In a job using thousands of GPUs, identifying this one straggler is incredibly difficult without specialized observability tools built specifically for AI workloads.

The report of XAI's low GPU utilization reveals a critical, non-obvious bottleneck in AI: it's not just about acquiring compute, but using it efficiently. This 'FLOPS utilization' problem, caused by architectural and load-balancing issues, means billions in hardware sits underused, creating an opportunity for companies that can optimize the compute stack.