Arbor Energy

Corpus Christi, Texas was on track to become the first American city to run out of water this year. Recent floods bought it another year but with industrial users responsible for more than half of regional demand, nobody there is calling it solved. That's the backdrop every new data center announcement lands against now. A single hyperscale campus can use up to 5 million gallons a day, comparable to a town of 10,000 to 50,000 people. In places like South Texas, that water often comes from the same system residents are already being asked to ration. From liquid cooling to reclaimed wastewater systems, solutions are already being deployed at scale. Novel power systems that generate water instead of consuming it go a step further. Read the full article to learn how Arbor Energy’s modular turbines help solve one of the hardest community issues for data centers: https://lnkd.in/gMwdv3kg Partner content, brought to you by Arbor Energy.

A single hyperscale data center can use up to 5 million gallons of water a day, roughly what a town of 10,000 to 50,000 people needs. Our CEO Brad Hartwig writes in Latitude Media about how developers are starting to solve for this issue: liquid cooling that skips evaporation entirely, closed-loop systems that recirculate instead of drawing new supply, reclaimed water sourced from treatment plants instead of the municipal system. Arbor's power systems don't pull from local water sources. Because they run on oxy-combustion, burning fuel with pure oxygen instead of ordinary air, the exhaust is mostly CO₂ and water vapor, and we condense that vapor into usable water. At 100 megawatts, that's roughly 100 million gallons a year, enough to run a co-located facility's entire cooling loop without drawing on the community's supply. Read the piece at the link in the comments.

Our engineering team passed significant turbomachinery milestones last week, and we gathered as a company to raise a toast. We'll have more to share about the specifics soon, but for now we can say that these past months of intense testing and development have produced capabilities that will make every test ahead faster. The team has learned to build up and control the system precisely under conditions few people have run, with abort and fail-safe logic that lets us push hard and repeat safely. We've cut failure recovery and diagnostic time by roughly a factor of ten, so a problem that surfaces during a test can be diagnosed and back under test in days instead of weeks. Strong supplier relationships and our own shop floor now turn new parts around fast enough to keep a campaign moving through hardware failures. Weeks like this are a good reminder of how far Arbor has come. Five years ago this company was an idea and a spreadsheet. We're grateful for the team that has spent the years since building toward clean, dependable baseload power, and proud of what they’ve built.

Arbor's 25 MW turbine is small enough to 3D print on existing equipment. That means the manufacturing supply chain that constrains the rest of the industry doesn't constrain us in the same way. Our own Josh Raycroft, who started his career at GE, walks through why flexibility matters and what 3D printing a turbine the size of a dinner plate unlocks. Head to the blog post in the comments for more on the broader baseload infrastructure picture.

AI data center developers used to face one core challenge, which was finding enough power. Now they face a second one just as pressing: earning the support of the communities hosting them. In this partner content, brought to you by Arbor Energy, Brad Hartwig argues that community concerns including air quality, water use, noise, and local economic impact, shouldn't be treated as a communications problem to manage after the fact. They should be engineering requirements, built into a project's design from day one, alongside grid constraints and permitting timelines. Arbor Energy points to Microsoft's 2-gigawatt Pecos, Texas campus as an example of the shift already underway, and details how it's applying that same approach to its own modular turbine facility in Louisiana, which highlights emissions-free operation and a semi-closed-loop water system designed to avoid drawing on local supply. Read the full article now: https://lnkd.in/gMBnVRVf

Legacy aerospace built its reputation on requirements handed down and rarely questioned. SpaceX famously upended that culture with a first-principles approach instead: push back on a requirement before you accept it, because once it's accepted, you're expected to meet it. If you can't, the design isn't done. Many of our engineers came from SpaceX, and we apply that same first-principles mindset here at Arbor. Our co-founder and CEO Brad Hartwig writes in Latitude Media about extending that discipline to the questions communities are asking about AI infrastructure: air quality, water, rates, roads. Those concerns arrive late in most projects, after the engineering is finished and the site is chosen. He argues they belong at the beginning. Find the article at the link in the comments.

Increasingly, a baseload power project doesn't move forward unless it clears three requirements at once: fast enough to meet the timeline, clean enough to keep the community on board, and free of the emissions liability that only compounds over time. No existing option has cleared all three at once, and each falls short for its own reason: gas on community impact and long-term emissions, nuclear and geothermal on speed or siting, renewables and storage on firm, around-the-clock output. Our latest blog post breaks down why that combination has been so hard to deliver, and what it takes to resolve all three at the level of the system itself. Find it at the link in the comments.

We're hiring a Senior Thermodynamics Performance Engineer to own the cycle simulation for our supercritical CO₂ power system, from steady-state performance through full dynamic behavior. You'll run the studies that precede each test campaign and reconcile your predictions against real data as the hardware comes to life this year. Over time, the models you build will grow into a digital-twin-quality platform underpinning controls, trip response, and hardware-in-the-loop testing. This is the simulation backbone for a first-of-a-kind machine, and foundational work on a system very few people have modeled before. If that's the kind of problem you've been looking for, find the role in the comments and reach out to Phillip Cabrera to learn more.

Last week we brought the whole Arbor team together at our El Segundo headquarters, gathering local and remote Arbornauts and pulling the test team home from our pilot site in San Bernardino. Most of the year, the team is deep in the hard problems: turbomachinery spinning near the critical point, combustion systems proven one test at a time, offtake discussions underway with some of the largest companies in the world. The zero-emission power systems we're building will take dedicated work and focus to bring fully into the world. But we also know the value of laughter and taking breaks to recharge; while we take our mission seriously, we try not to take ourselves too seriously. So we shared meals, unveiled our values wall, and held a town hall where the team hit leadership with every hardball question they could think of. Mostly, we enjoyed time to get to know our new teammates and catch up with old friends. If building dependable, clean baseload power alongside a team like this sounds like your kind of work, we're hiring. See all our open roles in the comments. Just know that an Arbor event always ends in karaoke.