Compute Availability Factor (CAF): A Framework for Measuring the Physical Deliverability of Artificial Intelligence Compute: From nominal accelerator capacity to effective and infrastructure-available compute

Q: What is the Compute Availability Factor (CAF)?

The Compute Availability Factor (CAF) is an empirical measurement framework developed by Richie Adomako at Serpens Inc. (ResearchGate Publication ID 414150400) that quantifies the fraction of hardware-effective AI compute that can actually be delivered in the physical world under real electrical (E), thermal (T), network (N), and operational (O) infrastructure constraints: CAF(t) = E(t) × T(t) × N(t) × O(t).

Q: What is the relationship between the CAF website, CAF Watcher, and the CAF Research Paper?

CAF is the formal mathematical methodology and research specification ('Compute Availability Factor: A Framework for Measuring the Physical Deliverability of Artificial Intelligence Compute', Adomako, 2026) rooted in the thesis 'When the Building Becomes the Computer'. This site hosts both the formal specification and CAF Watcher—the operational observability platform, 3D Spatial Desktop OS, Cluster Physics Simulator, and Workload Constraint Investigator built on that methodology.

Q: How does CAF calculate Physically Deliverable Compute and the Deliverability Coverage Ratio (DCR)?

CAF first converts Nominal Compute (C_N, raw installed accelerators) into Hardware-Effective Compute (C_E = C_N × U_H). It then multiplies by CAF = E × T × N × O to derive Physically Deliverable Compute (C_D = C_N × U_H × CAF). The Deliverability Coverage Ratio (DCR = C_D / C_R) compares verified deliverable compute against roadmap required compute (C_R) using continuous interval bounds [Lower, Upper].

Q: How does CAF handle missing or unverified public infrastructure data?

CAF enforces the epistemic axiom that 'Unknown does not equal Unavailable'. Missing public evidence is never treated as zero capacity; instead, it is classified as UNKNOWN, widens the [Lower, Upper] availability interval, and reduces the separate Evidence Confidence score (Q).

Adomako, Richie

↳ CAF WATCHER // CONTINUOUS PHYSICAL AUDIT LEDGER & RESEARCH SPECIFICATION

Compute Availability Factor (CAF)

What compute can actually be delivered, what constrains it, and what attributable evidence supports that assessment? Built on the formal paper by Richie Adomako (Serpens Inc., ResearchGate ID 414150400).

Empirical Substrate Verification (CAF = E × T × N × O)

CAF decomposes computing deliverability into four physical factors: Electrical Substation (E), Thermal Heat Dissipation (T), Cluster Optical Network (N), and Operational Reliability (O). Missing public evidence is treated as unknown—never assumed zero.

Frontier AI Substrate Registry (62 Audited Entities)

OpenAYA — Grade UNSCORED · Browser-delivered AI workspace and orchestration platform (Work OS, memory, evaluation harness, and external model gateway) — CAF’s transparent first-party internal test case
OpenAI — Grade C+ · Frontier reasoning pioneer confronting GPT-6 Astra recurrent-depth thermal loads, $500B Stargate consortia, and nuclear-scale utility dependencies
Anthropic — Grade B · Dual-architecture frontier pioneer (Mythos / Fable) deploying state-of-the-art agentic reasoning across AWS, Google, and Snowflake substrates
Google DeepMind — Grade A · End-to-end full-stack titan: silicon, datacenter buildings, undersea fiber, and research
Meta AI / FAIR — Grade A- · Open-weights superpower building massive 600k GPU clusters and custom MTIA silicon
xAI (Colossus) — Grade B- · Hyper-accelerated physical builder deploying 100k+ GPUs in record 122 days
Microsoft AI / Azure — Grade A · The world’s AI factory floor: massive global hyperscale infrastructure hosting OpenAI
NVIDIA (Hardware Substrate) — Grade A · Monopolist of the physical substrate: building the AI factory racks and NVLink copper spine
DeepSeek (Sovereign Frontier) — Grade A- · Hangzhou algorithmic titan delivering frontier-grade MoE & reasoning despite extreme silicon sanctions
Alibaba Cloud (Qwen Ecosystem) — Grade B+ · Asia’s hyperscale infrastructure backbone powering the ubiquitous Qwen open-model family
Amazon AWS (Hyperscale Substrate) — Grade A- · The global cloud colossus: backing Anthropic with Project Rainier, nuclear PPAs, and custom Trainium2 silicon
Huawei (Ascend AI Ecosystem) — Grade B+ · Sovereign Chinese semiconductor titan building full-stack hardware, CANN compiler, and mega-clusters

View all 62 monitored entities →