WillItSurvive
YC Fall 2025 · United States

Absurd

AI-agent studio that turns frontier video models into branded ad campaigns.

Protection under test
Assessed 27 July 2026 · Evidence as of 27 July 2026
Something worth protecting exists and is being tested right now.
On the recordSince 27 July 2026 · Checkable from Q1 2027 · Every call on the record Checkpoint: Next funding announcement and disclosed ARR or client-renewal rates
The call

Absurd is a six-person YC team wrapping frontier video models into branded ads for other startups, with no disclosed funding, revenue, or lock-in so far.

Absurd takes video models from companies like OpenAI and Google and layers software on top to help customers make ads faster. The problem is that Google, Anthropic, and other AI labs are now showing they can do ad-making themselves, which means Absurd's main job (connecting the pieces) could disappear if those labs decide to sell directly. The founding team has strong engineering talent from NVIDIA and Meta, but the company has not yet proven it can keep customers coming back or lock them in with data and workflows that would be hard to replace. What matters most is whether Absurd can shift from one-time video projects to repeat business where customers depend on their system, and whether the big AI labs bundle orchestration features into their base products.

Why
01
AI pressure: Frontier video-model labs (Google Veo/Gemini, Anthropic, Sora, Runway, Luma). Google and Anthropic are building in-house ad generation that could fold Absurd's whole orchestration layer into the model itself. Could AI do the core job? Yes, likely.
02
Bigger players: Google and Anthropic in-house ad production. If either lab ships a self-serve ad tool bundled with its model, Absurd's thin layer on top has nothing left to add. Could a giant copy it? Yes, easily.
03
What protects it. None apparent yet: no proprietary data, workflow lock-in, or distribution edge beyond YC-network warm intros.
What would change this call
Google or Anthropic ships a self serve 'Creative OS' ad tool bundled with its video model. Shows the frontier labs are closing the orchestration gap Absurd depends on.
Creatify or Arcads announces 5 to 10 new major brand deals with renewal data. Shows funded rivals are locking in the brand relationships Absurd needs.
Absurd discloses a Q1 2027 funding round with ARR or renewal rates for its 7 named clients. Shows whether Absurd is building recurring, outcome priced relationships or staying a one-off engagement shop.
Key numbers
  • 6-person team (2026-07-25)
  • 7 named clients incl. Hims, Kalshi, Brex (2026-05-17)
  • Kalshi ad topped 1M views (self-reported) (2025-10-14)
Named risks
  • No proprietary data, integration lock-in, or distribution moat identified beyond service quality. Severity: high
  • Frontier labs Google and Anthropic already produce AI ads in-house at near-zero marginal cost. Severity: high
  • All traction metrics (views, client count) are self-reported and inconsistently cited across vendor sources. Severity: medium
  • Small 6-person team with no disclosed funding round beyond standard YC investment as of July 2026. Severity: medium
  • Crowded, well-funded competitive category (Arcads, Creatify, AdCreative.ai, Icon, Koyal). Severity: medium
Rivals, as read in July 2026
RivalDirectionEvidence
Google (Gemini in-house ad production)losing toGoogle shipped an AI-assisted ad campaign using Gemini in July 2026
Anthropic (in-house ad production)losing toAnthropic ran its own AI-generated ad campaign in July 2026
Arcadscontesting300+ licensed AI actors, pricing from ~$110/mo, direct rival as of May 2026
Creatifygaining onWell-funded AI ad generator competing for same brand budgets as of mid-2026
KoyalcontestingYC-backed agentic AI filmmaking platform with paid label pilots as of July 2026
IconcontestingEnd-to-end AI ad platform combining actors, scripting, editing as of July 2026
Questions people ask
Is Absurd profitable in 2026?
No revenue figures are publicly disclosed for Absurd as of July 2026. The YC F25 company appears pre-revenue or revenue-undisclosed, with all traction limited to self-reported client names and view counts rather than financial figures.
How many people work at Absurd?
Y Combinator's directory listed Absurd at 6 employees as of July 25, 2026, roughly a year after its 2025 founding, suggesting a small founder-sized team rather than a scaled operation.
Who are Absurd's biggest competitors?
Absurd competes with funded AI ad generators Arcads, Creatify, AdCreative.ai, and Icon, plus YC peer Koyal, while frontier labs Google and Anthropic shipped their own in-house AI ad campaigns in July 2026, narrowing the market.
Could AI replace what Absurd does?
Yes, this is the core structural risk: Absurd is a thin orchestration layer atop frontier video models with no disclosed proprietary data or API lock-in, and Google and Anthropic both produced in-house AI ad campaigns in July 2026.
Who founded Absurd and what is their background?
Absurd was founded in 2025 by Philip Ho (CEO), Daniel Kathein (CTO, previously at NVIDIA), and Damian Chng (previously at Meta), per a February 2026 profile, as part of Y Combinator's Fall 2025 batch.
Has Absurd raised venture funding?
As of July 8, 2026, funding tracker Seedtable listed Absurd's total funding as Unknown, with no disclosed seed round, investor list, or valuation beyond the standard YC batch investment.
Timeline
  • 2026-07-25YC directory lists Absurd at 6 employees, HQ Los Angeles
  • 2026-07-14Anthropic runs its own AI-generated ad campaign
  • 2026-07-04Google ships AI-assisted ad campaign using Gemini
  • 2026-05-17Founder bio lists expanded client roster: Whop, Reforge, Micro1 added
  • 2026-04-15Adjacent competitor Hightouch reported reaching $100M ARR
  • 2026-03-02Job listing shows prior Founding Engineer role advertised
  • 2026-02-08Secondary blog covers Absurd's launch pitch and workflow
  • 2025-10-14Absurd publicly launches via YC Launch as 'AI ads powered by agents and creatives'
  • 2025Absurd founded by Philip Ho, Daniel Kathein, Damian Chng; joins YC Fall 2025 batch
Revision history
27 July 2026
First assessed. Protection under test. The text above is fixed as of this date and is not edited afterwards.
In the same band
Sources