For energy, infrastructure & deeptech investors
We find the critical assumptions your deal depends on and test whether they hold.
By Eden Djanashvili · 10+ years in deeptech · EIC jury expert
1,920 MW nuclear PPA to 2042 · public documents · evidence to 9 Oct 2026
Why now
Where an investor has its own benchmarks, it knows whether a number is reasonable. Where it doesn't (a new market, a technology with no precedent, a counterparty it hasn't worked with), the number goes into the model as presented.
Regulators are separating connection from firm power, and the rules on whose load is cut first in a shortage are still being written.
In one case we checked, the grant covers 75% of eligible costs, paid against invoices over seven years, with tens of millions of capex excluded.
The questions that decide value are the foundry, the qualification partner and the competitor's schedule.
Eight kinds of assumption, one method: grid and interconnection · approvals · first-of-a-kind build · capex · funding · price and revenue · capacity and performance · market window. Each one rated supported, look again, challenge or unsupported, with the evidence behind it.
How it works
You keep the decision and the valuation. DeepRadar gives you a defensible view on the assumptions underneath them.
Under NDA, read-only access to the model, the memo and the documents behind each assumption. You name the decision and the date.
Every material assumption tested against your documents and public filings. No citation, no flag.
The lines that change the decision, the one question for management, the path to the goal, and a verdict, signed by Eden Djanashvili.
Questions go to management. Conditions you can put in the term sheet.
Did the lead actually test it? Questions go to the lead, inside a short co-investment window.
Worked cases
Client work is confidential and never published. These show the method on cases anyone can verify.
The Talen × AWS check, laid out exactly as a client receives it.
Energy · PJM
1,920 MW nuclear PPA to 2042
Deeptech · EU state aid
Public grant · first production line
Who it's for
Entering a new market or asset type (a new grid region, data-centre power, storage), or screening a deal before committing to full diligence.
Direct deals and co-investments, often with no in-house technical team and a short decision window, relying on the lead investor's diligence.
Series B and later, where the valuation depends on a first production line, a scale-up or public funding that hasn't been delivered yet.
Pricing
Free shows you what a document proves. The Read tells you whether the case holds.
A first pass on one public document: every assumption found and rated.
For when you have a 40-page filing and no time to find what matters.
Five days
On one live decision: a new investment, a co-investment, a follow-on or a major capex.
For when part of your case rests on assumptions you can't benchmark.
Your documents stay in your data room and never appear anywhere else.
Where DeepRadar fits
Your advisors each check their part. DeepRadar checks the assumptions the valuation depends on, before the investment decision.
| Compared on | DeepRadar Read | Legal due diligence | Technical due diligence (independent engineer) | Commercial due diligence |
|---|---|---|---|---|
| Question it answers | Will the few assumptions the valuation depends on actually happen, at the volume and on the date in the model? | Are the contracts valid and enforceable? | Is the asset sound: design, construction, capacity, schedule, yield? | Is the market and demand case realistic? |
| Timing | Any time before the decision, including before full diligence is commissioned. Five days. | Usually once the deal is advanced | Usually once the deal is advanced, often required by lenders | Usually once the deal is advanced |
| Output | A dated verdict: commit, condition, reprice or pass, with the evidence behind it | Legal report | Technical report, often accepted by lenders | Market report |
| Stronger than DeepRadar on | Legal opinion and liability | Engineering depth, site visits, lender acceptance | Market depth, customer interviews | |
| Scope | Across technical, commercial and regulatory, limited to the assumptions that move the valuation | Contracts, corporate, regulatory compliance | Engineering and site | Market, customers, competitors |
| Sources | Your documents and public filings: grid operators, regulators, utilities and the companies themselves | Contracts and corporate records | Site, design and engineering data | Market data, customer interviews |
Not a legal opinion, not an engineering report, not a report lenders will accept.
When the deal is already delivering: an operating asset with a track record and nothing new left to happen. DeepRadar is for deals valued on something that hasn’t happened yet.
Who you work with
Founder of DeepRadar, with more than ten years in deeptech. Tests whether the assumptions a deal's valuation depends on will actually happen. Independent: no position in the companies or deals reviewed.
Send the assumption you're least sure about. In 30 minutes we'll know whether a check would change anything, and if it wouldn't, you'll hear that too.
Your sample is opening. If it didn't, use the link below.
Open the sample ReadHave a deal in mind? Book a 30-min call.
