0
This week brings a pleasing balance of electric mobility and deeptech news to break up the steady drumbeat of AI funding announcements — though of course there are plenty of those, too. To kick it off, Rivian spinoff Also announced a sizable Series C round just a year after its last fundraise to buoy its lineup of electric bikes and compact quad vehicles. There’s also fresh funding for Via Separations, which is working to electrify the kind of high-heat industrial processes that most of us depend on but never think about. And on the AI front, there’s new capital for data center flexibility platform Emerald AI and grid intelligence company ThinkLabs AI.
Our humble grid is sure getting complicated. Good thing there’s a whole host of companies now looking to build data centers in space! More on that, too.
Micromobility Startup Also Nabs $200 Million and a Partnership With DoorDash
In the U.S. over half of all car trips are under 6 miles, and about 80% are 15 miles or fewer. For many of these short journeys, a full-sized car with five seats, a spacious trunk, and precise climate control is simply not necessary. That’s where micromobility solutions come in — and where the Rivian spin-off Also sees its niche. The company is building smaller EVs from e-bikes to quads capable of carrying multiple passengers or hundreds of pounds of cargo while still fitting in the bike lane.
This week, the startup announced a $200 million Series C round led by Greenoaks Capital, pushing the company's valuation to $1 billion — not bad considering it spun out of Rivian just over a year ago. DoorDash joined the round as a strategic investor, inking a multi-year deal with Also to develop autonomous delivery vehicles to tackle last-mile challenges. “The intersection of roads and road-adjacent spaces, such as bike lanes, shoulders and curbsides, are the areas that make up the hardest part of the last-mile delivery puzzle,” the company states in its release, explaining these environments are where Also has “the greatest opportunity to perform.”
Also has an additional corporate partnership with Amazon, announced last fall, to design a pedal-assist cargo quad for deployment across Europe and the U.S. This vehicle is slated for launch this year, while the company’s bike is already available for pre-order and expected to begin shipping soon.
Via Separations, Maker of Industrial Membranes, Secures $36 Million to Cut Emissions in Refining and Chemicals
Industrial separations — the process of extracting a specific chemical or material from a mixture — may not immediately scream “climate tech.” It’s one of those foundational techniques that you rarely think about, yet somehow underpins everything from paper and pulp production to plastics and oil refining. But Via Separations thinks it’s found a way to perform this industrial necessity in a way that’s significantly less energy and emissions intensive — and this week it raised $36 million to do it at scale.
Today, industrial separations typically rely on heat-based processes like distillation, which sorts out substances based on their differing boiling points. But heating and reheating all that liquid requires boatloads of energy, and thermal separation as a whole accounts for roughly 12% of global energy use.
Via’s approach electrifies this process using membranes that allow only specific substances to pass through. It’s made advances in designing durable membranes that can perform under harsh industrial conditions, and now claims its process can cut energy use by up to 90% at the separation stage. Via has already demonstrated its tech at a Canadian pulp mill, where it’s operated for nearly two years. Now, as the startup moves into the much larger refining and chemicals industries, it says it’s completed a pilot at an unnamed Gulf Coast refinery and has hundreds of millions of dollars in projects lined up.
Climate Investment — a firm founded by a coalition of oil and gas companies — led the round alongside Aramco Ventures and Marathon Petroleum Corporation, which are all interested in putting Via’s tech to work in the oil refining and chemicals markets.
Emerald AI Lands $25 Million to Turn Power-Hungry Data Centers Into Grid Assets
It’s no secret that data centers are insatiable power consumers, and that our modern grid simply wasn’t built to handle the amount of new load they’re bringing online. As I wrote last summer, the startup Emerald AI is confident this challenge can be largely solved by turning data centers from “grid liabilities into flexible assets.” By slowing, pausing, or redirecting AI workloads when energy demand is peaking — a mere 0.5% of the time — Emerald estimates it could unlock up to 100 gigawatts of existing grid capacity, enough to power about 83 million U.S. homes for a year.
It’s a compelling vision, already backed by prominent investors including Nvidia’s venture arm, former U.S. Special Presidential Envoy for Climate John Kerry, Kleiner Perkins chair John Doerr, and Lowercarbon Capital. This week, Energy Impact Partners joined the mix, leading Emerald’s $25 million expansion round joined by other strategic investors such as GE Vernova and Siemens.
The funding follows last week’s CERAWeek announcement that Emerald and Nvidia are partnering to launch flexible-demand “AI factories” alongside energy companies including AES, Constellation, and NextEra Energy. To avoid the backlogged interconnection queues, these facilities will initially rely on co-located power. Then once they’re able to connect, their co-located energy and storage assets will flip to providing flexible grid services, storing excess cheap energy and providing power back to the grid during times of peak demand.
As Emerald’s CEO Varun Sivaram said in a press release about the partnership, “AI factories are too valuable to be treated as either passive loads or permanent islands.”
ThinkLabs AI Raises $28 Million for High-Speed Grid Modeling
AI-driven load growth is undoubtedly straining the limits of our outdated grid — but it’s also giving planners and operators new tools to run it more efficiently and reliably. This week, grid intelligence company ThinkLabs AI raised a $28 million Series A round, also led by Energy Impact Partners, to scale its software for modeling power flow on the grid.
In an era dominated by large language models, ThinkLabs says it’s doing something fundamentally different — training AI on physics-based simulators to model grid behavior in real time, making it possible to rapidly test a wide range of hypothetical scenarios. How rapidly? The startup says it can complete planning studies that once took months in a matter of minutes and run 10 million scenarios in 10 minutes, all while maintaining greater than 99.7% accuracy.
This allows utilities to proactively plan for emerging stresses — from new data centers and clusters of EV chargers coming online to extreme weather events that threaten critical infrastructure. "The legacy tools and processes utilities currently rely on can take months to complete a single study, cost tens to hundreds of thousands of dollars in engineering time, and the results are out of date the moment the study is finished,” Josh Wong, ThinkLabs’ CEO, said in a press release. “ThinkLabs’ AI-native high performance grid simulation model not only shows you the problems but also gives you the best solutions."
Bonus: Starcloud Blasts Past $1 Billion Valuation With Bet on Space-Based Data Centers
With SpaceX planning to go public and Artemis II on its way to the moon, the Earth feels abuzz with hope over extraterrestrial infrastructure. Now, the startup Starcloud wants to build data centers in space, and it just raised a $170 million Series A to help make it happen. Investors clearly don’t think the concept is as far-fetched as it sounds, given that they have valued the company at over $1 billion, a mere 17 months after its graduation from Y Combinator.
Worldly concerns such as grid interconnection queues, aging transmission systems, and mounting political opposition don’t apply to orbital data centers, though a laundry list of more technical challenges definitely do. But Starcloud appears undeterred, launching its first satellite equipped with an Nvidia GPU last November. It’s now preparing a more advanced satellite for later this year, outfitted with multiple GPUs and a Bitcoin-mining computer, of course.
