// 18 projects · 11 active · 4 sold · 3 closed · luxembourg · since 2012

Portfolio.

STATUS SECTOR showing 18 of 18
2022 · WEB3 · LUXFEATUREDEXITED
wagmi web3 Fund I SCSp
Luxembourg-domiciled AIF. On-chain, self-custodian, ETH-heavy thesis. Ran 35 months · +73.31% total return · 21.10% CAGR · wound down at ATH ahead of MiCA. The bet that named the holding company — and paid for the next one.
WEB3 · EXITED+73.31%
2026 · NUCLEAR · AI INFRACS · 02 →ACTIVE
Alpha Molten One SCSp
NGMI is an investor in Alpha Tech Research Corp — molten-salt thorium SMR for AI data centers. Walk-away safe by physics. $75M round · $400–500M pre · base case 587–734× over 10y. The forward shot: AI power demand is 3× by 2030 and the grid wasn't built for it.
INVESTMENT · NUCLEAR SMRQ1.2026 · INVESTOR
2018 · REAL ESTATEACTIVE
CoinFlip Properties
Personal real estate brand. Identifying and developing investment opportunities in Luxembourg.
REAL ESTATEinvestor
1950 · RESTAURANTACTIVE
Ristorante Roma
Est. 1950 — the first Italian restaurant in Luxembourg. Recognized by Michelin & Gault&Millau. Took over in 2024 to continue a historic family-style institution.
HOSPITALITY · ACQ. 2024operator
2023 · FITNESSACTIVE
Blocx Urban Fitness
Indoor movement & leisure center in Esch-sur-Alzette — Luxembourg’s largest bouldering gym. Bouldering, Ninja Warrior, parkour, calisthenics, fitness, plus kids’ area and on-site restaurant.
FITNESSoperating
2023 · RESTAURANTACTIVE
Los Monos
Mexican street-food restaurant inside BlocX, Esch-sur-Alzette. Bold Latin American flavors in a relaxed, modern setting.
HOSPITALITYoperating
2017 · RESTAURANTACTIVE
Pavillon Merl
Long-standing Luxembourg City restaurant in the Parc de Merl. Elegant terrace, relaxed bistro atmosphere, scenic park setting.
HOSPITALITYoperating
2017 · RESTAURANTACTIVE
Snooze Belval
Vibrant sports bar & burger destination in Esch-Belval. Bold flavors, extensive beer selection, energetic vibe.
HOSPITALITYoperating
2020 · COMMERCEACTIVE
Nerds.store
Custom merch & branded apparel studio for Luxembourg businesses. Sustainable, no-MOQ, print & embroidery in-house in Belvaux. Investor only.
RETAILinvestor
2019 · NIGHTLIFEACTIVE
LOA Festival
Luxembourg Open Air — the country’s premier electronic music festival. Two editions/year across Belval & Luxembourg City. 15,000+ attendees, international DJs, zero single-use plastic. Investor only.
NIGHTLIFEinvestor
2012 · AGENCYACTIVE
V&B Luxembourg
Sole agent for Villeroy & Boch Hospitality in Luxembourg. Every branded tableware sale flows through here.
COMMERCE14Y
2019 · APPARELACTIVE
Putain.
Own small clothing brand. A personal stake.
RETAILoperating
2017 · RESTAURANTCLOSED
Bronx Dogs & Burgers
Burger joint. Closed.
HOSPITALITYCLOSED
2016 · NIGHTLIFECLOSED
Epic Night Club
Club. Ran the cycle.
NIGHTLIFECLOSED
2015 → 2024 · RESTAURANTSOLD
Snooze City
Opened 2015. Sold in 2024 after nine years operating. Still one of the calmer exits.
HOSPITALITYSOLD ·2024
2017 · CAFÉSOLD
Gruppetto Café
Café concept. Sold.
HOSPITALITYSOLD
2015 · DELIVERYSOLD
Coffee Rocket
Coffee delivery. Sold.
HOSPITALITYSOLD
2012 · FOOD DELIVERYSOLD
Livraison.lu
Food delivery marketplace. Early bet. Exited to Wedely — now the market competitor.
FOOD DELIVERYEXIT → WEDELY
← back to grid  ·  ngmi/portfolio/wagmi-gp

wagmi web3 Fund I · SCSp

Luxembourg-domiciled AIF. Sep 2022 → Aug 2025. ETH-heavy, self-custodian, on-chain thesis across DeFi, NFTs, and metaverse primitives. Wound down profitably at ATH ahead of MiCA. +73.31% total · 21.10% CAGR · -22.60% max DD.
CASE STUDY · 01
launched
26 Sep 2022
category
Web3 AIF
role
Co-GP
horizon
35 months
outcome
Exited +73.31%
status
wound down

// 01 — THE THESIS

September 2022. The NFT cycle had peaked, the macro had turned, and most web3 funds were pivoting to "AI" or quietly closing. We launched anyway.

The thesis was specific: on-chain, self-custodian, ETH-heavy. No leverage, no illiquid bags we couldn't exit, no "we'll figure out custody later." A Luxembourg SCSp AIF structure — wagmi web3 Fund I — so LPs got real reporting and we got real governance. Six partners. Six signatures on every material decision.

// 02 — WHAT WE BUILT

A fund that actually held what it said it held. Every position on-chain, every wallet verifiable, NAV reconciled to the block. Exposure concentrated in ETH and quality L1/L2 infrastructure, with small sleeves for DeFi bluechips, select NFTs, and metaverse land.

The best call in the book: overweighting ETH and refusing to rotate into BTC when the consensus trade was "BTC dominance up." Captured most of the +60% Q2 upside on the right side of that call.

// 03 — HOW IT PERFORMED

Three years. $100 → $173.30 per unit. +73.31% total return. 21.10% CAGR. Max drawdown of -22.60% during the 2023 grind — painful but survivable, which was the whole point of the structure.

AUM at close: $1.64M. Not a generational fund. But every LP got back more than they put in, audited and on-chain, through a cycle that killed most of our peers. That's the bar.

// 04 — WHY WE WOUND IT DOWN

Summer 2025. Prices at new highs. MiCA compliance overhead about to change the math for a small AIF like ours. The rational move was to return capital at the top and build the next vehicle clean, rather than grind through a regulatory transition that would've eaten returns.

Closed by choice, at the high. A successor vehicle — post-MiCA, scaled structure — is the forward-looking work. WAGMI I did what it was supposed to do.

// NAV PER UNIT SEP 2022 → AUG 2025 · $100 → $173.30

$100 basis Q4.22 Q2.23 Q4.23 Q2.24 Q4.24 Q2.25 $70 $100 $130 $160 $180 LAUNCH · $100 DD -22.60% ETH OVERWEIGHT +60% EXIT · $173.30

// TIMELINE

2022 · Q1
LAUNCH
Project kicks off · brand, community, first events
2022 · Q3
PEAK
Highest engagement window · real IRL presence built
↑↑
2023 · Q1
PLATEAU
Growth flattens · we commit more effort, not less
2023 · Q3
DECLINE
Market appetite gone · we keep going anyway
2024
CLOSE
Wind down cleanly · write the memo · archive the brand
"We kept it open for a year longer than we should have,
because closing felt like admitting the joke was real.
It was. That was the lesson."
— NGMI · INTERNAL POST-MORTEM
// what we got right
Built a real community, ran real events, didn't hide behind the token. When it worked, it was because of the operating thesis, not the speculation.
// what we got wrong
Confused commitment with denial. We held on a year past the point of return because closing would've felt like losing. The story is the same either way — closing it earlier would've cost less.
// what we'd repeat
Taking the shot. The holding company that exists today is downstream of being willing to build something that might not make it. Name it, fund it, try it, close it, write it up, go again.
← back to grid  ·  ngmi/portfolio/alpha-molten-one  ·  ACTIVE DEAL · Q1 2026

Alpha Molten One · Nuclear AI Infra

NGMI is an investor in Alpha Tech Research Corp — a molten-salt thorium SMR platform purpose-built to baseload AI data centers. Walk-away safe by physics. $75M round · $400–500M pre. Base case 587–734× over 10y. This is the shot.
CASE STUDY · 02
stage
Series A
sector
Nuclear · SMR · MSR
role
Investor
horizon
10 years
target mult.
587–734×
status
live · diligence

// 01 — THE MACRO SHOT

AI compute demand is on track to 3× US data-center power draw by 2030. The grid was not built for this, and nobody is building enough grid fast enough. Natural gas turbines have multi-year lead times. Utility-scale solar + storage can't do 24/7 baseload without multiplying the capex.

Nuclear is the only physics-true answer: high density, low land, zero carbon, baseload. But legacy LWRs are 10-year permits and $10B+ overnight costs. The window opens for a new form factor — small, modular, inherently safe, factory-built — aimed directly at the hyperscaler load curve.

This is not a policy bet. This is a physics and supply-chain bet with a policy tailwind.

// 02 — WHAT ATRC IS BUILDING

A molten-salt reactor using thorium fuel, designed walk-away safe: fuel is liquid at operating temp, drain plug fails passively, no high-pressure vessel, no water-cooling catastrophe path. The physics cannot run away.

Target unit: ~100 MWe modules, factory-assembled, truck-shippable, sited at or adjacent to hyperscale campuses. Co-located heat offtake for direct-to-chip cooling and on-site hydrogen. One module anchors a ~75 MW AI cluster; stack modules for gigawatt campuses.

// 03 — WHY NOW, WHY US

Three things aligned: NRC Part 53 (risk-informed licensing for advanced reactors) is finally workable. DoE HALEU supply plus thorium breeder economics make fuel accessible. And the hyperscalers are actively signing PPAs — Microsoft, Amazon, Google have all telegraphed nuclear offtake in the last 12 months. The demand is not hypothetical.

NGMI is in this round because we bring operator DNA to an infra thesis — 14 years of building capital-intensive physical businesses that have to actually work when customers show up. We're not index-buying the SMR basket. We're picking one team, one fuel cycle, one regulatory path.

// 04 — THE DEAL

$75M Series A at $400–500M pre-money. NGMI participates as an investor in the round. Proceeds: core team build, NRC pre-app, HALEU + thorium fuel sourcing, first-module fabrication partnership, anchor-customer LOI pipeline.

Exit path is strategic acquisition by a hyperscaler, IPO on commercial operations, or continued private rounds to $20B+ valuation by first grid-tie. Base case is 587–734× on the entry round over a 10-year horizon. That's the math if the team executes.

BEAR
// regulatory slip / fuel delay
8–15×
Module demo works. Licensing pushes 2–3y. Exit: acquihire by utility or strategic. Check returns, thesis proven.
BASE
// first commercial unit by '33
587–734×
First grid-tie by 2033. 3–5 modules in pipeline with hyperscaler PPAs. IPO or strategic exit at $20–30B.
BULL
// thorium fuel cycle wins
1500×+
Platform becomes the default AI-infra baseload standard. Thorium supply advantage locks in. $100B+ category leader.

// US DATA CENTER POWER DEMAND 2024 → 2035 · GW · SOURCE: DOE / EPRI / NGMI

'24 '26 '28 '30 '32 '34 '35 0 100 GW 200 GW 300 GW TOTAL · 300 GW CONVENTIONAL SUPPLY THE GAP · ~180 GW SMR/MSR addressable 2024 · 35 GW 2030 · 3× INFLECTION 2035 · 300 GW

// CATALYST PATH · Q1 2026 → FIRST GRID TIE

Q1 · 2026
CLOSE
Series A closes · $75M in bank · NGMI participates
now
Q3 · 2026
TEAM
Chief Nuclear Officer hired · NRC pre-app team staffed · fuel MoUs signed
+6m
Q2 · 2027
REG
NRC Part 53 pre-application submitted · first hyperscaler LOI
+12m
 
 
 
 
2028
SERIES B
$200M+ at $1.5–2.5B · markup event · NGMI partial DPI window
+24m
2030
DEMO
Sub-critical test assembly · integrated thermal-hydraulic demo
+48m
2033
FIRST UNIT
First commercial module grid-tied · anchor hyperscaler PPA live · re-rate to $20B+
+84m
2035+
EXIT
IPO or strategic · base case realization · 587–734× return
+108m
// what can kill it
  • REG NRC certification slips past 2032 · fatal to IRR, not to thesis
  • FUEL HALEU / thorium supply chain fails to scale · dependency on DoE
  • CAPITAL nuclear capex markets freeze in a recession · bridge risk
  • COMPETITION X-energy / Oklo / Kairos wins a standard before ATRC ships
  • EXECUTION nuclear is hard · every SMR to date has slipped
// why we still lead
  • PHYSICS MSR + thorium is the only walk-away-safe baseload path
  • DEMAND hyperscalers already signing · not a policy bet
  • TEAM nuclear + hyperscaler operator pedigree, not pure academics
  • TIMING NRC Part 53 + DoE HALEU + PPA wave · 3 tailwinds at once
  • ASYMMETRY bear case still returns capital · bull is category-defining · we just need to be in the round
"Everything else is a hedge against AI power demand.
This one is the thing that meets the demand.
If we're right, it's the fund maker."
— NGMI · INVESTMENT MEMO · Q1 2026
// entry discipline
We are not buying the SMR basket. One team, one fuel cycle, one reg path. Conviction + concentration over optionality.
// downside we accept
10-year lockup, reg slip risk, potential re-ups through Series B/C to protect ownership. This is a patient-capital shot, not a trade.
// what winning looks like
First NGMI bet that's structurally category-defining rather than operationally excellent. The build-anyway thesis applied to infrastructure that matters.