AD FIELD NOTESAI advertising · the work behind the output

Case files / generative production

CoreWeave Unlocking AI’s Full Potential: build the system before the spectacle

AKQA used AI-built image libraries, model-informed brand guidelines and Veo 2 video for CoreWeave’s IPO-era website, OOH and national TV campaign.

Campaign period: 2025Source published: Date not statedArchive entry prepared: 15 September 2026

Retrospective coverage, prepared in September 2026 from the original record.

A blue CoreWeave billboard in Times Square reads ‘Powering All Innovations.’
CoreWeave’s Times Square out-of-home campaign, as documented by AKQA. CoreWeave / AKQA; image published by AKQA.

Reported execution: an AI company used AI as production infrastructure

AKQA describes a 360 campaign around CoreWeave’s $23 billion Nasdaq IPO. It says brand guidelines were reworked to inform models, a bespoke image library was made entirely with AI, and the work extended across a redesigned website, Times Square OOH and a national television campaign using AI-generated video and Veo 2.

The result is notable less for a single generated frame than for the attempt to turn a brand system into a repeatable production system. The visual language needed to survive a web landing page, TV, street-scale digital screens and a company-defining financial moment.

Our reading: the model brief is part of the identity brief

When a brand expects models to make assets, its guidelines cannot stop at a logo and palette. They need instructions a human and a generation workflow can both use: composition, subject matter, exclusions, tone and the evidence a visual is allowed to imply.

That can produce scale, but it does not eliminate art direction. The Times Square execution succeeds because the message is legible at a glance; a vast image library is useful only if selection remains disciplined.

Production lesson: distinguish the source library from the final claim

Keep a register of model outputs, prompt families, approved variants and the media placement each asset was built for. That makes corrections possible when an image overstates a product capability, resembles a protected work or fails a channel requirement.

For a technology brand, generated video needs extra review around demonstrations. Avoid letting an evocative synthetic scene read as literal product footage. The best governance is an ordinary production habit: label concepts, verify claims and give legal and subject experts a meaningful review point.

Evidence boundary: scale claims do not validate every asset

AKQA reports impressions, video views and clicks, and describes the campaign as part of CoreWeave’s IPO moment. Those are agency-reported outcomes, not an independent comparison of AI production against conventional production or evidence that every generated asset performed equally.

The enduring case is structural: a campaign used AI both as subject matter and as an organized making system. Its value depends on the human brand rules that narrowed a potentially limitless output space into recognizable public work.

Sources & limits

AKQA case study; reported reach figures are agency claims and its page gives year-level campaign context only.

  1. AKQA — CoreWeave: Unlocking AI’s full potential — source publication date not stated.

Send a correction with the passage and supporting source.

The weekly field note

One useful note for the next brief.

Campaigns worth studying, the work behind them, and questions to bring into your next creative review. Delivered weekly.

How we handle your email