Context gets lost between systems.
Market data, filings, policy documents and operational controls describe different parts of a decision. Combining them without their dates, definitions and limits can produce a convincing but unusable conclusion.
Investor overview / Research-led deeptech
AION Analytics builds domain-specific intelligence infrastructure for decisions that need more than a plausible answer. Our initial commercial focus is financial institutions and the software teams serving them.
Inspect the public evidenceFor angels, venture investors and strategic partners. Company information, not a public offer of securities.
01 / The business case
An institution also needs to know which source was used, what was current, which rule applied and where a person must take over. That is the layer AION is building around intelligence.
Market data, filings, policy documents and operational controls describe different parts of a decision. Combining them without their dates, definitions and limits can produce a convincing but unusable conclusion.
Institutional research teams, brokers, quantitative developers and financial-software businesses need components they can evaluate inside their own workflows, not just another standalone answer.
Our commercial thesis is that preserving context and evidence can reduce duplicated integration and review work. Customer evaluations should measure that value, rather than assume it from a model score.
02 / Indian market infrastructure
AION Intraday Infrastructure is designed for quantitative funds, proprietary trading desks, eligible research professionals and financial-software teams working with Indian markets. Supporting products have separate access conditions.
Rust Greeks and indicators, five proprietary strategy families, Edge Heuristics and native Edge ML form an analytical workflow for Indian markets, from permitted inputs to analytical delivery.
The engine is ready for professional deployment on founder confirmation. Clients can evaluate the complete Full Engine workflow or a White-Box arrangement that retains their own strategy and detector.
Full Engine and White-Box contracts, dedicated streams and replay remain proposed. Customer source rights, legal review, tenant-isolation acceptance and agreed operating responsibilities are separate from engine capability.
Explore the infrastructureWhat a customer evaluates
Conceptual product flow, not a live endpoint, latency measurement or SLA. AION does not resell raw exchange ticks or private broker feeds, and analytical delivery does not place broker orders.
Evaluate the complete analytical workflow, from permitted inputs through strategy and Edge assessment to contracted output.
Retain your own strategy and detector. Evaluate processed Greeks and indicators with an optional separate Edge assessment. AION strategy implementations are not distributed.
These are capability categories. Formulas, triggers, parameters and implementation remain proprietary.
The independently scoped Edge Engine REST product covers index options for NIFTY, BANK NIFTY and SENSEX.
Proposed Full Engine coverage for gold, silver and crude oil depends on instrument-specific readiness and licensed data rights. Commodity coverage remains proposed under this lease; inclusion does not grant commodity API access.
Each arrangement is subject to technical and contractual acceptance. Deployment scope and acceptance are agreed for each customer.
Separate IMI scope. Standalone IMI accepts caller-supplied events. Proposed institutional event ingestion is a separate scope and is not automatically bundled into an engine lease.
Customer-supplied financial events become structured sector-impact analysis through a hosted API.
View IMIDated financial, cash-flow, ownership and company-relationship evidence for human analyst review.
View SentinelPre-open market ranges and directional context, with dated public validation records and explicit limits.
View VeritasA device-resident execution agent: guided broker API onboarding, customer-controlled limits with mandatory stop loss and target, orders through the customer's own broker account, and a signal-to-order record on the customer's device. A customer-authorised evidence interface for brokers is proposed, not built. Institutional evaluation only.
View SetuStandalone Edge Engine access, API plans and open-source market calendars are documented separately. Current provisioning and instrument coverage must be checked on the relevant product page.
03 / India and Lithuania
India is the operating-company discussion. Lithuania is a separately planned market-entry discussion. Local buyers, validation, capital and technology rights must be assessed on their own terms. Indian product work demonstrates execution experience; it is not Lithuanian revenue or automatic UAB ownership.
AION Analytics Private Limited develops AI systems and software. Financial firms and their technology partners are our initial commercial audience.
Investment discussions concern the Indian company. Product subscriptions, commercial pilots and investment commitments are recorded separately.
The proposed Lithuanian venture starts with a specific workflow: source-linked tax calculations for individual-activity taxpayers, with accountants as a later professional channel. A public tax pilot and Lithuanian market-calendar library provide a starting point for local validation.
Innovation Agency / Startup Lithuania
Reference: R4-3078-(1.6 Mr)-2026
The decision confirms founder Lokesh Gupta's eligibility for the proposed startup activity. It does not issue a residence permit, incorporate a UAB, certify a product or commit capital.
Inspect the tax pilot. The approval decision can be reviewed during diligence. Any expression of interest would remain subject to incorporation, investor due diligence and agreed terms. No commitment or completed investment is represented.
Separate entities, separate decisions. An investment in one track does not automatically include equity, assets or intellectual property in the other. Any ownership, licence or service relationship must be documented. The Lithuanian company remains proposed in this overview.
04 / Commercial model
Commercial discussions start with a defined buyer problem and a testable scope. These are routes to revenue, not a statement of achieved revenue or signed customer commitments.
Product-specific access, usage and support. Published plans define the service; installing an SDK does not distribute proprietary inference models.
See API accessA complete analytical workflow or selected components, scoped to the customer's environment, permitted inputs and operational requirements.
See deployment scopeSeparately agreed onboarding, acceptance, maintenance and support. Delivery should become repeatable rather than depend on open-ended custom work.
Discuss a workflow05 / Evidence before conviction
Public research, source repositories, product documentation and dated outcome records make different parts of the work inspectable. None substitutes for customer validation.
In a study of 19 selected Indian index expiries, max pain was closer to the expiry-day underlying close than the previous close in 12 cases.
The paper did not claim a robust forecasting advantage. Keeping that result is part of the method: an appealing explanation has to survive the comparison.
Small, selected sample. Underlying close, not separately reconciled final settlement. This is a research result, not product-performance evidence.
Read Signal Provenance, Part 2Signal Provenance, Part 1 examines the upstream evidence a downstream transaction record does not contain.
SSRN 7118421 / Founder-authored researchContextual Intelligence examines country-specific structure around economic events and decisions.
SSRN 7596079 / Founder-authored researchRead the validation archive with its measurement definitions, misses and historical gaps.
Public record / Scope varies by releaseInspect calendar repositories and product documentation. Proprietary engine source and model weights are not distributed by those documentation repositories.
Canonical GitHub organisationResearch papers are attributed to Lokesh Gupta. SSRN preprints are not independent product certification. Explore the research library.
06 / The economics behind the engineering
Our research draws on incentives, information and decision-making. For an investor, the relevant question is how those ideas shape a useful product and a repeatable business.
Examine the roles of the data source, software provider, operator and customer. Make the decision authority explicit.
Preserve source, timing and uncertainty instead of treating an available data point as current, complete or authoritative.
Test changing inputs, missing evidence and operating conditions. Define what the system should withhold or hand to a person.
Turn validated methods, tests and integrations into repeatable delivery. Reuse must respect customer confidentiality and data rights.
Who has information, who can act, and which incentives could conflict with the customer's outcome?
Which inputs reduce uncertainty, and which merely add apparent precision?
When should the system proceed, abstain or request human review?
Are misses, counterexamples and uncertainty visible before people act?
How do feedback, dependencies and changing participant behaviour affect the workflow?
Which small changes can invalidate the conclusion? This is a prompt for stress testing, not a claim to predict chaotic markets.
What can improve through versioned evaluation without silently changing the live system's obligations?
How might fiscal, monetary and demand changes reach a sector? Proposed transmission channels still need evidence.
These are explanatory lenses, not a claim that each product implements every theory, or that theoretical framing establishes a trading edge.
07 / What should become measurable
The investor question is not how many projects exist. It is whether a defined product can be adopted, operated and renewed on sustainable terms.
Choose the buyer, deployment scope and acceptance criteria before the evaluation.
Measure: review effort, integration effort, output usefulness and behaviour when inputs fail.Document deployment, support boundaries, updates and recovery for the contracted workflow.
Measure: onboarding time, operating cost, support load and reproducible acceptance.Test paid adoption and continued use before expanding into further products or geographies.
Measure: conversion, retention, realised margin and the reasons customers renew or leave.These are proposed evaluation milestones, not reported traction. Revenue, pipeline and financing materials are discussed privately with their dates and evidence.
08 / Founder and diligence
Founder and technical lead of AION Analytics. Leads systems architecture and the research programme behind the company's domain-specific intelligence work.
Based in Ahmedabad, with a separately planned Lithuanian venture. Our invitation is to examine the work, its commercial use and the evidence behind each claim.
Founder profile09 / Questions investors should ask
AION Analytics is an AI systems and software company. Its initial commercial focus includes financial institutions and software teams serving them. Its analytical products are not a retail buy/sell recommendation service. AION Setu is a separately scoped, device-resident execution agent: it places orders through the customer's own broker account only within limits the customer sets, generates no signals and gives no advice, and keeps the signal-to-order record on the customer's device. It is separate from the analytical and infrastructure products.
The Innovation Agency decision of 23 October 2026 confirms Lokesh Gupta's eligibility for the proposed startup activity under Lithuania's Startup Visa route, with validity until 21 April 2027. It is not a residence permit, incorporation certificate, product certification, grant or investment commitment. The proposed UAB and local customer validation are separate steps.
No. The Indian company and the proposed Lithuanian venture are separate discussions with separate operating plans. Shared technology, licensing, services or ownership must be documented rather than assumed. No equity or cross-entity rights are offered on this page.
This page provides company information and public evidence. It does not offer securities or accept subscriptions. Any investment discussion takes place privately, subject to applicable law, diligence, current entity documentation and negotiated terms.
Start with the business that fits
Angels, venture investors and strategic partners are welcome to discuss the relevant company, market and operating plan.
[email protected]These links open your email application. Nothing is sent automatically. Please do not include sensitive financial or identity documents in an initial enquiry.