Exascale is now public. Exascale Labs Holdings Inc. has closed its business combination with D. Boral ARC Acquisition I Corp. Class A common stock and warrants begin trading on Nasdaq tomorrow, August 28 2026 under solana:XLnpFRQ3rSWupCRjuQfx74mgVoT3ezVJKE1CogRZxhH and $XLABW. We’re a next-generation AI infrastructure company. The constraint isn’t just GPUs- it’s deployable capacity: power, cooling, cluster management, and getting real workloads online. That’s what the platform is built for. As CEO Dr. Hoansoo Lee put it: "We’re entering the public markets at a time when demand is shaped by the ability to deploy reliable infrastructure at scale." Full announcement: globenewswire.com/news-relea…

Exascale Labs ($XLAB) announced its post-closing Board of Directors. Dr. Hoansoo Lee: the Board is built for the discipline it takes to scale a capital-intensive AI compute platform in public markets. Chair: founding investor Wenying Jia Lead Independent Director: Nobel Laureate David Card Also on the Board: Berkeley economist Shachar Kariv, Samsung Ventures’ Jaeyoung Shin, and CEO Hoansoo Lee Full release: globenewswire.com/news-relea…

We’ve signed a Letter of Intent with Compal Electronics (TWSE: 2324) to jointly develop a U.S. platform that validates native 800 VDC GPU systems in a full data-center environment. The idea is simple: test power, compute, cooling, and controls as one system- Compal’s production GPU servers and liquid cooling with our modular data center and SST architecture- before anyone talks large-scale manufacturing or certification. Full release: globenewswire.com/news-relea… #AIInfrastructure #DataCenters $XLAB

Exascale (Nasdaq: solana:XLnpFRQ3rSWupCRjuQfx74mgVoT3ezVJKE1CogRZxhH) and EnergyBank signed a non-binding MOU to explore grid-independent AI compute powered by floating offshore wind and long-duration storage. The work contemplates an initial pilot with Zephyros-One- the world’s first floating offshore wind turbine- targeting up to 800kW of compute-ready power on an existing turbine. The parties also intend to define a repeatable reference architecture for later multi-turbine projects, potentially targeting up to 10MW of firm compute per turbine. The constraint isn’t chips. It’s power that can actually be delivered on a commercial timeline. This collaboration is about assembling generation, storage, cooling, and modular compute as one system instead of waiting years for a grid connection. Full announcement: globenewswire.com/news-relea… #AIInfrastructure #OffshoreWind solana:XLnpFRQ3rSWupCRjuQfx74mgVoT3ezVJKE1CogRZxhH

Exascale Labs spent the day at Compal Server Austin discussing modular data centers, SST, next-gen AI servers. We all know chips get the headlines. But the AI buildout will be won by whoever ships infrastructure fastest. Our partnership delivers on that. More to come. #AIInfrastructure #DataCenters #compalserver

Our CEO Dr. Hoansoo Lee will join Needham’s Virtual AI Infrastructure 1x1 Conference on Aug 12 for meetings with institutional investors. To schedule: contact your Needham rep or email exascale@kcsa.com Full release: globenewswire.com/news-relea…

Hear it from the investor, @Taisu_Ventures. Discover what builds confidence in AI infrastructure execution. Exascale's operational thinking, handling complex challenges like GPU procurement and conversion, sets us apart. Learn why deep execution is rare and vital.

What makes the Exascale Labs team different? Mark Fidelman, Head of Marketing: “One thing we have that I haven’t seen anywhere else is a team with real talent and zero egos. This is a new space. We’re building things no one’s built before- and no one’s assembled before. The camaraderie and lack of egos is truly special.” Curious what CEO Hoansoo Lee has to say? Watch the video.

Access to a GPU is not the same as access to usable AI capacity. The industry has poured money into chips. But customers still face the hard part: power, high-density cooling, networking, integration, commissioning, and actually running clusters at high utilization. We address that with two complementary businesses: GPU-as-a-Service- reserved or on-demand compute with cluster management and optimization Infrastructure solutions- modular data centers, liquid cooling, high-voltage DC power, and related energy systems Founded on a simple observation in 2022. The goal remains the same: help customers bring compute online in months, not years- while staying disciplined and asset-light.

Most people think AI is the app. ChatGPT. Claude. Midjourney. But the real power sits underneath. GPUs in data centers. Massive power. High-density cooling. Networking. Uptime. Operators who actually know how to run it without turning the whole thing into a science experiment. AI isn’t floating in the cloud like some invisible genie. It’s more like a race car. The apps are what you see on the track. We’re helping build the track, the fuel systems, the pit crew, and the garage. Without that layer, nothing moves. Demand is exploding. Execution is hard. That’s the part of the stack we’re in.

Dr. Hoansoo Lee on what he hopes Exascale is known for in 5–10 years: Not infrastructure for its own sake- but for what it enables. Better AI products. Faster research. More reliable deployments. Companies that weren’t possible before. Watch the full answer ↓

Exascale Labs and D. Boral ARC Acquisition I Corp. (Nasdaq: BCAR) today announced that BCAR shareholders approved the previously announced business combination at today’s Extraordinary General Meeting. The parties expect the transaction to close shortly, subject to remaining closing conditions. Upon closing, the combined company is expected to operate as Exascale Labs Holdings Inc., with Class A common stock and warrants expected to trade on Nasdaq under the symbols “XLAB” and “XLABW.” Following redemptions, approximately $12 million remains in BCAR’s trust account (net of transaction expenses), satisfying the minimum cash condition. No additional financing is currently anticipated prior to closing. Full details: globenewswire.com/news-relea…

Exascale Labs CEO Hoansoo <@hoansoo> Lee on the real unit economics of a 1MW build: Market price for building + power + cooling (ex-GPUs & land) usually runs $12–15M. We’re highly disintermediated with direct supplier access, so our costs sit meaningfully lower. After a few layers of 30% contractor markups, the market is often paying more than 2x true cost. That gap is the surplus we share with customers. GPU leases: ~$3.40-$4.60/hr per chip (3-5 yr terms). Conservative pipeline view referenced: $300M. Sharing from the discussion. Not investment advice.

Just dropped the full Exascale Labs Fireside Chat with CEO Hoansoo Lee <@hoansoo>. We talk about the real bottleneck in AI right now- not just getting GPUs, but turning them into reliable, production-ready compute. Power, cooling, deployment speed, and how we’re approaching it. Full conversation here: piped.video/ljLVYdneZ54

From Computex Taiwan to AMD in San Francisco, innovation doesn’t stop at the conference floor. Together with Compal, we’re scaling AI infrastructure globally through our co-designed Modular Data Center platform- enabling deployments in as little as 6 months.

Big step forward on our $300M customer pipeline at Exascale Labs. - Signed MOU with RUTILEA for potential 20MW data center capacity - $71.4M multi-year GPU-as-a-Service deal with Dimension AI - $53M AI Nova MOU advancing Executing across both AI infrastructure and GPU-as-a-Service. Full details: globenewswire.com/news-relea…

The 36-Month Problem: GPUs ship in weeks. But turning a site into live AI capacity still takes years. That gap is quietly destroying capital. At Exascale, we’re fixing it by treating data centers as manufactured products- not serial construction projects. Factory integration + parallel site work = 4-6 months to live racks. Speed to revenue isn’t optional anymore. exascalelabs.ai/blog/the-36-… -By Hoansoo Lee, CEO

🚨 Exascale Labs Announces Major Compute Service Agreement SAN FRANCISCO, July 16, 2026- Exascale Labs Inc. has entered into a three-year Compute Service Agreement with Dimension AI Pte. Ltd. to procure ~US$71.4M of dedicated GPU compute capacity. This expands capacity on our GPU-as-a-Service and token factory platform amid strong AI demand. Full press release: finance.yahoo.com/technology…

While everyone is focused on AI capabilities, far fewer are talking about the growing compute constraints ahead. As Dr. Hoansoo Lee has highlighted, while Moore’s Law delivered transistor doubling roughly every two years, AI system complexity has been doubling approximately every three months. This rapid growth is creating significant infrastructure challenges sooner than many expect. At Exascale Labs, we’re building high-performance computing infrastructure to help address these barriers.