Skeleton Technologies raised a €33 million first close of a larger round as it prepares for a planned US IPO in 2027, according to EU-Startups. The Tallinn company says the financing brings total venture funding to €392 million and adds Axon Partners Group, SmartCap and Taiwania Capital to its investor base. Its pitch is direct: AI data centers are running into power constraints, and fast energy storage should be part of the fix.

Chips still get the glamour. Power gets the invoices. Skeleton is betting that the next phase of AI infrastructure will force investors to care about what happens between the grid and the rack.

The unsexy layer gets interesting fast.

Skeleton makes supercapacitors and SuperBattery systems based on its Curved Graphene technology. Unlike conventional batteries optimized mainly for energy density, supercapacitors are built for fast charge and discharge. That makes them useful where power spikes, grid stability and high-cycle performance matter. Data centers, industrial systems and mission-critical infrastructure all fit that pattern.

The company claims its systems can cut AI data center energy consumption by up to 44 percent. That is a big number, and customers will test it carefully. But even a smaller improvement would matter in a market where AI compute demand is straining grids, delaying projects and forcing operators to rethink everything from siting to backup power.

New investors include Axon Partners Group, SmartCap and Taiwania Capital. The Taiwania connection is notable because Taiwan sits at the center of the AI hardware supply chain. If power delivery becomes a bottleneck for AI infrastructure, strategic capital may start looking beyond chips into the systems that let chips run reliably.

Skeleton is one of the more industrial companies in the Nordic and Baltic tech landscape. Founded in 2009, it has spent years turning materials science into manufactured products. That is a very different rhythm from a typical AI application startup. The company operates manufacturing in Germany and has talked about a one-gigawatt SuperBattery factory in Leipzig.

This matters because Europe's AI competitiveness is often discussed as a software or model problem. Skeleton reframes part of it as an infrastructure problem. The region may not control every frontier model, but it can build enabling layers around power, cooling, grid stabilization and industrial systems. Germany Trade and Invest has pushed Germany as a location for energy and industrial technology investment, and Skeleton sits squarely in that lane.

There is a Baltic story too. Estonia is small, but it has repeatedly produced companies with outsized European relevance. Skeleton is not another SaaS export. It is a reminder that the region's tech ecosystem includes hard infrastructure, not only digital services.

A planned US IPO in 2027 gives the company a clock. Public-market investors will want growth, margins, customer concentration details and proof that the AI data center opportunity is converting into orders. The SEC path can reward infrastructure stories, but it also exposes them to harder questions than private investors sometimes ask.

Skeleton's advantage is that it can attach itself to a demand curve everyone can see. AI infrastructure buildout is not theoretical. Data center developers are hunting for power, utilities are warning about capacity, and countries are treating compute as strategic infrastructure. A company selling power-performance improvements can find receptive ears.

The challenge is proving that supercapacitors and SuperBattery systems are not a niche add-on but a necessary layer in AI facilities and industrial grids. That requires reference customers, measurable savings and manufacturing scale. It also requires education, because many investors still file energy storage into familiar battery categories that do not fully capture what Skeleton is selling.

This round is useful because it widens the frame. AI is not only models, GPUs and applications. It is substations, transformers, cooling loops, batteries, capacitors, permits and grid contracts. The companies that solve those constraints may not look like software startups, but they will shape how much AI infrastructure can actually be built.

For NordicTech readers, Skeleton is a reminder to watch the physical stack. The next bottleneck may not be model quality. It may be whether the power system can keep up when every company wants more compute at the same time.

If Skeleton can turn that pressure into public-market-ready growth, the IPO story will not just be about an Estonian company going to the US. It will be about energy storage becoming part of the AI economy's core plumbing.

Data center operators used to treat power as a site-selection constraint. Now it is becoming a board-level strategic question. Can the grid support the load? Can the facility handle spikes? Can backup systems respond quickly enough? Can efficiency gains reduce both cost and political pressure? Skeleton's products sit in that set of questions. They are not glamorous, but they touch the economics of every AI buildout.

The supercapacitor angle is especially relevant because AI infrastructure is not a smooth, polite electricity consumer. Training and inference workloads can create demanding power profiles, while cooling and backup systems add their own complexity. Fast-response storage can help manage those dynamics. The company still has to prove the scale of the benefit facility by facility, but the need for better power management is obvious.

The IPO ambition adds discipline. Private deeptech stories can survive for years on technical promise. Public-market stories need clearer translation: revenue growth, gross margin, backlog, customer proof and a credible path through manufacturing scale. Skeleton will have to explain itself not only to energy specialists but also to generalist investors who may know AI demand but not supercapacitor physics.

The wider lesson is that the AI economy is pulling old industrial categories into the tech spotlight. Grid equipment, thermal systems, energy storage and power electronics are becoming part of the same conversation as chips and models. That is good news for the Nordics and Baltics, where industrial engineering and clean-energy expertise already run deep. The AI boom may end up rewarding the places that know how to build the physical layer, not only the places that know how to demo the software layer.

Energy storage companies often sell to engineers first. The AI data center cycle may pull CFOs and boards into the conversation earlier. Power constraints can delay revenue. Efficiency improvements can change project economics. Grid upgrades can become political fights. If Skeleton can make a credible case that its systems reduce energy waste, smooth power demand or improve resilience, the buyer is not only the facility engineer. It is the executive responsible for whether a compute campus opens on schedule.

This is where the claimed 44 percent energy-consumption reduction becomes both powerful and dangerous. Big claims attract attention, then scrutiny. The company will need customer evidence that separates laboratory performance from real facility economics. Investors will ask where the savings show up, how repeatable they are, and whether the product is a must-have component or a useful optimization.

The manufacturing story matters too. Deeptech companies can win pilots and still struggle to scale production. Skeleton has spent years building industrial capacity, which may give it more credibility than a younger lab spinout. But public-market readiness requires a different level of operational transparency. Supply chain resilience, factory utilization and quality control will become part of the investor narrative.

For the region, the company is a useful counterweight to the idea that Europe's AI role is only regulation or application software. Power systems are strategic. Materials are strategic. Grid stability is strategic. Skeleton is trying to turn that reality into a growth company. If it succeeds, the next AI infrastructure winners may include more Baltic and Nordic industrial firms than the current hype cycle assumes.

The political backdrop is favorable but complicated. Governments want AI capacity, clean power and industrial competitiveness, often at the same time. Data centers can strain local grids and attract public resistance, especially when communities feel they bear the infrastructure burden without enough local benefit. Technologies that improve efficiency and resilience may become part of the social license for expansion.

Skeleton will also need to tell a simple story about a complex product. Supercapacitors, graphene and grid dynamics can lose a generalist investor quickly. The clearest version is this: AI needs power that is fast, stable and efficient, and Skeleton sells systems meant to make that possible. If the company can keep the narrative that direct while proving the numbers, the 2027 IPO target becomes more plausible.

The timing is fortunate. AI demand has made power constraints visible to investors who once ignored them. Skeleton now has to show that visibility can become orders, revenue and public-market confidence. Soon. Before the grid story becomes just another bottleneck investors learned to price in.