Poetiq, which leverages existing LLMs to create “expert agents” for specific tasks, and spent just $40K to achieve high ARC-AGI-2 scores, raised a $45.8M seed
PuckIan Krietzberg
Context & Ripple Effects
Poetiq’s seed round follows coverage of its claim that a small spend produced strong ARC-AGI-2 results, positioning the company around specialized agents built on existing models rather than new foundation models. The funding turns that technical-efficiency claim into a better-capitalized effort to build and deploy task-specific products.
Poetiq gains $45.8M to develop its expert-agent approach beyond the low-cost benchmark result and pursue deployment for specific tasks.
The round raises the burden on Poetiq to translate its reported ARC-AGI-2 efficiency into reliable, task-level agent performance.
Second-order effects
Other agent startups built atop existing LLMs face sharper pressure to demonstrate both specialized outcomes and capital efficiency, rather than relying on access to proprietary base models.
Enterprise buyers evaluating embedded agents gain another well-funded vendor category, making evidence of workflow fit and operational reliability more important in procurement.
Third-order effects
If task-specific agent companies continue attracting large rounds, more value may accrue to workflow expertise, evaluation, and integration around general-purpose models rather than to model training alone.
The pattern could also make agent markets more competitive and fragmented, with durable leaders likely determined by deployment and customer adoption rather than benchmark claims alone.
The trend:Enterprise AI is moving toward specialized agents that package general-purpose LLMs into deployable, measurable systems for particular workflows.
Super psyched to welcome Poetiq to YC family— we're over the moon to welcome Shumeet Baluja and welcome back YC W11 alum Ian Fischer. Getting to the top of ARC-AGI is no small feat, and recursive improvement a powerful milestone.
.@poetiq_ai is a new startup founded by former DeepMind researchers (@itfische & @sbpoetiq) that recently achieved a major jump on the ARC-AGI benchmark by layering a recursive self-improvement system on top of Gemini 3. In this conversation at NeurIPS, @FrancoisChauba1 sat down …
Congratulations @poetiq_ai on the monster $45M seed round! Have been so impressed with this team, and am thrilled to get to work with them them in an upcoming batch @ycombinator. [image]
Thanks @ycombinator and @FrancoisChauba1 for inviting me to talk with you about what we've been building at @poetiq_ai! I really enjoyed the conversation!
At @OpCo_VC, we're very excited and honored to partner with @poetiq_ai. They have raised a $45.8M seed round after beating industry-leading benchmarks with a small team of 6. Much more to come!
We're thrilled to announce a new chapter for Poetiq: We have closed $45.8M in Seed funding. It's a privilege to build alongside partners who understand the scale of our vision, including Surface, FYRFLY, @ycombinator, 468, Operator Collective, NeuronVC, and HICO.