Assessment Disclosure

An Independent Research System

Layer2C is an independent architectural research system. Assessments are opinionated analyses, not certifications, paid rankings, procurement recommendations, or permanent conclusions.

The system has four public properties. Layer2C assesses authority across the 4+1 Layer AI Infrastructure Model. Fourth Cloud assesses operating-model readiness, FC-0 through FC-4. Labs validates where authority actually holds, through scoped real-hardware builds. StackBuilder composes reference architectures from the assessed evidence. The System Map lays out how they fit.

What These Assessments Are Not

• Not rankings. Vendors are not scored against each other or ranked in order. There is no Layer2C top-ten list.
• Not certifications. No vendor pays to receive a certification, badge, or seal of approval from Layer2C or The CTO Advisor.
• Not procurement recommendations. These assessments do not tell you which vendor to buy. They map the architectural and authority implications of each vendor's platform so buyers can ask better questions.
• Not lab validations by default. Most cells rest on product documentation. Where a published lab result exists, from Layer2C Labs or a third party, the cell cites it, and the lab speaks only to the surface it tested.
• Not exhaustive product reviews. Each assessment evaluates authority placement across the 4+1 AI Infrastructure Model. It does not cover every product feature, SKU, or configuration option.
• Not permanent. AI infrastructure evolves rapidly. Assessments reflect a point in time and are updated as architectures change. Each assessment displays its date and version.

Assessment Type

Layer2C assessments are opinionated architectural analyses. They evaluate where decision authority resides across the 4+1 Layer AI Infrastructure Model — from compute and networking through data, orchestration, runtime, reasoning, and applications — and classify each vendor component using the Decision Authority Placement Model (DAPM).

The assessments are produced by The CTO Advisor LLC, an independent analyst firm. The primary author is Keith Townsend. The CTO Advisor's independent architectural judgment is one of the evidence inputs; it is not the only one.

Evidence Sources

Assessments draw on:

• Official product documentation: datasheets, admin and user guides, release notes, and API references
• Vendor posts and press releases
• GitHub repositories and open-source project documentation
• Published lab results, from Layer2C Labs or a third party

Vendor briefings, conference sessions, earnings calls, analyst coverage, and trade press can point to evidence. They never move a grade on their own. Keith Townsend rules every contested call, and the rulings are recorded in each assessment's audit notes. Each assessment lists its primary sources. Where a vendor has reviewed an assessment prior to publication, this is noted on the assessment page.

How Assessments Are Produced

Assessments are drafted with AI assistance and decided by a human. Claude Code (Anthropic) researches sources and drafts each cell, and Keith Townsend grades every cell with it one at a time. Each locked cell is then attacked by ChatGPT (OpenAI, through its codex command-line tool), and each assembled row by Antigravity (Google, through its agy command-line tool). Every claim either reviewer raises is checked against the source it cites and ruled by Keith before it changes a cell.

Anthropic, OpenAI, and Google are also assessed on this site, including at Layer 3, where Claude Code, ChatGPT, and Antigravity are graded as applications. Using a tool doesn't change how its vendor is graded: the same rules apply to every row, and the overlap is disclosed here so readers can weigh it.

Update Cadence

Assessments are updated when material architectural changes occur — new product launches, significant partner changes, acquisition of capabilities relevant to a layer, or substantial platform evolution. Every content change, a corrected fact included, increments the version and adds an audit note to the assessment's sources.

The date and version displayed on each assessment reflect the most recent substantive update. Assessments are not updated on a fixed schedule — they are updated when the architecture changes.

Vendor Correction Process

Vendors who identify factual errors in their assessment may submit corrections to The CTO Advisor at thectoadvisor.com. Factual corrections — incorrect product names, inaccurate capability descriptions, outdated information — are reviewed and incorporated promptly.

Vendor review does not imply vendor approval or endorsement. Editorial positions — architectural judgments, DAPM classifications, gap analysis conclusions — are retained at The CTO Advisor's discretion. Vendors do not have final approval over assessment content.

Commercial Relationship Policy

Vendors do not pay to be assessed on Layer2C. Assessment inclusion and findings are not influenced by commercial relationships.

The CTO Advisor has a commercial relationship with a vendor when that vendor paid The CTO Advisor LLC, or signed to pay it, for any work in the 12 months before an assessment's publication date. Work means advisory or analyst retainers, consulting, sponsored content, Layer2C Labs and design challenges, and Buyer Room sessions. The test is an invoice or a signed agreement.

These don't count: an unsigned proposal, a briefing, an event invitation or analyst program access, The CTO Advisor buying a vendor's products as a customer, and a vendor using Layer2C's vocabulary. A relationship that ended more than 12 months before publication isn't disclosed.

Where a commercial relationship exists, it's disclosed on that vendor's assessment page. When a parent company and its subsidiary each have their own assessment, the disclosure sits on the page of the one that paid. The absence of a disclosure note means no such relationship existed in the 12 months before the assessment's publication date.

What DAPM Does and Does Not Mean

The Decision Authority Placement Model (DAPM) classifies each assessed component on one axis: can the enterprise take its accumulated opinions at this layer — configs, policies, governance logic, operational model — and operate them independently of this vendor without rebuilding?

Retained means yes. The capability rests on a commodity substrate (x86 compute substitutable across OEMs) or open-source software with real alternatives. Also applies when the vendor provides nothing at a layer — the enterprise owns the function by default.

Delegated means partially. The capability is provided by a substitutable partner or open-source tooling the enterprise could swap without rebuilding. The enterprise keeps the right to change providers.

Ceded means no. A closed system is a closed system. The vendor's opinions are proprietary with no open exit, regardless of who operates the software or where it runs.

DAPM is not a quality score. A Ceded component can be excellent — cloud vendors routinely provide world-class capability that enterprises consume without governance authority. The classification describes the governance relationship, not the capability's fitness for purpose. Completeness and authority are orthogonal. A Retained classification does not mean the capability is good. A Ceded classification does not mean the vendor should be avoided.