The Agentic and Data Infrastructure Powering Luxury Travel
Scroll
CONFIDENTIAL
01 Origin Story
We ran luxury travel manually first. On purpose.
QT
2021 to 2025
Quantum Temple
Regenerative travel experience company. Government partnerships across Indonesia, Egypt, Morocco, Panama, and UAE. Institutional credibility that capital alone cannot replicate.
The Inflection
Clients Asked for Everything
UHNW clients wanted full travel management. We spent 2025 manually validating concierge workflows across FIT, incentive, and corporate segments.
◆ Government & UNESCO◆ IATA & Virtuoso◆ Direct Hotel Contracts◆ DMC Network · 12+ Countries◆ 36K Waitlist · $0 Marketing
02 Traction
Zero tech. Zero marketing. All manual.
Delivered by Quantum Temple, by hand. Workflows and unit economics validated before a single line of product code.
$5.74M
FY2025 Revenue
642%
YoY Growth
~$250K
Monthly Run Rate
36K
Waitlist
$0
Paid Marketing
Revenue Composition
◆ One whale family office: ~$250K/month commission ◆ Organic UHNW pipeline: $150K to $200K/month ◆ All manual concierge, deliberate workflow validation ◆ Zero paid acquisition, pure QT ecosystem + referrals
What This Proves
Willingness to pay at extreme price points ($120K+ ACV) Workflows validated across FIT, incentive, corporate Supplier infrastructure live and operational Client trust compounds. Scope expands organically
03 The Problem
Siloed inventory is breaking operations and customer experience.
Luxury travel has no unified data layer. Inventory lives across disconnected systems, PDFs, and human memory. Every proposal starts from scratch.
The Client
Net Worth$10M–$5B+
Annual Trips8–20
Annual Spend$250K–$2M+
Concierge Cost$48K–$150K/yr
40-72h
Per Proposal
Hours of manual research across disconnected systems. Every proposal starts from scratch.
0
Personalization
All time goes to sourcing and booking. None left for tailoring the experience to who the client actually is.
Status: Live in alpha with investors and selected clients. Founding member onboarding Q1-Q2 2026. Personalization and memory layers shipping Q2-Q3.
2,000+ Luxury Hotels · Direct Partnerships & Exclusive Benefits
05b JMPR
The first MCP server for luxury travel.
Luxury travel, accessible by any AI agent. JMPR opens Obelisk inventory to the agentic ecosystem through the Model Context Protocol.
Every AI Agent = A Sales Channel
MCP-native. Any AI assistant can search and book our inventory. Zero integration cost per channel.
100x the Addressable Market
Affluent travelers ($10K–$100K/yr) who want curated luxury without a concierge price tag.
Pay Per Request, Not Per Seat
USDC micropayments + commission on bookings. No subscriptions. Revenue scales with usage.
Same Engine, Broader Reach
Powered by Obelisk — same curated inventory and booking rails as KARNAK, different interface.
09b Agentic Commerce
Payments built for agents. Settled in stablecoins.
Two payment layers: programmatic agent payments via MPP, and USDC credit balances for client settlement.
Agentic Payments
Powered by MPP (Machine Payments Protocol) · Stripe & Tempo
Every interaction between an AI agent and our platform is a paid transaction. Inventory lookups, matching requests, booking confirmations — each settles programmatically via HTTP 402. No accounts, no subscriptions, no human in the loop.
→Booking fees — $15–$20 per confirmed booking through JMPR
→Commission layer — percentage on GMV for high-value transactions
→Stripe-grade compliance — tax, fraud, reporting, refunds built in
Client Settlement Layer
USDC Stablecoin · Pre-funded credit balance
Clients deposit $100K–$500K in USDC as a pre-funded credit balance. We draw against it per booking — jets, hotels, experiences. Value locked at time of deposit. No wire delays, no FX friction, no 2–3% card fees on six-figure sums.
→On-chain transparency — every spend has a hash ID linked to a specific booking
→Client-visible wallet — full provenance trail from deposit to each transaction
→The future of invoicing — blockchain-native receipts replace PDFs and wire confirmations
Why this matters: Tempo raised $500M at $5B valuation to build MPP. Coinbase’s x402 processed 100M+ agent transactions in 2025. We’re not building payment rails — we’re the first luxury vertical to plug into them. Every AI agent becomes a paying customer, every client gets full on-chain transparency.
06 Obelisk
Our proprietary data engine. Deep profiling meets perfect matching.
Obelisk is our proprietary intelligence layer. We deeply profile every client and every supplier across all verticals, then match them with narrative precision.
DEMAND INTELLIGENCE
Preferences
Room type, dietary, family, budget
Behavioral Patterns
Booking cadence, spend, lead time
Experience History
Loved, repeated, avoided
Life Context
Anniversaries, companions, seasonal
MATCHING
ENGINE
Obelisk Core
OUTPUT
3 perfect options with narrative
SUPPLY INTELLIGENCE
Property Attributes
Specific suites, views, renovation year
Service Reality
Genuinely child-friendly or just says so?
Negotiation Dynamics
Rate flexibility, VIP thresholds, upgrade culture
Expert Network Intel
Curated intel via our on-the-ground experts network
Structured data is commodity. OTAs have had it for 30 years. The unstructured layer, the qualitative signals that tell you which suite has the best view, which chef will do a private dinner, which property is actually worth the rate, that’s where 90% of the value lives. And no one else has it.
07 Why Now
Three structural shifts just converged. Travel hasn't caught up.
A once-in-a-decade window to build the infrastructure layer for luxury travel.
Shift 1
AI Can Now Book End-to-End
Agents orchestrate flights, hotels, dining, and transport in one chain. Production-grade reliability hit in early 2026.
Shift 2
MCP Turns Every Agent Into a Channel
One protocol, infinite distribution. Our MCP server makes luxury inventory available to every AI assistant.
Shift 3
No One Owns the Data Layer
GDS covers flights and hotels. Dining, experiences, VIP access, ground transport — no structured layer exists. First to build it wins.
Our edge: 4 years of relationships and ops knowledge AI-native companies can't replicate. Now agentic AI lets us scale what we already proved manually.
08 Market
Three markets. One platform.
KARNAK · UHNW Concierge
$30B
Serviceable Addressable Market
~395K UHNW individuals globally spending $75K+ annually on managed travel. Currently served by fragmented concierge services and personal EAs with no tech.
JMPR · Premium Mass Travel
$200B
Serviceable Addressable Market
Affluent travelers spending $10K to $100K per year who want AI-powered curation without the concierge price tag. 100x the KARNAK addressable base.
Obelisk · B2B Travel Tech
$12B
Serviceable Addressable Market
Traditional travel advisors and agencies with deep luxury networks but no AI tools, data infrastructure, or technical capability. They have the clients. We have the platform.
Conservative 0.1% capture by Year 3 = ~400 members × $130K blended ARPU = $52M annual revenue
Year 3 Target
$52M
09 Business Model
Three revenue engines. Two monetization layers.
KARNAK and JMPR monetize membership + transactions. Obelisk captures transactional revenue from B2B partners.
KARNAK
B2C · UHNW
Membership: $12K to $180K/yr
Commission: ~15% on GMV
Virtuoso preferred rates
SOVEREIGN
$12K
$1K/mo
WHALE
$120K
$10K/mo
CORPORATE
$180K
$15K/mo · Unlimited
JMPR
B2C · Premium Mass · Q1 2027 Beta
Membership: $1.2K to $6K/yr
Commission: ~15% on GMV
Brokerage rates (not Virtuoso)
EXPLORER
$1.2K
NAVIGATOR
$3.6K
VOYAGER
$6K
OBELISK
B2B · Advisors & Agencies
Transaction fee per booking
SaaS access fee (planned)
Data licensing via MCP
Advisors bring the clients. Obelisk powers inventory, AI curation, and booking rails. We capture a fee on every transaction.
KARNAK Unit Economics
$192,000
Annual Revenue / Sovereign User
Membership · 6%$12,000
Commission on $100K/mo GMV @ ~15% · 94%$180,000
Member spends $1.2M/yr on travel. $12K membership = 1% of spend.
10 Go-to-Market
Build the audience first. Then leverage the infrastructure.
1
KARNAK Direct
NOW
Invite-only closed cohort. Billionaires and HNW families. Every client deepens the data flywheel. Growth through trust-based referrals. Strategic investor network as distribution.
2
B2B White-Label
NEXT
KARNAK deployed inside luxury hotel loyalty programs. Marriott Ambassador, Rosewood, Fairmont, Ritz-Carlton, Banyan Tree. Their top 5%, their whales. "Powered by KARNAK" = distribution at zero acquisition cost.
3
JMPR + Obelisk B2B
SCALING
MCP server opens inventory to all AI agents. Obelisk platform gives independent advisors unified workflow. Every channel feeds the same data engine. The flywheel accelerates with every new partner.
Unfair supply advantage: Government partnerships across Indonesia, Morocco, Egypt, Panama, UAE. 400+ luxury properties. Fully human-curated dining layer. IATA license. Direct hotel contracts in 12+ countries. No competitor has this access.
10 Competitive Landscape
Everyone serves luxury travelers. No one owns the data layer.
Legacy concierge services lack technology. AI startups lack relationships and inventory. We have both.
BOOKING INFRA
AI / AGENTS
PROPRIETARY DATA
LUXURY NETWORK
UHNW FOCUS
STATUS
QTG / KARNAK
●
●
●
●
●
$5.7M rev · AI + Ops + Data
Velocity Black
●
○
○
●
●
Raised $64M · Acquired by Capital One (2023). No AI.
Quintessentially
●
○
○
●
●
$2M raised · Est. ~$342M rev · 60+ offices, 1500 staff. No tech moat.
Ten Group
●
○
○
●
◐
Public (AIM: TENG) · $91M mkt cap · $91M rev. HNW, not UHNW.
Amex Centurion
●
○
◐
●
◐
Amex internal. Card perk, not core product. Outsourced.
Mindtrip
○
●
○
○
○
Raised $22.5M (Seed/A) · Mass market. No inventory.
Layla AI
◐
●
○
○
○
Raised ~$3.3M (Seed) · Mass market. No booking rails.
Aurora
○
●
○
◐
●
Raised $3.8M (Pre-Seed) · Ex-Velocity Black founder. No booking rails.
● Has capability◐ Partial○ Missing
The gap: Legacy concierge companies have relationships but no technology. AI startups have models but no inventory or supplier trust. QTG has both.
11 Financial Projections
Revenue roadmap. Three-year path to scale.
FY2026 is a transition year: QT winds down (~$1M), KARNAK scales (~$5M). Growth compounds from members, B2B, and AI margins.
FY2025 ACTUAL
FY2026 PROJECTED
FY2027 PROJECTED
FY2028 PROJECTED
Revenue
$5.7M
$6M
$22M
$52M
KARNAK Members
~10
100+
250+
400+
B2B Partners
-
3
15+
40+
Whale Accounts
1
4
8+
15+
Gross Margin
~45%
~55%
~65%
~72%
REVENUE MIX SHIFT
Whale: 60% today to under 25% by FY2028. Members + B2B diversify.
MARGIN EXPANSION
AI agents replace manual ops. Each new member costs less to serve than the last.
KEY ASSUMPTION
0.1% UHNW capture by Year 3. Conservative vs. Velocity Black's peak ~2K members.
12 Use of Funds
Product-led. Capital-efficient.
65% goes directly into Obelisk and KARNAK product. Growth spending activates trust-based channels, not paid ads.
Obelisk 40%
KARNAK 25%
Hires 20%
Growth 10%
Infra 5%
$2M
Obelisk Data Platform
MCP server live by Q3 with 3 pilot B2B partners. Sabre GDS integration. 10-region DMC automation. AI enrichment pipeline processing 10K+ properties.
$1.25M
KARNAK App & AI
Voice-first mobile GA by Q2. Service Agent handling 80% of routine requests. Personalization engine live across 100+ member profiles.
$1M
Key Hires & Ops
Head of Ops onboarded May. Head of Product by Q3. IATA/GDS credentialing, payment rails, and compliance infrastructure.
$500K
Growth & Partnerships
Activate 36K waitlist into 100+ founding members by Q4. EA/house manager outreach. Referral programs. Zero paid acquisition.
$250K
Infrastructure
Cloud, security, encrypted memory layer, monitoring, CI/CD. SOC 2 Type I by Q4. Built for high-value transactions.
13 Team
Humans set direction. Agents execute.
Team reduced from 35 humans and 2 AI agents to 12 humans and 20 AI agents. OPEX reduced 65% in 6 months.
Leadership
Linda Adami
Founder & CEO
Hunter Soik
Chairman
Vincent Guittet
Head of Growth Tech & Obelisk
Hao Cheng
Head of Product & Engineering
AI Agent Layer
Kiona
Chief of Staff Agent
Reports to: Linda
AI · Orchestration
Aria
Customer Concierge
Reports to: Kiona
AI · Service
Proposal Agent
Smart Match + Generate
Reports to: Kiona
AI · Matching
Marco
Restaurant Intel Agent
Reports to: Vincent
AI · Sourcing
CRM Brain
Client Intelligence
Reports to: Kiona
AI · Enrichment
Infrastructure Agents
GDS API Agents
Flights & Hotel Inventory
Reports to: Proposal Agent
API · Sourcing
Scraper Agents
Rate & Availability Mining
Reports to: Marco
AI · Sourcing
Browser Agents
Live Web Discovery
Reports to: Vincent
AI · Sourcing
Enrichment Engine
Data Profiling & Scoring
Reports to: Hao Cheng
AI · Enrichment
MCP Server Agents
JMPR Distribution Layer
Reports to: Hao Cheng
AI · Distribution
Ops
Head of Ops
TBD
Pipeline · Senior
?
TBD
Head of Product
Pipeline · Senior
Agentic profitability: Traditional concierge = more users → more staff. Our model: AI agents handle the work. User growth increases margin, not headcount.
Brand optionality: Spinning up a new brand on Obelisk is fast and cheap. Adventure, wellness, heritage. Same infrastructure, different audience.
14 The Ask
Join us at the inflection point.
$5,000,000
Pre-A SAFE · $100M Pre-Money · Quantum Temple Group
$1M
COMMITTED
$5M
TARGET
$4M open to co-investors
Lead investor committed $1M · YC SAFE · March 2026
12-Month Milestones
→ 100+ active KARNAK members → $2M+ monthly GMV → Obelisk API live · 3+ B2B clients → 3+ whale accounts → JMPR beta launch
Why $100M Valuation
→ $5.74M revenue at 642% YoY = ~17x → Vertical AI SaaS: 20 to 40x at this growth → Inspirato peak: $1.1B on ~$200M rev → Cap One / Velocity Black: ~$82M acq → Proprietary data moat + 4yr ops
What You're Buying Into
→ The only AI-native luxury platform → 4-year booking infrastructure → Data that exists nowhere else → Multiple independent exit paths → Agentic team that scales without cost
Contact
Linda Adami · linda@quantumtemple.io · karnak.co
A1 Obelisk Detail
The intelligence layer that doesn't exist anywhere else.
Every travel platform runs on parametric search. Obelisk replaces filters with intelligence.
TODAY
Traditional Advisor / Every OTA
Parametric Search
Advisor searches 5+ platforms by location, dates, star rating. Assembles proposal in spreadsheets. Starts from scratch every trip. Nothing remembers who the client is.
✕5+ disconnected tools per advisor
✕Zero memory between trips or clients
✕400 generic results sorted by price
✕Hours per proposal, every single time
"Hotel in Paris, 2 guests, June 15" → 400 results
OBELISK
AI-Powered Platform
Curated Search
Obelisk matches enriched inventory against profile, behavioral history, and implicit preferences. Proposal in minutes. Every interaction makes the next one smarter.
1
Unified Inventory
Five verticals in a single searchable layer. No more switching between disconnected tools.
∞
Compounding Memory
Every booking and rejection trains the next recommendation. Trip 10 is smarter than trip 1.
3
Curated Matching
3 perfect-fit options instead of 400 generic results sorted by price.
Min
Instant Proposals
Multi-segment proposals in minutes. No spreadsheets, no starting from scratch.
Same traveler → 3 hotels that match who they are
A2 Obelisk Architecture
How data flows from raw to curated. Every layer we own.
Raw feeds enter left. AI agents enrich and structure. Curated intelligence flows to our service layers and customers.
RAW SOURCES
GDS / NDCLIVE
Travel BrokerageLIVE
PDFs & DecksRAW
DMCs · 10 RegionsEXCL
AI DISCOVERY
Scraper Agents
Structured extraction from hotel sites, review platforms, DMC portals
Browser Agents
Live web discovery for new properties, experiences, pop-ups
OBELISK
ENRICHMENT CORE
PARAMETRIC LAYER
Structured data any GDS can provide. Necessary but not defensible.
+
NARRATIVE LAYER
What it feels like, who it suits, and why. Built from 4 years of operations. Proprietary.
SERVICE LAYER
Service Agent
Concierge AI
MCP & API
B2B interface
CUSTOMERS
KARNAK
B2C · UHNW
JMPR
B2C · Premium Mass
B2B Advisors
Travel agencies · Q4
A3 Strategic Value
Three credible exit paths. Different assets, different buyers.
Each path monetizes a different layer of the platform. All three are independently viable.
PATH 1 · HIGHEST STRATEGIC VALUE
Luxury Groups (LVMH, Kering, Richemont)
UHNW spending and preference data: the richest behavioral signal in luxury. The data alone justifies acquisition.
PATH 2 · STRONGEST PRECEDENT
Banks & Neobanks
Lifestyle as a card benefit. Precedent: Capital One acquired Velocity Black. Every premium issuer needs this and none have built it.
PATH 3 · INFRASTRUCTURE PLAY
Travel Tech / GDS (Amadeus, Sabre)
Curated luxury inventory and AI enrichment. They've spent years trying to move upmarket. Obelisk is the layer they cannot build.
.jpg)
