When Sam Ramadori first met Jean-Simon Venne, he wasn’t looking for a career change. He was deep in private equity, investing in traditional industries—manufacturing, industrial operations, and companies built on physical infrastructure, not software. He knew how hard it was to make even modest energy efficiency gains.

Shaving 3% off energy costs in a factory meant a team of 10, a multi-million-dollar budget, months of disruption, and a painful approval process. That was just the reality of dealing with physical industries notoriously slow to embrace anything that disrupts business as usual.

Then Venne laid out his vision for BrainBox AI—and Ramadori nearly fell off his chair.

Instead of installing expensive new hardware, ripping out legacy systems, or forcing complex retrofits, BrainBox AI would layer intelligence on top of what already existed. It would tap into a building’s control systems, optimize them in real time, predict problems before they happen, and slash energy waste—all without turning a single screw.

The promise wasn’t a few percentage points in savings. It was 10, 20, and sometimes even 30% reductions in energy use and carbon emissions. It was possible because BrainBox AI offered customers an autonomous AI system that continuously digested data and continuously learned how to do its job better.

Sam knew what he had to do: explain to his wife why a guy with no technical background was leaving his financially secure future to run an artificial intelligence startup.

For years, the climate conversation about emissions has focused on cars, coal plants, and industry. But buildings are among the biggest culprits—accounting for nearly 40% of global carbon emissions.

Half of that comes from heating and cooling.

HVAC systems are dumb, reactive, and wasteful. They don’t anticipate; they just respond, often working against the building’s needs and the energy grid.

A self-driving car anticipates obstacles, reroutes and adjusts in real time. HVAC systems? They’re stuck in the past, mindlessly burning through energy.

BrainBox AI changes that.

A thermostat, even a "smart" one, only knows what’s happening now. BrainBox AI sees six hours into the future—with 99% accuracy.

It factors in: 

✅ Real-time weather data (Is the sun about to hit those west-facing windows?) 

✅ Building occupancy shifts (Is an office floor emptying out for the day?)

✅ Energy prices (Is electricity about to get more expensive?) 

✅ Carbon intensity (Is the grid producing renewable energy or mostly fossil fuels?)

By adjusting proactively, BrainBox AI smooths operations, navigates around costly energy spikes, and reduces the overall carbon footprint.

Ramadori and Venne understand—and the industry is slow to realize—that buildings are drowning in data, but nearly all of it is discarded.

Every thermostat reading, fan speed adjustment, and heating and cooling cycle—millions of data points go to waste daily.

In an era where data is the new oil, buildings might be the biggest wasters of intelligence on the planet.

BrainBox AI’s approach isn’t just smart—it’s fast, seamless, and cost-effective from Day One.

When Dollar Tree signed on with BrainBox AI, 600 locations across 18 states were operational on its platform within months—without heavy equipment, installation downtime, or disruption to store operations.

The AI simply plugged into the existing systems, started learning, and began optimizing.

One Dollar Tree executive put it admiringly:

“BrainBox AI helps you offset costs from day one. It’s also incredibly flexible; you get new equipment, it adapts. You move to a new location, it moves with you. You grow, it grows with you.”

That’s the real superpower of BrainBox AI in buildings: It keeps learning.

If a retailer remodels, BrainBox AI relearns.
If an office gets new windows, BrainBox AI recalibrates.
If energy prices spike, BrainBox AI shifts its strategy in real-time.

For building owners and operators, this isn’t just better energy efficiency; it’s a business advantage.

The HVAC industry has another problem: There aren’t enough skilled technicians.

North America alone employs 400,000 technicians but still has a deficit of 100,000 unfilled jobs.

That means thousands of buildings—offices, warehouses, stores, hospitals—operate inefficiently simply because no one has time to fix them.

BrainBox AI’s virtual engineer, ARIA, is changing that.

ARIA acts as an on-demand AI assistant for maintenance teams. It can:
🔹 Diagnose issues remotely (so a technician knows what’s wrong before they arrive).
🔹 Recommend fixes and suggest spare parts (so there are no wasted trips).
🔹 Predict machine and part lifespans (so problems are prevented before they happen).

In an industry where labor is stretched thin, ARIA isn’t replacing workers—it’s making them radically more efficient.

BrainBox AI’s ambition goes beyond individual buildings.

Historically, power grids have focused on supply, but preparing them to satisfy future needs requires developing new intelligent demand capabilities.

Next up for BrainboxAI: weaving the brains of all of its AI-powered buildings into a network to scale a demand-response business that can:

With Trane Technologies’ acquisition of BrainBox AI last month, the goal is clear: scale the AI-driven transformation as widely and quickly as possible.

Listen to this podcast episode on Apple, Spotify, YouTube, and all other platforms.

↓

That’s the square feet of commercial real estate that BrainBox AI operates—equivalent to about half the office space in Manhattan. It’s achieved this remarkable scale in just under six years, making its first sale in May 2019.

“AI isn’t replacing technicians—it’s making them radically more effective. Instead of walking into a building blind, they get a full diagnostic before they even arrive.”

—Sam Ramadori, CEO of BrainBox AI

↓

As solutions like BrainBox AI expand, we’re tracking the data centers powering AI’s growth. We recently spoke with Nazar Khan, CTO of TeraWulf, a company building hydro-powered AI data centers in New York—proof that low-carbon, high-performance computing (HPC) is now possible.

Still, data centers account for 2–3% of global emissions, and demand is rising. But few in the industry are standing still.

Take Nvidia, the world’s most valuable company and the central actor in AI’s exponential growth. Its newest GPU, Blackwell, is 96% more energy efficient than Hopper, released just two years ago. Over the past decade, Nvidia’s GPUs have become 45,000 times more energy efficient.

Yet efficiency alone won’t curb AI’s energy appetite. The Jevons Paradox tells us that as AI becomes cheaper and more accessible, usage will skyrocket, likely offsetting efficiency gains.

That’s why a new wave of startups is reimagining all aspects of AI digital infrastructure.

Cooling is one of AI computing’s biggest hidden costs and emissions sources. It consumes nearly 30% of a data center’s electricity to prevent chips from overheating. In response, data centers are shifting from air cooling to more effective yet technically complex liquid cooling.

We’re tracking Switzerland-based startup Corintis. Its breakthrough is microfluidic cooling embedded directly in the chip, which extracts ten times more heat than conventional liquid cooling solutions.

The impact:
✅ More powerful AI without overheating.
✅ Lower cooling costs and water use.
✅ A significant cut in energy waste—and emissions.

As AI’s energy demand soars, every efficiency gain matters.

↓

Hit the button below for proven solutions that cut carbon, boost the bottom and improve modern life—delivered weekly to your inbox.

🌐