Arqis
Solutions ▾
Neoclouds & GPU cloudsGPU supply and customer offtake Data-centre & AI infrastructure operatorsLeases, power and tenant concentration OEM & hardware suppliersDelivery, allocation and warranty exposure Banks, funds & private creditGPU-backed financing and covenants Infrastructure funds & investorsCross-border AI infrastructure capital Insurers & reinsurersStructured exposure and accumulation AI labs & compute buyersFrontier training and inference transactions
How Arqis is used Risk coverage Pricing Extensions Security
Client login
Risk intelligence & underwriting

The risk layer
for AI infrastructure

Arqis is the risk intelligence and underwriting layer for AI infrastructure. We help AI infrastructure operators, lenders, investors, insurers and strategic counterparties understand, structure and transfer the new economic risks that are not well covered by traditional insurance today.

See a worked example
Running with five design partners across Asia and the GCC today.
Contract workspace
Drop a contract to underwrite it
GPU_Supply_Agreement.pdf, extracting terms…
ContractCounterpartyValueExposureStatus
GPU supply agreementNeocloud, US$500.0m$12.4mJust added
Compute leaseData centre op., SA$84.0m$8.1mMonitored
Colocation MSACloud operator, UAE$61.2m$2.3mMonitored
Transformer supplyOEM, delivery slipped$14.5m$8.4mRe-rated
ONE RISK RECORD PER CONTRACT, PRICED IN DOLLARS
How Arqis is used

Three ways in. One risk record.

One platform, one risk methodology, one audit trail. Only the entry point changes: most customers start with the contracts they are most worried about, then extend to the rest of the book.

01 · NEW TRANSACTION

Should we do this transaction, and on what terms?

In origination, structuring or approval. Not yet committed.

What you get
  • An approval-ready decision memo
  • Quantified $ exposure per risk
  • Ranked negotiation positions
  • Obligations captured from day one
Measured on cycle time to approval, and adviser hours per transaction.
02 · EXISTING TRANSACTION

What did we actually agree to, and where are we exposed?

Signed under time pressure, inherited, or never diligenced to this standard.

What you get
  • A remediation plan per contract
  • Recoverable exposure, ranked
  • The lever still available to you
  • An obligation register you can run
Measured on exposure closed against identified, and value recovered at amendment.
03 · FACTS CHANGED

Which live transactions just got worse, and what do we do?

Credit, ownership, sanctions, power, delivery, hardware or regulation moves.

What you get
  • Alerts scoped to affected transactions
  • Revised exposure within hours
  • The right that applies, and its deadline
  • Remediation routed to an owner
Measured on time from change to impact assessed, and actions taken inside the window.
How a transaction moves through Arqis
1 · Transaction enters ArqisCustomer, contract, GPU config, location, financing assumptions
2 · Modules runOfftaker · deployment · residual value · utilization · contract
3 · Risk PassportProceed / conditions / escalate / decline with quantified exposures
4 · Risk routingContract, insurance, financing, OEM support, structured capital
5 · MonitoringTriggers, covenants, credit changes, regulatory and asset-value alerts
Who we work with
Transaction counsel and legal opsDiligence and negotiation positions
Risk and complianceThe register and the alerts
FinanceExposure priced in dollars and MW
Asset and portfolio teamsObligations and deadlines
A change on a live transaction re-enters diligence automatically
Worked examples

The rating was never the operator’s. Watch it move.

Two contracts in an illustrative book trace to the same ultimate parent. The parent is downgraded two notches and the guarantee test fails.

Offtaker / customer risk
CoreWeave’s $8.5bn DDTL 4.0 facility was rated A3 by Moody’s and A (low) by DBRS. What made it investment grade was the creditworthiness of the customer on the other side of the contract, not the operator, whose own profile is speculative grade.
Public trigger: reported ratings on CoreWeave’s DDTL 4.0 facility; roughly $35bn of operator debt at Q2 2026 · portfolio below is illustrative
Live book · 4 contracts
ContractCounterpartyValueExposureStatus
GPU supply agreementNeocloud, US$500.0m$12.4mMonitored
Compute leaseData centre op., SA$84.0m$8.1mMonitored
Colocation MSACloud operator, UAE$61.2m$2.3mMonitored
Transformer supplyOEM, Korea$14.5m$1.2mMonitored
PORTFOLIO EXPOSURE $24.0m THREE OF FOUR CONTRACTS DEPEND ON THAT DATE
Risk coverage

Eight risks traditional insurance was not built for.

Eight risks sit inside every transaction. Offtaker / customer, compute utilization, GPU residual value and regulatory deployment risk are covered today, on one graph, with the other four built on the same record so nothing is re-entered as they arrive.

01
Offtaker / Customer Risk
Covered today
Counterparty and demand

A customer weakens, defaults, downsizes or fails to renew after infrastructure has been committed around them.

LeverTake-or-pay · parent guarantee · collateral · exposure cap · credit insurance
02
Compute Utilization / Demand Risk
Covered today
Utilization and stranded capacity

You build the GPU or MW capacity, but actual demand and utilization fall below the level required to support the investment case.

LeverCapacity reservation · utilization floor · anchor offtake · minimum revenue commitment · phased deployment
03
GPU Residual Value / Obsolescence
Covered today
Asset value

The hardware still works, but newer chips or architectures reduce its earning power, resale value or useful economic life faster than expected.

LeverOEM buyback · residual-value guarantee · leasing · refresh rights · asset-value protection
04
Compute Price / Revenue Risk
Market and revenue

GPU-hour, token or compute pricing declines faster than expected, compressing revenue even when utilization remains high.

LeverFixed-price offtake · price floor · collar · long-term reservation · revenue hedge
05
Regulatory Deployment / Stranding Risk
Covered today
Regulatory and geopolitical

Export controls, sanctions or AI rules change after deployment and restrict which chips, customers, workloads or jurisdictions can be served.

LeverChange-in-law · relocation rights · customer representations · termination compensation · regulatory cover
06
Contract Bankability Risk
Contract and financing

A large headline contract provides less real protection than expected because termination, volume, pricing or guarantee provisions weaken the revenue.

LeverTake-or-pay · termination payment · parent guarantee · step-in rights · tighter covenants
07
Technology Substitution Risk
Technology

A new GPU, accelerator architecture or software stack makes existing infrastructure economically uncompetitive before the end of its expected life.

LeverUpgrade rights · OEM refresh commitment · buyback · shorter asset tenor · technology diversification
08
AI Efficiency / Compute Compression Risk
Demand and technology

Models and inference software become much more efficient, so the same AI workload requires significantly less compute than originally assumed.

LeverShorter commitments · volume floors · workload diversification · flexible capacity · repricing rights
One counterparty across nine transactions is one exposure, not nine
See each module in full →
Who it's for

Built for the parties that build, sell, finance, insure and buy AI infrastructure.

The eight risks are the same everywhere. What differs is which one breaks your book first, and that is where each brief starts.

Neoclouds & GPU clouds

Your obligations are contracted for five years. Your revenue isn't.

Duration mismatch between long GPU and power commitments and much shorter customer contracts, priced in dollars and GPU-hours.

Read the brief →
Data-centre & AI infrastructure operators

You underwrite a fifteen-year lease in six weeks, then never look again.

Tenant credit, grid access, EPC delivery and regulation, read as one graph instead of five disconnected systems.

Read the brief →
OEM & hardware suppliers

You promised delivery on a hardware roadmap that resets before the order ships.

Allocation commitments, lead-time risk and warranty exposure on the orders that sit upstream of every AI infrastructure transaction.

Read the brief →
Banks, funds & private credit

The collateral behind your GPU financing depreciates faster than your covenant package assumes.

Borrowing-base tests, covenant resets and residual-value risk on GPU, data-centre and compute-offtake financing.

Read the brief →
Infrastructure funds & investors

You're underwriting a $2B AI infrastructure platform across three jurisdictions with one diligence team.

Counterparty, delivery and regulatory risk on cross-border AI infrastructure capital, kept current for the life of the hold, not just at close.

Read the brief →
Insurers & reinsurers

The insurer cannot price what it cannot see.

Structured exposure and evidence an underwriter accepts, with accumulation across tenants, OEMs and grids made visible before it is written.

Read the brief →
AI labs & compute buyers

You committed to a five-year training cluster on a roadmap that won't survive the next model generation.

The buy-side view: screen the supplier, price the commitment and structure the agreement before you commit capital to it.

Read the brief →
The business case

The return is in the contract, not the licence.

Whose budget it comes from, what it displaces, and what we need from you to prove it on your own book in the first thirty days.

Whose budget
Legal and risk operating spend, alongside the $3–5M a year that typically goes to external counsel and advisers on this work
What it displaces
$0.6–2.0M of repeatable review work not sent out again, and 1,500–3,000 internal review and coordination hours released
First 30 days
Your signed book ingested read-only, one module scoped, and a remediation plan per contract with exposure ranked
What we ask of you
Executed contracts, an access list, and one named owner for remediation. No integration work before the pilot
Scoped against your own book in the pilot, on your numbers rather than ours
Next step

Fourteen days from contract access. One number at the end.

Run Arqis alongside your current process: one new transaction, one signed contract set, or your live portfolio under monitoring. You baseline it, we run it, and you compare. The fee is at risk against a number you set on day one.

Day 1 · BaselineYour current hours, cycle time, adviser cost and known risks. You set the bar the pilot has to clear
Days 2–10 · RunDiligence, mitigate, negotiate and monitor inside Arqis
Days 11–13 · CompareRisks found that you did not have, terms improved, hours released, each one traced to the clause behind it
Day 14 · DecideOne dollar figure for value found, against one price. Miss the bar you set and the fee is refunded in full