LUNARTULIP LAB
EN/CN
RESEARCH/GLOBAL AI

INDEPENDENT PUBLIC-EQUITIES RESEARCH · GLOBAL AI TECHNOLOGY

Discretionary Fundamental× Systematic QuantTwo Research Engines.Judgment under Continuous Test.

Lunartulip Lab is an AI-native independent research institution focused on global AI technology equities.Discretionary fundamental research explains industry causality, earnings and expectations; systematic quant tests signals, portfolio discipline and risk-adjusted value.For professional investors and institutional research teams seeking faster change detection, explicit thesis review and traceable improvement in judgment quality.

Read flagship research View the Authority Ledger
银色月面新月环抱郁金香的 Lunartulip Lab 标志
THESIS PATHQUESTION → HYPOTHESIS → TEST

HUMAN JUDGMENT IN LOOP

EVIDENCE SYSTEMSOURCE → CLAIM → CAUSE

TRACEABLE / UPDATEABLE

ACTIVE RESEARCH SYSTEM
LUNARTULIP LAB · 2026

WHO WE AREIndependent public-equities research
ENGINE / FUNDAMENTALDiscretionary fundamental research
ENGINE / QUANTSystematic quantitative research
RESEARCH WORKSPACEAlways-On Research Desk

RESEARCH ENGINES / 01

Two research engines continuously explain and test market opportunity.

Discretionary fundamental research addresses industry causality, company earnings and expectation gaps; systematic quant tests signals, portfolio discipline and risk-adjusted value. Together they produce traceable judgment, validation records and outcome feedback.

01

ENGINE 01 / DISCRETIONARY FUNDAMENTAL

Discretionary Fundamental Research

Trace how technology change travels through industry structure, company earnings, valuation and expectations—producing company deep dives, value-capture maps and forward tests.
Read flagship Deep Dives
02

ENGINE 02 / SYSTEMATIC QUANT

Systematic Quantitative Research

Test whether signals, weights, regimes and portfolio rules offer reproducible predictive or risk-adjusted value, with explicit validation design, sample boundaries, post-cost results and failures.
Explore quant research

THEMATIC INVESTMENT INTELLIGENCE / AI-NATIVE RESEARCH WORKSPACE

Keep judgment from both engines continuously updated in one workspace.

LunarTulip Research Desk is an always-on thematic investment intelligence service for AI technology, continuously maintaining investment-relevant evidence, thesis states and signals across a focused set of themes, with delivery through institutional research subscriptions and machine-readable thematic data APIs.Explore the real interface and nine-layer architecture
01 / RESEARCHGlobal information, industry causality and investment theses
02 / STRATEGYFactor research, backtests, stress tests and market regimes
03 / PORTFOLIOPortfolio state, risk budgets, constraints and monitoring
04 / LEARNINGDecision memory, attribution and rule updates
PUBLIC RESEARCH SERIES / LUNARTULIP ALPHAMAPFrom semantic graphs to cross-sectional stock factors, with dated and auditable methods research

Lunartulip AlphaMap public research preserves graph vintages, samples, signal definitions, negative results and citation metadata. Articles are public citation objects; underlying point-in-time data is delivered separately by use case and license.

Explore Lunartulip AlphaMap

PROOF 01 / RESEARCH OBJECTS

01

Lunartulip Deep Dive

Versioned deep research on specific companies and industries, preserving core theses, key evidence, risk boundaries, as-of dates and update histories at stable URLs.

BEST FOR
For institutions and professional investors evaluating company-research depth, industry-transmission judgment and update discipline.

WHAT IT INCLUDES

  • Timestamped thesis
  • Key evidence and risk boundary
  • Update history and corrections
POINT-IN-TIME / VERSIONEDContinuously updated bilingual canonical Research Objects
View Deep Dives

PROOF 02 / AUTHORITY LEDGER

02

Calls & Outcomes Ledger

Aggregate subsequent outcomes, dual benchmarks, cohort definitions and methodology so readers can evaluate prior judgment under one consistent rule set.

BEST FOR
For professional readers examining prior outcomes, validation discipline, sample transparency and payoff asymmetry.

WHAT IT INCLUDES

  • Separate reconstructed and discipline cohorts
  • Hit-rate, excess and payoff methodology
  • Source, as-of date and generated version
AGGREGATE / METHODOLOGY-LEDGenerated from the canonical ledger with reproducible methodology and timestamps
Open Authority Ledger
PROFESSIONAL & INSTITUTIONAL ACCESSBegin with public research, then move into human, machine and system delivery

After evaluating the research through Deep Dives and the Authority Ledger, move by coverage into Always-On Research Desk, the machine-readable Research API, alternative datasets for quant investors, or a six-session workshop that installs the team’s own research framework.

Explore products and access

RESEARCH DISCIPLINE / 02

Two engines validate independently and meet in the investment-learning layer.

Fundamental research preserves full explanations of industry causality, company earnings and expectation gaps. Quant independently tests signals, stability and post-cost value. Both meet through shared evidence, decision records and outcome attribution.

01 / POINT-IN-TIME

Judgment needs a timestamp before the future can test it.

02 / DUAL ENGINE

Industry causality
× statistical validation

Fundamental research addresses mechanisms and expectation gaps; systematic quant addresses reproducibility, stability and post-cost value.

03 / CORRECTION

Corrections are
research assets

New evidence triggers an upstream correction and regeneration of every affected version, keeping published judgment aligned with current evidence and an explicit point in time.

RESEARCH VALUE CHAIN / 03

Make research judgment traceable, testable and reusable.

From information intake, industry causality and thesis management to quantitative validation and outcome attribution, the connected research chain reduces information loss and repeated validation work while turning experience into reusable institutional knowledge.

01 / THESIS ENGINE

Thesis Engine

RESEARCH CHALLENGE

Investment intuition is scattered across meetings, chats and personal notes.

RESEARCH METHOD

Encode core assumptions, disconfirming evidence and time windows as updateable theses.

RESEARCH OUTPUT

Hypothesis cards and explicit next validation points

View method detail

Frame the core hypothesis, disconfirming evidence and critical milestones around a real investment question—preserving human judgment while making its objectives and boundaries legible to AI.

INPUTQuestion / intuition / market disagreementOUTPUTStructured thesis / falsification criteria
01 Thesis decomposition02 Expectation-gap mapping03 Milestone design

CONTINUOUS RESEARCH OPERATIONS / 04

Identify material change faster and update critical judgment in time.

The system continuously connects market change, key evidence, thesis state and outcome feedback into clear daily priorities and next validation actions—shifting research attention from chasing information to updating judgment.

01SENSE

Market change enters the event ledger

The system captures material change within the authorized scope; the researcher confirms whether it should enter the research state.

System capture · human confirmation
02UPDATE

Evidence updates the relevant thesis

New evidence links to a specific thesis, records what it strengthens or weakens, and updates falsification criteria and watchpoints.

Agent organization · researcher judgment
03BRIEF

Form today’s research priorities

The Research Desk compresses change into priorities, thesis states and the next validation actions.

System draft · PM review
04RECORD

Judgment enters decision memory

Preserve the information set, owner, risk boundaries and subsequent action behind each judgment.

Human decision · system record
05LEARN

Outcome feedback updates the rules

Separate judgment, execution and external noise, then update evidence weights, thesis states and the next operating cycle.

Joint review · version update

VALUE REALIZATION / 05

See research quality now.
Judge long-term value through capital outcomes.

Institutions and professional investors can first evaluate our research quality through published work, version histories, corrections and outcome records. Over time, both engines will remain accountable to real capital outcomes; any future asset-management activity will operate separately under the appropriate entity, qualifications and compliance framework.

Read the decision-attribution researchRead the system field notes
RESEARCH ACCESS

Evaluate research quality first

Use Deep Dives, research methods, update histories and the Authority Ledger to assess quality, then choose the human-readable Research Desk, machine-readable API, alternative datasets or the research-framework workshop.

Explore research products and services
LONG-TERM DIRECTION

Long-term capital-management direction

Real capital outcomes will test the two engines’ judgment quality, risk discipline and learning capacity over time. Related activity will operate separately under the appropriate entity, qualifications and compliance framework.

Discuss a strategic institutional partnership

DEEP RESEARCH / 06

See the depth of research through specific company and industry questions.

Continuously updated bilingual Research Objects address specific company and industry questions while preserving the research question, core thesis, key evidence, risk boundary, original publication date and update history.

DEEP DIVE / PLTR

Palantir: hard commercialization evidence, zero valuation tolerance

Reconcile the AI-sovereignty narrative and 65x TTM P/S through US commercial growth, contract value, retention and cash conversion.

Read the research

DEEP DIVE / NVDA

NVIDIA: the supercycle’s reconciliation variables have shifted

Demand is validated by hard numbers; risk has shifted to margins, supply commitments, guarantee concentration and receivables.

Read the research

DEEP DIVE / SNOW

Snowflake: hard acceleration data, AI attribution still awaits dollars

Reconcile consumption growth, AI-product accounts and reverse DCF while separating fact, attribution and market implication.

Read the research

DEEP DIVE / NET

Cloudflare: machine traffic is over half; monetization still lacks dollars

Separate business acceleration and agentic traffic from catalysts without financial terms, then reconcile at Q3.

Read the research

DEEP DIVE / TEAM

Atlassian: after three catalysts, Rovo still lacks a dollar metric

Reconcile the beat, Cloud acceleration and profit inflection with FY27 deceleration guidance, FCF and Rovo monetization disclosure.

Read the research

RESEARCH NOTES / 07

Research methods, system practice and long-term thinking

Ongoing methodological research for professional readers across discretionary fundamental, systematic quant, AI-native investing and investment decision systems.

View all research notes

ENGAGE LUNARTULIP LAB / CHINA & GLOBAL

Start with a research question worth testing.

Tell us about your institution, market focus and a specific research question, and we will begin with the most relevant research format.

China & global institutional partnershipschief@lunartuliplab.com