In February 2026, MARA acquired a majority stake in Exaion, a European private cloud company delivering secure, enterprise-grade AI infrastructure. The acquisition brings together two companies that share a vision for what private AI infrastructure can become. While many focus on the models themselves, the infrastructure layer underpinning them represents an emerging opportunity.
Exaion has been building that layer: a private cloud platform designed for AI workloads that enterprises cannot entrust to shared, public infrastructure. We spoke with Fatih about where private AI is headed and what this partnership means. The following are his responses, condensed from our conversation.
What problem was Exaion created to solve?
Exaion was born inside EDF Group for a specific reason. Our core team had spent years experimenting with emerging technologies: blockchain, AI, IoT, and cybersecurity. At the same time, we were employees inside one of the most sensitive enterprises in Europe – a company managing more than 55 nuclear power plants.
We kept bringing breakthrough ideas back to EDF, and at some point they said: you understand how startups think and move, but you also understand the values and processes of a large enterprise. We need you to build a bridge between those two worlds.
The goal was to make sure EDF never fell behind on emerging technology, with the agility to evaluate it and implement what mattered. We became a kind of commando unit inside EDF. That's why Exaion exists, and why EDF has chosen to remain a customer and shareholder after MARA’s investment.
Why are enterprises moving toward private cloud?
The private cloud trend is strong right now, and we are just at the beginning of what I consider to be phase two of AI for enterprises. Phase one started with Claude, Gemini, ChatGPT, and so on. Enterprises wanted to know: can it help me translate a document, write a report, or build a presentation? And at that stage, you are not sharing anything particularly sensitive.
But then enterprises hit a wall. AI helps with simple tasks, but it does not solve deeper problems. It does not help you cut costs by 30% on a specific workflow or move significantly faster on something that really matters. To get there, you cannot share your most sensitive code or proprietary models on a public platform. That is where private cloud comes in, and because we started building for these requirements in Europe years before it became a mainstream conversation, we believe we are years ahead of the market.
Where does Exaion fit in the cloud landscape?
The cloud industry is segmenting into three pieces: public cloud, sovereign cloud, and private cloud. The hyperscalers (AWS, Azure, GCP, and in France, Scaleway and OVH) own the public layer. Sovereign cloud is a live challenge in Europe, though less so in the US where the hyperscalers are already considered sovereign solutions. Private cloud is the newest and where Exaion operates.
Think of a pharmaceutical company developing drugs. Their emails live on the public cloud. Their client data lives in a sovereign cloud with compliance guarantees. But their drug formulas, the proprietary information they cannot share with anyone, live on a private cloud. Exaion operates on this third layer.
Exaion builds the infrastructure, the software, the cloud orchestration, the platform, everything embedded inside the company. The result is an internal private AI cloud where that pharmaceutical company can run parallel calculations on their most sensitive data, working toward the next generation of solutions.
Why team up with MARA?
At the macro level, we share the same vision and the same priorities. MARA is a world-leading Bitcoin mining company deliberately integrating AI and enterprise-grade solutions, and we have the same product roadmap for getting there.
The technological complementarity is also significant. MARA has the land, the power, and deep expertise in containerizing compute at scale. Now apply that same model to AI inference for enterprises and layer in Exaion's cloud orchestration expertise, our security capabilities built for institutions, and our experience integrating AI models into enterprise platforms. Add a shared commitment to renewable energy and you have something extremely powerful. What MARA did for Bitcoin mining, we believe we can do together for private cloud AI infrastructure.
What are you most excited to build with MARA?
My dream is simple: to become Europe's leading AI inference private cloud company. If we can do for private AI cloud what MARA did for Bitcoin mining, it would be truly remarkable. And I believe we can.
What gives me confidence is not just the technology. There is also a cultural complementarity that I find powerful: the go-getter, scale-focused business mentality of the US combined with the precise engineering mindset of France. Those two approaches together are a strong foundation for building something great.
The cloud industry is segmenting, inference AI is arriving in a serious way, and we already have much of the solution in place. The combined history and expertise of MARA and Exaion creates something both strong and genuinely needed. And honestly, sometimes it is hard for me to sleep at night. There is so much we can build together that I find myself looking forward to the next day, the next meeting, the next step. That energy does not go away.
To learn more about MARA's vision for digital energy and AI infrastructure, read our latest Shareholder Letter.
Investor Notice
Investing in our securities involves a high degree of risk. Before making an investment decision, you should carefully consider the risks, uncertainties and forward-looking statements described under the heading "Risk Factors" in our most recent annual report on Form 10-K and any other periodic reports that we may file with the U.S. Securities and Exchange Commission (the "SEC"). If any of these risks were to occur, our business, financial condition or results of operations would likely suffer. In that event, the value of our securities could decline, and you could lose part or all of your investment. The risks and uncertainties we describe are not the only ones facing us. Additional risks not presently known to us or that we currently deem immaterial may also impair our business operations. In addition, our past financial performance may not be a reliable indicator of future performance, and historical trends should not be used to anticipate results in the future. See "Forward-Looking Statements" below.
Forward-Looking Statements
This press release contains forward-looking statements within the meaning of the federal securities laws. All statements, other than statements of historical fact, included in this press release are forward-looking statements. The words "may," "will," "could," "anticipate," "expect," "intend," "believe," "continue," "target" and similar expressions or variations or negatives of these words are intended to identify forward-looking statements, although not all forward-looking statements contain these identifying words. Such forward-looking statements are based on management’s current expectations about future events as of the date hereof and involve many risks and uncertainties that could cause our actual results to differ materially from those expressed or implied in our forward-looking statements. Subsequent events and developments, including actual results or changes in our assumptions, may cause our views to change. We do not undertake to update our forward-looking statements except to the extent required by applicable law. Readers are cautioned not to place undue reliance on such forward-looking statements. All forward-looking statements included herein are expressly qualified in their entirety by these cautionary statements. Our actual results and outcomes could differ materially from those included in these forward-looking statements as a result of various factors, including, but not limited to, the factors set forth under the heading “Risk Factors” in our most recent annual report on Form 10-K, and any other periodic reports that we may file with the SEC.
