The problem we are solving

Three trades, one missing feedback loop.

3D concrete printing has been treated as three separate trades that hand off to each other through paper specs and human translation: a mix is formulated in one lab, printed on someone else’s machine with parameters that are not fully documented, and tested as cured cubes that don’t reflect print anisotropy. The handoffs lose information. The field-wide signal disappears in the gaps.

The RILEM TC 304-ADC inter-laboratory study has now quantified the cost: across 30 labs — each printing its own mix design on its own printer — printed specimen strengths spanned roughly 3.3 to 130 MPa (per-lab 28-day means ranged 27–96 MPa), a spread driven by materials and printing process together. Process variables are not noise; they are most of the signal.

What we build

One company across all three stages.

Sunnyday operates all three stages of the closed loop as one system, so the data that moves between them is structured rather than reformatted at every handoff.

• CEMFORGE — AI mix formulation, available at cemforge.ai.
• M3-CRETE — open-source printing hardware, CERN-OHL-W-2.0, at m3-crete.com.
• Open3DCP — open printability dataset and schema at open3dcp.org, anchoring process-controlled training data across the field.
• CADCLAW — the open-source CAD validation tooling that produces M3-CRETE’s assemblies, at cadclaw.io.

The logic: process-controlled training data compounds. Every cycle of predict-print-prove tightens the model for the operator running it, and contributes structured data back to the field-wide schema. The loop is the product.

Founder

Nick Sonnentag

Sunnyday Technologies was founded in 2023 by Nick Sonnentag — a materials and process engineer with more than 30 years of engineering experience spanning formulation chemistry, metallurgy, and additive manufacturing across the chemical, aerospace and defense forging, and specialty-vehicle industries.

Read more on the founder page.