The 2026 enterprise AI infrastructure vendor SLA conversation resolves on five dimensions (uptime commitment with the denominator named explicitly, latency commitment at P95 and P99 negotiated into the MSA addendum because public SLAs typically omit it, support response tier per severity level, credit calculation scope and cap, exclusions list scope including scheduled-maintenance window, content-policy actions, capacity constraints, partial-availability events, and third-party-source outages); the publicly disclosed headline numbers (AWS Bedrock 99.9% monthly with 10/25/100% credit tiers, Azure OpenAI Service inheriting Azure platform 99.9% with PTU separate availability, Google Vertex AI 99.5%-99.9% varying per-model and per-region, OpenAI Enterprise 99.9%-99.99% per-customer in MSA, Anthropic Enterprise commitments per-customer with no public uniform tier) understate the year-two operational reality because exclusions list scope and credit calculation scope vary materially across vendors; the buying-committee discipline is to populate the per-vendor matrix at short-list rather than discover the gaps at year-one operational experience or year-two renewal.
Anchored on (a) AWS Bedrock SLA at aws.amazon.com/bedrock/sla/ (99.9% monthly uptime, 10/25/100% credit tiers, regional scope); (b) Azure OpenAI Service SLA inherited from Azure platform at 99.9% with standard Azure credit tiers and PTU-specific commitments per learn.microsoft.com/en-us/azure/ai-services/openai/concepts/service-levels; (c) Google Vertex AI SLA at cloud.google.com/vertex-ai/sla with varying per-model per-region commitments; (d) OpenAI Status page at status.openai.com for historical record and OpenAI Enterprise documented MSA-addendum SLA targets (99.9%-99.99%); (e) Anthropic Status page at status.anthropic.com plus Enterprise-tier per-customer SLA observations from 2025-2026 procurement-team interactions. The exclusion-scope characterisation is from current public SLA documentation; specific exclusions (content-policy actions for Azure OpenAI, model-update windows for Vertex AI, scheduled-maintenance windows across the hyperscalers) are the rows most likely to surprise at renewal. 60-day review cadence (26 Jul 2026). Trigger conditions: (1) any major vendor publishing a new SLA tier or restructuring the credit mechanism materially shifts the matrix; (2) a published industry-wide outage event (2026 analog to Dec 2024 Bedrock content-filter outage or Jan 2025 Azure OpenAI degradation) provides precedent that changes exclusions-list framing; (3) NIST AI RMF or sector-specific cybersecurity rules requiring specific SLA commitments for regulated workloads warrants matrix update; (4) Anthropic and OpenAI Enterprise SLA documentation reaching same transparency level as hyperscaler tier changes the model-vendor comparison materially. Sibling AM-174 covers cost-side calculation; AM-167 covers contract-side instruments; /agentic-ai-sla-architecture/ covers the customer-side SLA architecture this piece is the supply-side companion to (the architecture piece has 48 Copilot citations and is the inbound referrer for this comparison).
/holding/AM-179/Embed this claimiframe + oEmbed
The card auto-updates when the claim's status, last-reviewed date, or correction log changes. Embedders never need to refresh — the card is rendered live from the canonical record.
Email-me when AM-179's status, next review date, or correction log changes. One email per change. No newsletter subscription, no other mail.
The claim: The 2026 enterprise AI infrastructure vendor SLA conversation resolves on five dimensions (uptime commitment with the denominator named explicitly, latency commitment at P95 and P99 negotiated into the MSA addendum because public SLAs typically omit it, support response tier per severity level, credit calculation scope and cap, exclusions list scope including scheduled-maintenance window, content-policy actions, capacity constraints, partial-availability events, and third-party-source outages); the publicly disclosed headline numbers (AWS Bedrock 99.9% monthly with 10/25/100% credit tiers, Azure OpenAI Service inheriting Azure platform 99.9% with PTU separate availability, Google Vertex AI 99.5%-99.9% varying per-model and per-region, OpenAI Enterprise 99.9%-99.99% per-customer in MSA, Anthropic Enterprise commitments per-customer with no public uniform tier) understate the year-two operational reality because exclusions list scope and credit calculation scope vary materially across vendors; the buying-committee discipline is to populate the per-vendor matrix at short-list rather than discover the gaps at year-one operational experience or year-two renewal.
About this register
The Reporting register tracks claims published from articles addressed to senior enterprise IT leaders — CIOs, IT directors, heads of platform. Claims are reviewed on a 30–90 day cadence; each review either reaffirms the claim, marks one substantive part as Partial, or marks it Not holding once the underlying evidence has been overtaken.
Recent corrections in Reporting
- AM-008 · Partial · 17 Jun 2026
Source-text figure re-review: Google's 2024 Environmental Report reports a 28% year-over-year increase to 8.1 billion gallons, not the 33% (from a 6.1 billion 2023 base) asserted at publish. The 8.1B 2024 figure and the Microsoft WUE 0.30 L/kWh / 39%-improvement figure are unchanged and verified. Article corrected to 28% and the unsupported 6.1B base removed; the claim text retains the original figure with this correction per the Holding-up protocol.
- AM-132 · Partial · 10 Jun 2026
One of four legs unanchored on re-review. The claim text attributes '12% of deployments clearing 300%+ ROI with 88% at or below break-even at 12-18 months' to the Stanford DEL 2026 Enterprise AI Playbook. Full-text verification on 10 Jun 2026 found no such figure in that source: the playbook (Pereira, Graylin, Brynjolfsson, Apr 2026) studies 51 successful deployments by design and contains no ROI distribution, no 300%-plus cohort, and no break-even measurement point (full finding at AM-029, correction of 10 Jun 2026). The only verified figure carrying the same 12/88 numerals is IDC research with Lenovo (via CIO.com, Mar 2025): roughly 88% of AI proof-of-concepts never reach production and roughly 12% graduate — a pilot-to-production graduation metric, not an ROI distribution. The Gartner 28%, McKinsey 23%/17%, and MIT NANDA 95% legs verify; they support a small high-performing tail and a large struggling body, but none documents the two-peak bimodal shape the claim asserts. Status Up -> Partial.
- AM-129 · Partial · 10 Jun 2026
One of three read-against anchors unanchored on re-review. The claim text cites 'Stanford Digital Economy Lab Enterprise AI Playbook (12/88 bimodal ROI distribution at 12-18 months)' and frames the realistic ROI band around 'the highest-discipline 12% cohort'. Full-text verification on 10 Jun 2026 found the playbook contains no 12/88 distribution, no bimodal ROI shape, and no 12-18-month ROI measurement point (full finding at AM-029, correction of 10 Jun 2026). The claim's core negative finding — no mid-market enterprise has produced a documented +240% ROI in 90 days under audited conditions — is unaffected; the McKinsey State of AI 2025 and MIT NANDA legs verify and continue to support it. The '12% cohort' framing has no verifiable referent. The only verified figure carrying the 12/88 numerals is IDC's pilot-graduation finding (roughly 88% of AI proof-of-concepts never reach production; via CIO.com, Mar 2025), a different metric. Status Up -> Partial.
Reviews coming up in Reporting
- AM-063 · Holding · next +9d (27 Jun 2026)
AI agents executing financial transactions need a four-control bundle (action-approval gates by blast radius, kill-swit…
- AM-061 · Holding · next +9d (27 Jun 2026)
Production agentic-AI costs at scale routinely run multiples of POC projections, and a layered optimisation programme c…
- AM-003 · Partial · next +9d (27 Jun 2026)
GPT-5 Pro's tiered-subscription model forces enterprises to classify problems by computational difficulty — $200/month…