Skip to main content
spacbrain

DBCA merger with Exascale Labs Inc.

Exascale Labs is an asset-light, software-defined AI infrastructure provider (United States)Revenue $13M (LTM 3/31/2026) as reported.

StatusTerminated
Announced deal valuenot stated in the filings we hold

Announced 1 July 2026.

Shareholder voteno vote date filed yet
IndustryTechnology — AI Infrastructure / GPU-as-a-Service (asset-light neocloud)

S-4 for Exascale Labs Inc. was WITHDRAWN under this entity (RW filed 2026-07-01). The Exascale business combination proceeds under a separate entity, D. Boral ARC Acquisition I Corp (BCAR) — not yet tracked. Verified vs EDGAR submissions API.


Structure & dilution

SEC-primary terms

The headline number ignores the shares that did not pay $10 — the founder promote, PIPE stock and warrants. This is the same deal with all equity claims counted.

Deal structureSEC-primary — BCA 8-K / S-4 / DEFM14A
Min-cash condition
$5M

An effective (post-dilution) figure needs either a stated pro-forma share count or the headline value plus the promote terms; the filings we hold do not yet state enough, and we will not print an estimate built on inventions.

Why headline and effective values differ is covered in headline vs effective deal value, in plain English.


The target: Exascale Labs Inc.

from DEFM14A

The business actually being bought — described from SEC primary filings, with projections labelled as projections.

Exascale Labs is an asset-light, software-defined AI infrastructure provider. Its revenue-generating business is GPU-as-a-Service (GaaS): it resells reserved and on-demand access to high-performance GPU compute that it sources from THIRD-PARTY data centers (it owns no core hardware), plus GPU cluster management and optimization services for AI data centre operators. It also markets modular data centre, high-density liquid cooling, HVDC power and energy-storage designs, but the filing states these have NOT YET GENERATED ANY REVENUE. Exascale IS revenue-generating - $7.0m in FY2025 and $10.6m in the nine months to 31 March 2026 - but it is a very small, loss-making, thin-margin reseller: gross margin is only ~16%, it has 10 full-time employees (4 in the US, 6 remote in Singapore), and it has negative shareholders' equity of $(20.9)m against a $500m deal value. It has never been profitable, with an accumulated deficit of $21.1m at 31 March 2026.

SectorTechnology — AI Infrastructure / GPU-as-a-Service (asset-light neocloud)
HeadquartersHouston, United States

Founded 2022.

Revenue$13M (LTM 3/31/2026)

A reported actual.

Employees10

source: 0001829126-26-007326opens on sec.gov in a new tab

Exascale Labs Inc. — every SPAC that has bid for it, and its listed peers


Expensive or cheap?

vs 2 listed peers

A price only means something next to what the same kind of business costs on the stock market. This divides what the buyers are paying by what Exascale Labs Inc. actually sells, and sets the answer against its closest listed comparables — or says plainly when that cannot be done.

SpacBrain’s read on the price

No multiple can be computed

No revenue multiple can be computed for Exascale Labs Inc. — no enterprise value to strike the multiple on.

No price has been announced for the deal, so there is no number to compare.

What the buyers are paying for the whole companyno price announced

No announced deal value, so there is no price to divide.

Divided by what the company actually sells in a year$13.1M

LTM 3/31/2026 — a reported actual.

= what this deal pays for every dollar of those salesno multiple

Not computable — no enterprise value to strike the multiple on.

What the stock market pays for its closest listed peers33.49×

$1 of their sales costs $33.49 on the open market. Median of 2 listed companies we judged a true comparable, which individually run from 13.79× to 53.18×. Their share prices are from 15 August 2026, not today.

What qualifies the figures above

  • No announced deal value — nothing to strike a multiple on.
  • AZIO, SHAZ, AIB, BTBT, WYFI, ALP, SUPX, TSSI, QMLS, VIP, BIRD have no revenue to divide by, so they are shown but left out of the peer median.
  • APLD, IREN, CORZ, WULF, GLXY shown for context only — not close enough to move the median.
The 18 listed companies it is measured against, and why
  • CRWV13.79× revenue

    CoreWeave is the pure-play GPU-as-a-Service neocloud and the closest business-model match - reserved and on-demand GPU compute for LLM training and inference. Caveat: CoreWeave's revenue is roughly three orders of magnitude larger, so it is a model peer, not a scale peer.

  • AZIOno revenue multiple

    Direct comp: Auto, Truck & Motorcycle Parts (NEC); micro-cap ($2m); shares gpu, compute, density, infrastructure, data, modular with the target's own description; no forward EV/Sales published — counted as a peer, excluded from the median.

  • NBIS53.18× revenue

    Nebius Group is an AI-cloud/GPU-compute provider selling the same reserved and on-demand GPU capacity to AI workloads; identical revenue model, again dramatically larger scale.

  • SHAZno revenue multiple

    Operational comp: IT Services & Consulting (NEC); micro-cap ($23m); shares gpu, compute, data, infrastructure, storage, high with the target's own description; forward EV/Sales 10.7x.

  • APLD23.65× revenuecontext only — left out of the median

    Applied Digital sells HPC/AI data centre capacity and GPU cloud services, the nearest listed name in the AI-infrastructure mid-cap bucket - but it OWNS and builds its data centres, the opposite of Exascale's asset-light no-hardware model.

  • AIBno revenue multiple

    Operational comp: IT Services & Consulting (NEC); shares compute, density, data, centers, infrastructure, high with the target's own description; forward EV/Sales 11.5x.

  • IREN23.44× revenuecontext only — left out of the median

    IREN operates AI cloud services on GPU fleets alongside bitcoin mining; relevant as an AI-compute revenue comparable but it owns its own power and data centre infrastructure, so the cost structure and capital intensity are not comparable.

  • BTBTno revenue multiple

    Operational comp: Cryptocurrency Mining; small-cap ($612m); shares gpu, cooling, data, centers, liquid, asset with the target's own description; forward EV/Sales 6.9x.

  • CORZ20.43× revenuecontext only — left out of the median

    Core Scientific provides colocation and hosting for HPC/AI compute; a valuation reference for AI compute capacity but asset-heavy and far larger, with a fundamentally different margin profile.

  • WYFIno revenue multiple

    Operational comp: IT Services & Consulting (NEC); small-cap ($605m); shares gpu, data, cooling, infrastructure, centers, high with the target's own description; forward EV/Sales 8.2x.

  • WULF65.11× revenuecontext only — left out of the median

    TeraWulf is converting mining sites into AI/HPC hosting capacity; useful as an AI-infrastructure sentiment comparable only - it is an infrastructure owner, not an asset-light compute reseller.

  • ALPno revenue multiple

    Operational comp: IT Services & Consulting (NEC); micro-cap ($7m); shares gpu, compute, infrastructure, data, hardware, service with the target's own description; no forward EV/Sales published — counted as a peer, excluded from the median.

  • GLXY0.15× revenuecontext only — left out of the median

    Galaxy Digital's Helios data centre business is a listed AI-compute landlord comparable; included for completeness but the business mix (digital asset trading and asset management) makes it a poor read-across.

  • SUPXno revenue multiple

    Operational comp: IT Services & Consulting (NEC); micro-cap ($224m); shares density, cooling, high, infrastructure, singapore, liquid with the target's own description; no forward EV/Sales published — counted as a peer, excluded from the median.

  • TSSIno revenue multiple

    Operational comp: IT Services & Consulting (NEC); micro-cap ($204m); shares data, modular, centers, hardware, infrastructure, party with the target's own description; forward EV/Sales 1.1x.

  • QMLSno revenue multiple

    Operational comp: IT Services & Consulting (NEC); shares gpu, compute, infrastructure, power, small, software with the target's own description; no forward EV/Sales published — counted as a peer, excluded from the median.

  • VIPno revenue multiple

    Operational comp: IT Services & Consulting (NEC); micro-cap ($23m); shares infrastructure, data, power, high, centers, performance with the target's own description; no forward EV/Sales published — counted as a peer, excluded from the median.

  • BIRDno revenue multiple

    Operational comp: IT Services & Consulting (NEC); micro-cap ($25m); shares cluster, infrastructure, hardware, performance, designs, not with the target's own description; no forward EV/Sales published — counted as a peer, excluded from the median.

Which companies count as comparable is our judgement, written out above so you can disagree with it. The median is what these shares happened to trade at on the date given — not a price anyone is offering for this deal.


In plain English

No floor / floorlessthe cash guarantee is gone — the price is unprotected

A SPAC's downside protection is not the cash in trust; it is your right to demand that cash. Once the redemption window closes, the cash stays with the company and the share can trade anywhere.

Dilutionhow much of the company new shares take from you

Sponsor promote, PIPE shares, warrants and rights all issue stock that did not pay $10 for it. The headline deal value is before that; the effective value is after.

Pro-forma equitywhat the company is valued at once the deal closes

The combined company's equity value assuming the announced terms and the redemptions that have actually happened.

De-SPACthe day the SPAC becomes the real company

The shares stop being a claim on a pot of cash and start being equity in an operating business. Roughly 80% of recent de-SPACs traded below $10 within a year.

Outside datethe contractual long-stop for closing the deal

A deadline between the SPAC and its target, not between the SPAC and you. It confers no right to cash, which is exactly why it must never be counted as a redemption window.