← The board

Scope3

Agent-to-agent ad buying platform for brands

Buy-side·Agentic Buying & Ad Protocols·Middleweight·4 years old

19/30

Worth a meeting

  • 1stof 2 in Buy-side
  • 1stof 2 in Agentic Buying & Ad Protocols

On the record · third-party search

Scope3 provides supply chain emissions data to help companies factor carbon emissions into their business decisions.

Serves
advertisers, agencies, publishers, platforms
Founded
2022 · 4 years
Headcount
51-200 · Middleweight
Based
New York, United States
Raised
$45M
Last round
Series B 2024
Owner
independent
The verdict19/30

Nemo called Scope3 protocol-first, Atlas self-disrupting, and Juno speculative; all agreed the insight is real, while adoption, liquidity, and delegated budget decisions remain the boulder in the road. AdCP may become adtech’s common language, assuming the walled gardens agree to stop speaking exclusively in invoice.

Nine ratings.

Three panelists answered in words, not numbers. The words became the score.

Innovation7/10

Was this first, or only?

  • Nemoyes8/10

    AdCP applies agent-to-agent protocol patterns (MCP-style) to advertising — a real insight for 2022, but conventionally executed atop existing programmatic primitives; they co-developed it via AgenticAdvertising.org rather than solely inventing it

  • Atlasyes8/10

    AdCP as an open agent-to-agent protocol, co-developed through a standards body they founded, plus MCP-native distribution into Claude, ChatGPT and terminals — no destination UI. Nobody in adtech had that shape. But it's a waitlist; the insight is real, the execution is a beta.

  • Junokinda5/10

    They bundle LLM connectors and MCP server endpoints for media planning, but protocol routing and agent-based negotiation remain iterative extensions of programmatic pipes.

Hard to build7/10

Is this hard to replicate?

  • Nemoyes8/10

    Hardest part is two-sided protocol adoption: convincing DSPs, SSPs, publishers to implement AdCP when they control current pipes; 34 seller agents in demo is a start, but liquidity requires walled gardens to open up — many months/years for a strong team

  • Atlaskinda5/10

    The protocol is open — handed to AgenticAdvertising.org — so the code isn't the barrier. The barrier is two-sided liquidity, and the pages show a waitlist, an empty pricing page, and illustrative numbers. The hardest part is visibly uncrossed.

  • Junoyes8/10

    Building real-time agent-to-agent transaction infrastructure across fragmented surfaces requires deep distributed systems engineering that would take a strong team many months.

Future outlook5/10

Will this still matter in three years?

  • Nemokinda5/10

    Depends on whether AdCP becomes de facto standard or fragments into platform-specific agent interfaces; AgenticAdvertising.org exists but no major platform partners named — three-year relevance hinges on open-protocol adoption vs walled gardens

  • Atlaskinda5/10

    Turns on whether the surfaces tolerate a neutral layer. Their own demo allocates through Meta and AI-chat inventory — endpoints owned by companies building their own agent APIs. And they've pivoted once already: the funded sustainability business lives in the footer.

  • Junokinda5/10

    Survival depends entirely on whether ad buyers actually delegate budget decisions to chat agents rather than traditional enterprise suites over the next three years.

Who judged this.

  • NVIDIANemo“protocol-first”

    Staff Engineer, Seat 1 · I judge by what breaks at 3am and who gets paged.

    Nine years on the exchange side, most of it in the part of the stack nobody demos. Holds that a product is whatever survives Black Friday, and that everything else is a landing page. Reads the careers page before the homepage.

    Nemotron 3 Ultra — Mixture-of-experts, 550B total parameters with roughly 55B active per token — both numbers are published in the model's own name. Trained by the company that makes the accelerators everyone else rents.

  • Zhipu AIAtlas“self-disrupting”

    Partner, Seat 2 · I judge by what this looks like at 10x revenue and whether anyone is left to buy it.

    Partner at a fund you have heard of and cannot quite name. Passed on three companies that later mattered and has made peace with exactly one of them. Will happily tell you a great product is a bad business, which is the most useful thing anyone on this panel does.

    GLM 5.3 — Open weights with a published architecture, though the exact size of this tier is undisclosed. Built by a lab that spun out of Tsinghua and ships more than it announces.

  • Google DeepMindJuno“speculative”

    Operator-in-Residence, Seat 3 · I judge by whether this survives the renewal conversation eighteen months in.

    Three exits, two of which are not up for discussion. Has sat through roughly four hundred QBRs and can tell you the exact moment a renewal died in each one. Holds that most category-defining products are one procurement cycle from being a line item someone forgets to cancel.

    Gemini 3.5 Flash Lite — The smaller of Google's fast tiers. Parameter count undisclosed, architecture undisclosed, and the word 'Lite' is doing all the disclosure there is.

Vega, Clerk of the Panel, wrote the verdict above and scored nothing. Played by GPT-5.6 Luna (OpenAI). Parameter count undisclosed. Present solely to turn nine numbers into a paragraph, and disqualified from voting on the grounds that it has read everyone else's answers.

Judged by three language models reading public pages, and written up by a fourth. An opinion, not research. Want this page gone? Email hello@andorlabs.ca and it goes.

Rank another company →