Building energy optimization, built by someone who managed the buildings it was supposed to fix.
Celaxis started in Portland in 2025 with a specific observation: commercial buildings were paying hundreds of thousands of dollars per year in peak demand charges that a better-timed HVAC schedule could eliminate — and the BMS tools available required an engineering consultant to configure and months to deploy.
Ingrid Larsson
Before founding Celaxis, Ingrid spent seven years in energy procurement and facilities systems management for commercial property portfolios across Oregon and Washington. The inflection point came in 2024: she pulled a year of PGE interval data for a 90,000 sq ft office building and found that roughly $340,000 of the past year's demand charges had been driven by six predictable Monday-morning warm-up spikes — all caused by a BMS schedule that hadn't been adjusted since the building opened in 2017.
The tools that existed required an HVAC controls engineer to configure, or were embedded inside larger energy management platforms that took 12–18 months and significant capital to deploy. Neither fit the reality of a facilities team managing 8 buildings with one energy manager.
Celaxis is the tool she needed. It connects to the BMS you already have, builds a thermal model of your specific building from observed data, and starts issuing pre-conditioning commands in under 30 days. No hardware, no HVAC engineer required, no 18-month deployment. A facilities manager can evaluate, approve, and operate it directly.
Contact Ingrid directly
Small team. Deep domain.
Three people building something we want to use. No sales team, no account management layer — direct line from customer feedback to product iteration.
Seven years in energy procurement and facilities systems for Pacific Northwest commercial property portfolios. Founded Celaxis in 2025 after tracing six months of demand charge spikes to a BMS schedule that hadn't been updated since 2017.
Former distributed systems engineer with background in industrial IoT and time-series data pipelines. Designed Celaxis's BACnet adapter layer and thermal model inference engine.
Building thermal modeling and occupancy inference. Designed Celaxis's per-zone regression model — trained from each building's own BMS observation data, not industry averages. Every building gets its own thermal fingerprint.
Portland is an ideal proving ground for this kind of software.
Portland General Electric and Pacific Power both operate time-of-use and peak demand tariffs that create real dollar impact when HVAC systems are poorly optimized. Oregon's commercial building stock — heavy on converted mid-century office and mixed-use retail — is also exactly the kind of building that benefits most from thermal inertia pre-conditioning.
Oregon's Clean Energy Plan creates regulatory incentives for measurable efficiency gains that commercial tenants and building owners both care about. That means Celaxis customers have a downstream reporting use case beyond just the energy bill savings.
We're local, which means we can sit in the same room as the facilities manager during onboarding and the first week of optimization. That matters when you're putting software in front of 25-year-old building systems.
Want to talk before you evaluate?
We are a small team working with a focused set of buildings. Ingrid responds directly to conversations with facilities managers evaluating Celaxis — no sales process, no SDR, no demo-to-proposal pipeline. If your building isn't a fit, we'll say so.
Reach out directly