Co-Founder Agreements in India (2026): A Comprehensive Drafting Guide

Last verified: April 2026

On 30 September 2024, a settlement was signed in a dispute that ran for over two and a half years. The co-founder of a leading Indian fintech unicorn agreed, under that settlement, to no longer be associated with the company in any capacity. Certain shares moved to a Resilient Growth Trust and to another co-founder. Both sides withdrew pending criminal complaints, civil proceedings, and the FIR that had been registered with the Economic Offences Wing in May 2023. Stripped of the litigation, the settlement read like the bad-leaver clause that should have been in the original Co-Founder Agreement in India when the company was incorporated.

The catch? Either it never was, or, if it existed in any disciplined form, nobody re-read it.

Three other recent founder events fold in as supporting beats. Four co-founders of a leading Indian healthcare-commerce platform stepped down from executive roles in January 2025. A fifth founder from the same company exited in August 2025. The co-founder of a leading Indian hyperlocal-delivery startup departed in January 2025 amid vendor and employee dues litigation, after which the company wound down. Each of those exits ended differently. Each one tested clauses the original founders’ agreement either had or did not have.

The pattern is the lesson. None of these disputes were exotic. Each one turned on a question a Co-Founder Agreement in India is built to answer in advance. Who can leave. On what terms. What happens to their shares. Whether their non-compete and confidentiality obligations survive the exit. Which forum decides the dispute when the WhatsApp group goes quiet and the lawyers send their first notice. Indian founder culture has historically run on handshakes and honour codes, which is fine until valuations correct, board positions become contested, or DPDP Act 2023 personal liability lands on someone who is no longer in the room.

Here’s the thing. The cohort of founders who have signed a tightly drafted agreement before incorporation isn’t large. The cohort that has signed one and then re-read it before a fundraise is smaller still. And the cohort that has had it reviewed against post-2025 case law (the Delhi High Court’s May 2025 ruling on Section 27 of the Indian Contract Act, the Supreme Court’s December 2023 Constitution Bench in Cox & Kings, the 2023 Mediation Act, the DPDP Rules of 2025-26) is smaller again. The fastest-moving Indian startup lawyers are quietly building a checklist for that cohort. This guide is that checklist.

What follows is the document those founders wished they’d drafted properly the first time. It walks through every clause an Indian co-founder agreement needs in 2026, why it needs that clause, how each one has been tested in court, and where the trap doors are. Founders who use this as their reference don’t just avoid the next dispute. They land their Series A on cleaner paper.

A co-founder agreement in India is a private contract executed at or before incorporation, governed by the Indian Contract Act, 1872. It records each founder’s equity, vesting (four years with a one-year cliff), roles, IP assignment, confidentiality, departure mechanics, and dispute resolution forum. It is binding under Section 10 ICA and survives DPDP Act 2023 founder obligations.

The rest of this guide unpacks each of those elements: what to draft, what to avoid, and which Indian case or statute will adjudicate the clause if it is ever tested. Use the table of contents below to jump to the section you need.

1. What is a co-founder agreement in India?

Most Indian startups still incorporate without a written co-founder agreement. The reflex is familiar. Friends meet at college, at a previous employer, at a hackathon. The pitch deck arrives before the term sheet. The cap table arrives before the agreement that should explain it. A practising commercial lawyer in Bengaluru will tell you that the single biggest predictor of a clean Series A is whether the founders signed a real document at incorporation, not three months later when the term sheet arrived.

A co-founder agreement is a private, written contract between two or more founders of a startup. It records, at minimum, each founder’s equity, role, time commitment, intellectual property assignment to the company, confidentiality undertakings, vesting schedule, leaver mechanics on departure, and dispute resolution forum. It is executed at or before incorporation. It is governed by the Indian Contract Act, 1872, and operates alongside the Articles of Association of the company once incorporated.

Is it legally binding? Yes. Under Section 10 of the Indian Contract Act, 1872, an agreement is enforceable when it is made by parties competent to contract, for a lawful consideration, with a lawful object, and is not expressly declared void by law. A co-founder agreement signed by two adults, supported by mutual promises (equity, time, IP) and lawful object, satisfies that test. The fact that no specific statute requires it does not make it less enforceable. It makes the founders’ choice to sign or not entirely a question of risk.

[HISTORICAL] India’s framework here builds in layers. The Companies Act, 2013 replaced the 1956 Act and tightened the corporate governance backdrop against which founders’ agreements operate (sections on memorandum, voting, share transfer, and related-party transactions all became routine reserved-matter triggers). Then the 2016 Startup India launch and DPIIT recognition framework pushed structured incorporation early. Around 2017-2019, Indian VCs converged on the four-year vesting / one-year cliff norm. By 2023, the Mediation Act and the DPDP Act added two more layers founders could no longer ignore. The document an Indian founder signs in 2026 sits on top of that fifteen-year accumulation.

In practice, what experienced practitioners watch for is the gap between “we have an agreement” and “we have an agreement that has been read in the last 12 months.” Investors at the seed stage often discover that the document was last opened the day it was signed, by founders who have since renegotiated equity informally over WhatsApp, hired a third co-founder without amending the agreement, or built a product on IP that lives, technically, on a personal GitHub account.

A common question on Quora threads about Indian startups: are oral co-founder agreements valid? The short answer is yes. Section 10 ICA does not require writing for most contracts. But evidentiary value is another matter altogether. An oral agreement produces a hearing where each side narrates a different chronology. A written agreement produces a hearing where each side reads from the same document. You don’t want to be in the first kind of hearing.

The pitfall founders tend to walk into is partnership inference. Where two or more people work together for shared profit without any written agreement, a Sec 4 Indian Partnership Act, 1932 claim can attach: a court may infer that they were carrying on a partnership rather than operating a company. That changes everything: liability, share of profits, the ability of one founder to bind the others. Sign before you start trading. Always.

2. Why most Indian co-founder agreements fail in practice

So here’s the question worth asking: why do disputes happen even when there is an agreement? The honest answer has four parts, and the fourth one accounts for most of the recent fintech and healthtech messes.

The first failure mode is no document. Founders trust each other, the lawyer’s bill seems unnecessary, and they delay. By the time a dispute arises, there is an email chain, a few WhatsApp screenshots, and a memory war.

The second failure mode is a vague document. The agreement exists but uses language like “the founders shall act in good faith” without defining what “good faith” means in the context of removal, share transfer, or competing employment. Vague clauses are unenforceable clauses.

The third failure mode is a document that contradicts the Articles of Association. Indian courts will read the AoA as the controlling instrument on matters of share transfer, director appointment, and shareholder rights. If the founders’ agreement says one thing and the AoA says another, the AoA generally wins on those subjects. Under Section 58 of the Companies Act, 2013, transfer restrictions in private companies operate via the AoA; a side agreement that doesn’t match the AoA struggles to bite.

The fourth failure mode (and this is the one nobody wants to hear) is a document that nobody re-reads at Series A. The original agreement gets superseded clause-by-clause by the SHA without anyone formally amending the founders’ agreement. Six months later, two clauses are in conflict, and the company is one term sheet away from a circular dispute about which document governs.

[SECOND-ORDER] What follows is a cascade. Sloppy IP assignment in the original founders’ agreement means the company’s title to its core IP cannot be cleanly traced at Series A. Investors run a “title trace” through every line of code, every design asset, every domain name. Anything still personally held by a founder becomes a Schedule of Disclosures item. The fix in 2026 (re-papering all material IP) costs roughly Rs. 5-15 lakh in legal time and three to six weeks of fundraise delay. Founders rarely build that into the runway.

[HISTORICAL] The 2019 wave of Indian founder corrections (publicly known across consumer-tech, hospitality-tech, and e-commerce) made vesting clawback and share buyback standard. The 2024-25 wave is doing the same for leaver and DR clauses. Each generation of founder learns by watching the previous generation’s mess. The next one learns by reading them.

There is also a quiet legal risk that catches founders without an agreement. Where two people work together on a venture without formally constituting a company or partnership, a court can find that they were de facto partners under Section 4 of the Indian Partnership Act, 1932. The consequence: shared liability for each other’s actions, joint ownership of partnership property, and rights to a share of profits regardless of what one party privately intended. Founders who never signed anything sometimes discover, mid-dispute, that they have inadvertently created a partnership rather than a company.

But what about removal? Can the board simply vote a co-founder out, or does that automatically attract an oppression claim? The Supreme Court answered this in Tata Consultancy Services Ltd. v. Cyrus Investments Pvt. Ltd., (2021) 9 SCC 449. Removal of a director or chairperson by a board majority, on its own, does not constitute oppression and mismanagement under Section 241 read with Section 242 of the Companies Act, 2013, where the AoA permits such removal. The case anchors a clean removal-clause architecture: drafted into the founders’ agreement, mirrored in the AoA, supported by a procedural fairness record, and survives a Sec 241/242 challenge.

2.1. Lessons from recent Indian co-founder disputes

Three recent Indian disputes illustrate the failure modes in real time. The fintech unicorn settlement of September 2024 turned, in substance, on the absence of a tightly drafted bad-leaver and removal clause: a founder was alleged to have engaged in misconduct, criminal complaints were registered, civil proceedings followed, and the resolution was negotiated rather than adjudicated. A clean leaver clause with a price formula and a removal-with-cause trigger would have shortened that drama by roughly two years.

The January 2025 healthcare-commerce platform exit involved four co-founders stepping down from executive roles, with a fifth exit following in August 2025. The reported friction sat at the intersection of valuation correction, executive role definition, and continuing equity rights. A reserved-matters list and a clear “if you leave the executive role, what vesting accelerates and what does not” provision would have made the negotiation cleaner.

The hyperlocal delivery startup wind-down in January 2025 illustrates a third failure pattern: the absence of disciplined cap-table and dispute mechanics in the original agreement made vendor and employee dues litigation harder to manage on the way down. None of the three were exotic. Each one turned on a question the agreement was supposed to answer.

3. The 17 must-have clauses in a co-founder agreement

If you read no other section of this guide, read this one. Below are the seventeen clauses every Indian co-founder agreement signed in 2026 should contain. Some clauses get a deep H3 here; others get a deeper deep-dive in their dedicated H2 later. The full list, optimised for a quick reference scan:

1. Parties and capacity (and a pre-incorporation recital)
2. Recitals: history of the venture, IP background, prior contributions
3. Equity split and capitalisation table reference
4. Roles, responsibilities, and minimum time commitments
5. Reserved matters and decision-making thresholds
6. Transfer restrictions: ROFR, ROFO, tag-along, drag-along
7. Confidentiality
8. IP assignment to the company (with pre-incorporation IP carve-in)
9. DPDP Act 2023 data fiduciary recital and inter-founder indemnity
10. Non-compete (during term only) and non-solicitation
11. Vesting schedule and reverse vesting on already-issued shares
12. Acceleration triggers (single or double, post-2025 norm)
13. Leaver mechanics: good leaver, bad leaver, death, disability, removal
14. Buyback price formulae (face value, FMV, Rule 11UA cross-reference)
15. Dispute resolution waterfall (mediation first, then arbitration; seat; institution)
16. Governing law and jurisdiction
17. Boilerplate (notices, severability, amendment, supersession by SHA, counterparts and e-signing)

That headline list is the first table of contents an investor’s lawyer flips to when reviewing the document at term sheet stage. Get all seventeen present. Then make each one substantive.

A subtle drafting question that does come up: where the Companies Act, 2013 default and a founders agreement override differ, which prevails? On most subjects covered by the Companies Act (transferability, voting, shareholder rights), the Act and the AoA prevail unless the founders’ agreement clauses are mirrored into the AoA. For pure inter-founder commercial undertakings (time commitment, IP assignment, confidentiality, non-compete during term), the agreement stands on its own as a contract. The drafting move that eliminates ambiguity is to mirror the corporate-law-touching clauses into the AoA at incorporation.

3.1. Parties, capacity, and pre-incorporation status

The opening recital matters more than founders think. Each party should be identified by full legal name, address, PAN, and (where relevant) DIN. Capacity should be stated: each founder is signing in personal capacity and warrants competence to contract under Section 11 of the Indian Contract Act, 1872.

Pre-incorporation status is the trap. An agreement signed before the company exists technically binds the founders as individuals. The clause that converts those personal obligations into corporate obligations on incorporation (a pre-incorporation contract ratification clause) is what makes the document work as the founders’ agreement of the company once the company exists. Without it, the founders have a bilateral contract between themselves and not an agreement that the company can enforce.

In practice, what experienced drafters do is structure the agreement as a deed between the founders that contemplates the imminent incorporation of a target company, and that automatically extends benefit and burden to the target company on its incorporation, subject to the company expressly adopting the agreement at its first board meeting. Quick, clean, two-paragraph fix.

3.2. Equity split and capitalisation table reference

The equity split clause does two jobs. It records the split itself (60/40, 40/40/20, etc.) and it cross-references the capitalisation table maintained by the company. The cross-reference is not decorative. As the company issues shares to employees, advisors, and investors, the cap table becomes the single authoritative document. The founders’ agreement should be drafted so that any conflict between the agreement and a properly board-resolved cap table update is resolved in favour of the cap table.

A common practitioner habit in 2026 is to maintain the cap table on a SaaS tool (EquityList, Qapita, Trica are commonly used in India) and to cross-reference that tool by name in the agreement. This is not a contractual oddity. It is a forward-looking move because investors at Series A will run their diligence against that cap table; if the source of truth is unambiguous, the diligence is a 30-minute exercise rather than a three-day reconciliation.

3.3. Roles, responsibilities, and time commitments

The roles clause should specify each founder’s title, scope of authority, and minimum weekly time commitment to the company. The time commitment is the part founders most often skip. It does work later. If a founder commits to “at least 40 hours per week, exclusive of any other paid engagement” and then takes a side consulting role, the breach is documented; the bad-leaver path becomes available without the messy “is this a real breach?” debate.

The clause should also reserve the company’s right to vary the role (with the founder’s consent) and should anticipate that the role will change as the company grows. Hard-coded titles (“Chief Technology Officer in perpetuity”) create more problems than they solve.

3.4. Reserved matters and decision-making thresholds

Reserved matters are the list of decisions that cannot be taken without the consent of all founders (or a specified supermajority). Common reserved matters include changes to the equity structure, related-party transactions, taking on debt above a threshold, hiring or removing senior leadership, changing the company’s principal business, and entering into binding term sheets. The list should be tight, not generic.

For two-founder companies, a deadlock-breaker is essential: a senior independent advisor who casts the deciding vote, a buy-sell mechanism, or an automatic mediation referral. For three-or-more-founder companies, the structure is different: a simple majority (with a supermajority reserved for the largest decisions) usually works, and the deadlock-breaker becomes a fallback rather than the primary tool.

3.5. Transfer restrictions: ROFR, ROFO, tag-along, drag-along

This is the clause investors will mark up most heavily at Series A, so getting it right at incorporation pays off. Right of First Refusal (ROFR) gives existing founders the right to match a third-party offer for another founder’s shares before the third party can buy. Right of First Offer (ROFO) is the lighter version: the seller must first offer to the other founders before approaching the market. Tag-along protects minority founders by letting them join a sale on the same terms; drag-along protects majority founders by forcing minority founders to sell when a qualified buyer wants the whole company.

Are ROFR/ROFO clauses enforceable in India for a private limited company? Yes. The Bombay High Court in Western Maharashtra Development Corporation Ltd. v. Bajaj Auto Ltd., (2010) 154 Comp Cas 593 (Bom) confirmed that pre-emptive rights in a shareholders’ agreement are valid and not violative of free transferability under the Companies Act. The clause survives. What matters is mirroring the restriction into the AoA so Section 58 of the Companies Act, 2013 gives it teeth at the company level.

A common Quora-style question: what does a drag-along right do to a minority founder? It compels them to sell on the same terms as the majority when a qualifying buyer triggers the clause. It can be useful (a clean exit for everyone) or punishing (the minority founder forced out at a price they do not love). The drafting fix in 2026 is to attach a price floor or a valuation-method gate to the drag-along, so the minority founder is not forced into a fire sale.

Cox & Kings Ltd. v. SAP India Pvt. Ltd., (2024) 4 SCC 1 also matters here on the multi-entity side: founders who sign the agreement at the operating company level while also holding shares in a holding company need the agreement to reach both entities. The Constitution Bench’s group-of-companies doctrine means non-signatory affiliates can be bound where consent is inferred from conduct and group structure, but the cleaner drafting move is to include all relevant entities as parties.

3.6. Confidentiality, IP assignment, and DPDP carve-outs

Confidentiality is one of two restrictive covenants that survive the Section 27 line on restraint of trade (the other is during-term non-compete). It should cover all confidential information, define exclusions narrowly (publicly available, independently developed, lawfully obtained from a third party), and survive termination. Founders pairing the agreement with a standalone NDA should review the broader drafting framework in a clause-by-clause NDA drafting guide for India.

IP assignment is non-negotiable. Every line of code, design asset, brand element, and trademark application created by a founder before or during the engagement, that relates to the business of the company, must be assigned to the company. The clause should cover (a) pre-incorporation IP that needs to be assigned in, (b) ongoing IP created during the engagement, automatically assigned, and (c) post-departure work that the departing founder must hand over.

DPDP carve-outs become non-trivial in 2026. The full treatment lives in H2 6, but the introductory clause needs to flag the data fiduciary recital, the inter-founder indemnity, and the leaver-clause carve-out for ongoing data exposure.

3.7. Non-compete and non-solicitation

Full treatment in H2 7. The introductory clause should record (a) a during-term non-compete (enforceable per the Gujarat Bottling line), (b) a tightly drafted non-solicitation that survives Section 27, and (c) confidentiality obligations that survive termination. Anything broader risks being void.

3.8. Vesting and reverse vesting

Full treatment in H2 5. The introductory clause records the four-year vest with a one-year cliff, monthly accrual after cliff, and the reverse-vesting structure on already-issued shares (where the company has the right to repurchase unvested shares at face value or fair market value depending on leaver category).

3.9. Leaver mechanics

Full treatment in H2 8. The introductory clause records the leaver categories (good, bad, death, disability, removal), the consequences for vested and unvested shares, and the buyback price formula.

3.10. Dispute resolution

Full treatment in H2 10. The introductory clause records the mediation-first waterfall (30-60 days), the arbitration fallback with a defined seat, and the carve-out for interim relief.

4. How co-founders should split equity in 2026 India

Here’s the question that derails more cohorts than any other. Two friends start a company. The product sketch is largely one founder’s; the seed capital is largely the other’s. What’s the right split? “50/50, obviously, we trust each other” is the fastest way to a frozen company in 18 months. The harder, better answer requires three frameworks and a documented rationale.

The first framework is equal split with a tie-breaker. Equity is divided equally; one founder is designated the “tie-breaker” on operational deadlocks. This works only if both founders genuinely accept the tie-breaker structure and the rationale is documented. A simple equal split without a tie-breaker is the structure that produces the deadlocked two-founder companies which become impossible to sell at Series A.

The second framework is role-weighted. Equity reflects the relative weight of roles: CEO 40, CTO 35, COO 25 for a three-founder company. The advantage is that it captures the asymmetry of contribution; the disadvantage is that role weight changes over time, so the agreement must include a documented rationale and a periodic review mechanism (informal, not a contractual entitlement to renegotiate).

The third framework is value-weighted. Each founder’s contribution (idea, capital, network, prior IP, future time commitment) is scored, and equity is allocated accordingly. The Slicing Pie / dynamic equity model is one version of this. It works in theory; in India, it works only until the first round of external capital, when the dynamic must freeze.

The Quora-style question that comes up constantly: one founder contributed only the idea, another the capital, is unequal split fair? The answer is contextual. An idea is worth roughly nothing without execution; capital is worth roughly nothing without an executable plan; sweat is worth roughly nothing without idea and capital. Most reasonable splits in this configuration land somewhere between 40/60 and 50/50, with vesting schedules that lock in the result over four years. What investors look for, at the seed stage, is a split that reflects actual contribution and a documented rationale memo (one to two pages) explaining why.

[HISTORICAL] The Indian VC norm on founder equity has moved through three phases since 2013. Phase one (2013-2016): equity splits were informal, often 50/50, set without reference to vesting. Phase two (2017-2019): vesting became standard at the seed round (US-imported four-year / one-year cliff norms), and the Slicing Pie debate hit Indian startup forums. Phase three (2020-2026): equity splits at incorporation are now expected to reflect a documented rationale; investors at the seed stage will quietly insist on re-papering equity if the original split looks reflexive.

What experienced practitioners do, in our view, is run founders through a one-hour conversation before the equity table is finalised. The questions are blunt. Who came up with the idea, and what was the idea worth at that moment? Who put in capital, and on what terms? Who is committing how many hours per week for the next four years? Who has prior IP that’s being assigned in? Who has the relevant network? The answers don’t always change the split, but they always change how cleanly the agreement holds up under stress.

The pitfall founders walk into is freezing the split too early without vesting. A 50/50 split signed on day one without vesting means that on day 30, if one founder leaves, they walk away with 50% of the company forever. That’s not a co-founder agreement. That’s a gift.

4.1. When 50/50 works and when it kills the company

50/50 works when the two founders have a track record together (they have built and sold a product before, they have weathered a real disagreement and resolved it, they have complementary skills neither can substitute) and when they have a contractual deadlock-breaker that they both genuinely accept. 50/50 kills the company when those preconditions don’t hold and when a deadlock arises early.

[SECOND-ORDER] The downstream effect of an unbroken 50/50 deadlock is brutal. The company cannot raise external capital because every investor wants to see a controlling structure. The cap table cannot be amended because both founders must consent. Senior hires walk away because they cannot get clarity on who decides. The company freezes mid-air. By the time the founders agree to fix it, six to nine months have been lost.

The fix is to draft the deadlock-breaker into the agreement at day one. Options include: an external advisor’s casting vote on operational matters, a buy-sell (“Texas shoot-out”) mechanism for terminal deadlocks, an automatic mediation referral, or a shifting-control mechanism tied to performance milestones. The right choice depends on the founder pair; the wrong choice is silence.

4.2. Slicing Pie and dynamic equity, and why it must freeze before Series A

Slicing Pie is a dynamic equity model where contributions (time, capital, IP, network) are tracked over time and equity is reallocated based on running totals. It produces a fairer split in early days because it doesn’t require founders to predict, on day one, exactly what each one will contribute over the next two years.

But Slicing Pie has no Indian tax safe harbour. Every reallocation is potentially a fresh share issuance subject to Section 56(2)(viib) fair-market-value scrutiny, and the Income Tax Act, 1961 has no carve-out for this model. More importantly, no investor at Series A will fund a company whose cap table is still in motion. The dynamic must freeze before the first external round, and the founders’ agreement should include the freeze trigger explicitly: “the equity allocations under this agreement shall be deemed final on the earlier of (a) the company’s first equity issuance to a non-founder, or (b) [date].”

In our view, Slicing Pie is a useful internal accounting framework for the first 6-12 months of a startup. It is not a long-term cap-table architecture. Treat it as a way to track contributions while the founders decide on the final split, not as the split itself.

5. Founder vesting in India: 4-year vest, 1-year cliff, and beyond

Vesting is the single most important commercial protection in the agreement. It exists for one reason: to make sure that a founder who leaves early does not walk away with shares they have not earned. The mechanics, by 2026, are standard across Indian startups, but the enforcement is anything but.

The four-year vest with a one-year cliff works like this. On signing, each founder is allotted (or held to be entitled to) the agreed equity percentage. Nothing vests for the first 12 months. At month 12, 25% of the founder’s equity vests in a single tranche (the “cliff”). From month 13 to month 48, the remaining 75% vests in equal monthly tranches (1/48 of the total per month, or roughly 2.083% of the founder’s pool every month). At month 48, the founder is fully vested.

Why the cliff? It’s a forcing function. If a founder isn’t going to be in the company for at least 12 months, the founder doesn’t earn any equity. It filters out the cohort who treat the engagement as optional. From month 13 onwards, the monthly accrual incentivises continuity.

A common comparison question: time-based vesting vs milestone-based vesting, which to use? Time-based is the default and is well understood by investors. Milestone-based ties vesting to specific deliverables (revenue thresholds, product launches, regulatory approvals). The disadvantage is that milestones are subjective and create dispute risk; the advantage is that they reward outcomes rather than time-served. Most Indian agreements in 2026 use time-based with a milestone-based overlay for one specific founder (often the technical co-founder) where it makes sense.

Investors at Series A will require vesting to be “re-papered.” That means the original four-year clock either restarts or is materially amended. Four-year vesting that was signed two years before Series A often becomes a fresh four-year clock from the Series A close, with credit for the time already served capped at one year. Founders who have not anticipated this in the original agreement get blindsided. The agreement should expressly contemplate Series A re-papering.

5.1. Reverse vesting on already-issued founder shares

Reverse vesting solves a structural problem in India. Companies often issue all founder shares at incorporation (because share issuance is administratively simpler than tranche-based allotment). But that means the founder owns the shares from day one. If the founder leaves at month 6, they technically own them.

Reverse vesting layers a contractual repurchase right on top. The shares are issued, but the company retains the right to repurchase the unvested portion at face value (or fair market value, depending on leaver category) if the founder leaves before the vesting clock matures. The reverse-vesting clause should specify (a) the trigger events, (b) the price formula, (c) the procedural mechanics (notice, payment, share transfer), and (d) the buyback authority under Section 67 of the Companies Act, 2013 and Section 68.

In practice, what experienced drafters do is mirror the reverse-vesting clause into the AoA, so the company-law machinery for share buyback (Section 68) is available rather than a contractual transfer that depends on the departing founder’s cooperation.

5.2. Acceleration: single trigger vs double trigger

Acceleration is the clause that protects founders against bad outcomes (termination without cause, change of control, board removal). Single-trigger acceleration means one event (often change of control: the company is acquired) automatically vests the remaining shares. Double-trigger means two events must occur (change of control AND termination without cause within a window after the change).

[FUTURE] The 2026 trend in Indian agreements is a quiet shift from single-trigger to double-trigger. Single-trigger was the norm 2017-2022 because acquirers used it as a retention tool. As 2024-2025 down-rounds normalised, acquirers and investors began pushing back, arguing that single-trigger acceleration on a change of control creates perverse incentives for founders to support a sale at suboptimal valuations. Double-trigger acceleration is becoming the standard in 2026-2028 drafting. Founders signing today should anticipate this shift and draft the clause with both options available, defaulting to double-trigger.

Are accelerated-vesting terms enforceable in India? The clauses are commercial and contractual, not statutory; their enforceability rests on Section 10 ICA. Where the clause specifies clear trigger events and clear vesting consequences, courts have given effect to them. The risk is in vague drafting (what counts as a “termination without cause”?), and the fix is precise definition.

5.3. Enforcing a vesting clawback when a founder refuses

[SECOND-ORDER] Here’s the part that gets ugly. The founder agreement says the company has the right to repurchase unvested shares at face value if the founder leaves before vesting. The founder leaves and refuses to sign the share transfer form. What now?

The first move is interim relief under Section 9 of the Arbitration and Conciliation Act, 1996 (if the dispute resolution clause specifies arbitration) or Section 14 of the Specific Relief Act, 1963. The company seeks an injunction restraining the founder from transferring the shares to a third party and an order directing the founder to execute the share transfer form. The second move is to enforce the AoA-level transfer restriction, which (if properly mirrored) gives the company a corporate-law route via Section 58.

But, in practice, the enforcement is messy. Indian courts move slowly on these matters; the arbitration route can take 12-18 months for a final award. The fastest practical fix is to structure the original agreement so that the company holds physical and dematerialised possession of unvested founder shares in escrow, with release contingent on vesting. Few Indian founders do this, but those who do save themselves the enforcement nightmare.

6. IP assignment, confidentiality, and the DPDP Act 2023 founder layer

The IP assignment problem is older than any DPDP question, but the DPDP layer is what makes IP and confidentiality drafting in 2026 different from drafting in 2023. Both have to be addressed in the same set of clauses.

The IP problem is structural. Pre-incorporation work product (code written before the company existed, designs sketched on personal devices, brand mood-boards drafted on a personal laptop) defaults to the founder personally under Section 17 of the Copyright Act, 1957. The work-for-hire doctrine attaches only where there is a contract of service between the creator and the company, and the company didn’t exist when the work was done. Without a positive assignment, that IP stays with the founder.

The fix is a tightly drafted IP-assignment clause with three components. First, an assignment-in clause: at signing, each founder assigns to the company all pre-incorporation IP that relates to the business. Second, an ongoing assignment clause: any IP created during the engagement, that relates to the business, is automatically assigned. Third, a residual handover clause: on departure, the founder hands over passwords, repository access, design files, and any related credentials. The lawful object test under Section 23 of the Indian Contract Act, 1872 is satisfied because the assignment is supported by the consideration of equity allocation and joint enterprise.

Confidentiality drafting in 2026 needs a DPDP overlay. The founder is a “data fiduciary” under the DPDP Act, 2023, on the same footing as the company itself for personal data the founder processes in the course of running the venture. Where the founder leaves and continues to hold customer data on personal devices, the remaining founders are exposed to penalty under Section 33 of the Digital Personal Data Protection Act, 2023 (up to Rs. 250 crore on the data fiduciary). The clause must address this directly.

In our view, the cleanest drafting move is to add a DPDP recital and an inter-founder indemnity. The recital states that each founder acknowledges joint data fiduciary obligations and undertakes to handle all personal data in accordance with Section 4 (lawful processing) and Section 9 (data fiduciary obligations) of the DPDP Act, 2023. The indemnity covers any DPDP penalty, regulator action, or data principal claim arising from a founder’s mishandling of personal data, including post-departure mishandling. Founders pairing this with the company’s external data processing template should review the data processing agreement template under the DPDP Act to make sure the inter-founder layer and the third-party-processor layer talk to each other.

[FUTURE] The DPDP Rules of 2025-26 (phased operationalisation) introduce tier classification for “Significant Data Fiduciaries.” Companies likely to fall into that tier (consumer apps with large data sets, fintech, healthtech) carry additional obligations: a Data Protection Officer, regular audits, impact assessments. The founders’ agreement signed in 2026 should anticipate this tier classification and reserve the company’s right to nominate a DPO from among the founders or to hire externally. Practitioners expect the Significant Data Fiduciary tier to expand materially through 2027-2028.

A common Quora question: can a co-founder hold the IP hostage during a dispute? They can try. If the IP was never properly assigned, the founder’s personal title stays intact and the company has limited remedies. Where the IP was assigned, the founder’s only leverage is procedural delay (refusing to sign handover documents, withholding repository credentials). The fix is to (a) assign IP in writing at incorporation, (b) maintain repository credentials in a multi-signature vault that no single founder controls, and (c) include a confidentiality-injunction clause supported by the approach in Desiccant Rotors International Pvt. Ltd. v. Bappaditya Sarkar (Delhi HC, 14 July 2009) to interim injunctive relief.

The pitfall in 2026 is treating DPDP as the company’s problem rather than the founders’ problem. The Act creates personal liability. A clean clause anticipates that.

6.1. The data fiduciary recital and inter-founder indemnity

Sample drafting language (illustrative only; tailor to your facts): “The Founders acknowledge that, in their joint capacity as Data Fiduciaries within the meaning of the Digital Personal Data Protection Act, 2023, each Founder bears responsibility for the lawful processing of Personal Data by the Company in accordance with Section 4 and Section 9 of the said Act. Each Founder hereby indemnifies the Company and the other Founders against any penalty, regulator action, or claim under Section 33 of the said Act arising from such Founder’s failure (whether during or after the term of this Agreement) to comply with applicable Data Fiduciary obligations.”

The clause should be paired with a leaver-clause carve-out: a departing founder remains bound by the DPDP indemnity for any personal data processed during their tenure or carried forward post-departure. That carve-out closes the loop on the bad-leaver-plus-DPDP convergence.

6.2. Pre-incorporation IP: the title-trace problem at Series A

[SECOND-ORDER] Investors at Series A run a title trace. They map every line of code, every domain, every design asset, every trademark application back to a chain of ownership. Anything that traces back to a founder’s personal capacity (rather than the company) becomes a Schedule of Disclosures item. In bad cases, the round is conditional on re-papering all material IP before close. The fix is preventive: assign in at incorporation, and document the assignment with consideration recitals.

A practical drafting move is to attach a Schedule of Pre-Incorporation IP to the founders’ agreement, listing every asset (with author, date of creation, and assignment status) and recording the assignment. The schedule is updated through the life of the company. At Series A diligence, the company hands over the schedule and the diligence is a 30-minute exercise rather than a three-week archaeology project.

7. Non-compete, non-solicit, and confidentiality after Varun Tyagi (2025)

If you’ve read US-style founder templates, you’ve seen broad post-employment non-compete language. None of it works in India. The Section 27 line is unforgiving and the Delhi High Court reaffirmed it in May 2025. So what survives?

The doctrinal anchor is Section 27 of the Indian Contract Act, 1872: any agreement by which a person is restrained from exercising a lawful profession, trade, or business is void to that extent. The exception is narrow (sale of goodwill in connection with sale of business). Founders’ agreements are not goodwill sales. They are joint enterprises. The Section 27 line bites with full force.

[HISTORICAL] The jurisprudence runs through a clear arc. Niranjan Shankar Golikari v. The Century Spinning and Manufacturing Co. Ltd., AIR 1967 SC 1098 set the foundation: restraints during employment are valid; restraints post-termination are void unless they protect trade secrets in narrow circumstances. Superintendence Co. of India (P) Ltd. v. Krishan Murgai, (1981) 2 SCC 246 confirmed it for the post-employment context. Percept D’Mark (India) Pvt. Ltd. v. Zaheer Khan, (2006) 4 SCC 227 reaffirmed it for service-provider style covenants. Gujarat Bottling Co. Ltd. v. Coca Cola Co., (1995) 5 SCC 545 carved out the during-term exception: negative covenants operating during the term of a contract do not amount to a Section 27 restraint and are enforceable. The Desiccant Rotors ruling narrowed the injunctive relief available even for confidentiality and customer-poaching post-employment. And then Varun Tyagi v. Daffodil Software Pvt. Ltd. (Delhi HC, 28 May 2025) reconfirmed that any term restricting a person’s right to be employed post-termination is void under Section 27.

So what does this mean for founders drafting in 2026? Three things. First, during-term non-compete clauses are enforceable; founders cannot start a competing business while still inside the venture. Draft this clearly, with carve-outs for permitted activities. Second, post-term non-compete is unenforceable as a blanket clause; do not draft it as a blanket. Third, confidentiality and non-solicitation (narrowly tailored, with reasonable scope and duration) survive the Section 27 line and are the workhorse restrictive covenants in 2026. Founders looking for the longer-form India-vs-UK comparative are the Indian position on post-employment non-compete enforceability.

A finer question: Section 27 ICA vs Section 41 of the Specific Relief Act, 1963, can a founder enforce a non-compete by injunction? The short answer is rarely, and only for narrowly drafted confidentiality or non-solicitation breaches with concrete protectable interest. Section 14 of the Specific Relief Act, 1963 excludes contracts that depend on the personal qualifications of a party from specific performance, and Section 41 of the same Act limits injunctive relief where the contract itself is unenforceable. Asking a court to injunct a founder from working for a competitor under a void Section 27 clause is asking for a refusal.

In practice, the drafting strategy that works in 2026 has three components: (a) a tight during-term non-compete with permitted-activity carve-outs, (b) a post-term confidentiality undertaking with no time limit (confidentiality survives termination), and (c) a 12-to-18 month customer non-solicitation and employee non-solicitation, with concrete definitions of “customer” and “employee” tied to the company’s books at the relevant date. Anything broader risks being void.

The pitfall is drafting a broad post-term non-compete and assuming it scares the departing founder enough to be useful as a deterrent. It doesn’t. Once the founder consults a competent lawyer, the bluff fails. Worse, the broad clause can taint the whole restrictive-covenant block: a court that finds one clause unenforceable may read the others narrowly.

7.1. During-term vs post-term restraints (Gujarat Bottling line)

The Gujarat Bottling line is the cleanest articulation of the during-term exception. Negative covenants that operate while the contract is alive (not after termination) are not Section 27 restraints because they are part of the affirmative obligation of the contract, not a restriction on a person’s future right to trade. A founder who agrees not to work for a competitor while still inside the venture is honouring an affirmative obligation, not surrendering a future right.

The drafting move is to anchor the during-term clause clearly to the term of the agreement and to specify that the obligation falls away on termination (replaced by confidentiality and non-solicitation, not non-compete).

7.2. What survives after Varun Tyagi 2025

[FUTURE] After the May 2025 Delhi High Court ruling, the drafting trend is unmistakable. Broad post-term non-compete language is being clipped from Indian agreements. What stands in its place is a layered structure: confidentiality (perpetual, narrow definition), customer non-solicitation (12-18 months, defined customer list), employee non-solicitation (12 months, defined employee list at termination), and IP-misuse undertakings. Litigators with experience in restrictive-covenant injunction proceedings expect this layered approach to dominate 2026-2028 drafting.

The litigation question worth thinking about is whether, even with a narrow non-solicit, a Section 41 injunction is realistic. The practical answer for an early-stage startup is that an injunction is expensive, slow, and uncertain. The faster fix is to draft the clause clearly enough that the founder doesn’t breach in the first place, and to maintain customer and employee lists as Schedules to the agreement so the boundaries of the obligation are not in doubt.

8. Exit and leaver mechanics: good leaver, bad leaver, death, and removal

Every Indian co-founder agreement in 2026 needs a disciplined leaver clause. Without one, a departing founder retains all their shares, and the company is stuck with a passive shareholder who has no operational engagement and possibly hostile interests. The fix is a well-drafted leaver mechanic with four pillars: definitions, vesting consequences, share buyback, and governance treatment.

The leaver categories that an Indian agreement should recognise are: good leaver (death, permanent disability, removal without cause, mutual termination), bad leaver (voluntary resignation before vest matures, termination for cause, material breach of the agreement, breach of restrictive covenants), death and disability as a sub-category, and removal as a separate category.

For each category, the agreement should specify the vesting consequence. Good leavers typically retain vested shares and may be entitled to accelerated vesting for the unvested portion (depending on circumstances). Bad leavers forfeit unvested shares (the company has the right to repurchase at face value) and may also be required to sell vested shares back to the company at fair market value (depending on the breach severity).

8.1. Good leaver vs bad leaver

Defining good leaver and bad leaver is where most Indian agreements get fuzzy. The discipline that works is binary trigger drafting: list the events that trigger each category, with no overlap. Good leaver triggers: (a) death; (b) permanent disability medically certified; (c) removal by the company without cause; (d) mutual termination by written agreement. Bad leaver triggers: (a) voluntary resignation by the founder before vesting matures; (b) termination by the company for cause (fraud, misconduct, conviction, material breach); (c) material breach of the restrictive covenants (confidentiality, non-solicitation, IP assignment).

The contested category is “termination for cause.” Loose drafting (“the company may terminate for any reason it deems sufficient”) is a recipe for litigation. Tight drafting (an enumerated list of cause events, each capable of objective verification) survives Section 241/242 challenge.

8.2. Pricing the buyback: face value, fair market value, and the Rule 11UA cross-reference

[SECOND-ORDER] The price formula is the single most litigated leaver clause. The norm in 2026: face value for bad leaver (the founder forfeits the gain on their shares), fair market value for good leaver (the founder is bought out at a fair price). FMV is determined per Rule 11UA of the Income-tax Rules (or by an independent registered valuer agreed by both sides).

The buyback authority lives in Section 68 of the Companies Act, 2013. The clause should reference the Section 68 procedural framework (board resolution, shareholder approval where applicable, payment within 12 months). Agreements that don’t reference Section 68 sometimes find themselves trying to enforce a contractual transfer that the company cannot execute without further corporate action.

The pitfall is price-formula opacity. A clause that says “fair market value as determined by the board” invites a Section 241 oppression claim by the departing founder. The fix is to specify (a) the valuation method (Rule 11UA, DCF, comparable company), (b) the valuer (a SEBI-registered merchant banker or a registered valuer), and (c) a tie-breaker if the parties dispute the valuation (e.g., averaging two valuers’ opinions).

8.3. Death, disability, and transmission of founder shares

What if a founder dies? Without a clause, shares transmit to legal heirs under Section 72 of the Companies Act, 2013 (transmission via nomination or by operation of succession law). The heirs become passive shareholders with no operational tie to the company.

The clause that addresses this: on death, the company has the right (not the obligation) to repurchase the founder’s vested shares from the legal heirs at fair market value, subject to Section 68. Disability is treated similarly. The clause should also address transmission via nomination (the founder may nominate a successor) and the mechanics of valuation when the heirs and the company disagree.

A common practitioner question: what if the founder stops working but refuses to sell? This is the bad-leaver-resigning-before-vest scenario. The agreement gives the company the right to repurchase unvested shares at face value; the leaver also forfeits any acceleration. Vested shares stay with the founder unless a separate “compulsory transfer” clause is triggered by, say, a material breach of confidentiality. The enforcement is as discussed in H3 5.3.

8.4. Removal: how the Tata Sons ruling shaped the clause

The clean removal clause owes its architecture to Tata Consultancy Services Ltd. v. Cyrus Investments Pvt. Ltd., (2021) 9 SCC 449. The Supreme Court confirmed that removal of a director or chairperson by a board majority, on its own, does not constitute oppression and mismanagement under Section 241 read with Section 242 of the Companies Act, 2013, where the AoA permits such removal. The drafting move that survives a Section 241/242 challenge is: (a) a clear removal clause in the founders’ agreement; (b) the same removal mechanic mirrored into the AoA; (c) a procedural fairness record (notice, opportunity to be heard, board resolution on grounds documented in the minutes); (d) reference to Section 169 of the Companies Act, 2013 (removal of directors by ordinary resolution).

The clause should also address the compensation consequences of removal: removal without cause typically activates the good leaver pricing (FMV), removal for cause activates the bad leaver pricing (face value).

9. Drafting walkthrough: 10 steps to a co-founder agreement that survives a fundraise

The clause-by-clause exposition is one thing. Sequencing the drafting work is another. Here are the ten steps a founder should follow, in order, before signing:

Step 1. Decide pre-incorporation MoU vs binding founders agreement. If the company doesn’t yet exist and the founders are still socialising the idea, a non-binding MoU recording intent and confidentiality is the right first document. Once the founders are committed and the company is being incorporated, the binding agreement comes next. Don’t skip the MoU step if the cohort isn’t fully decided. Don’t skip the binding agreement if the cohort is.

Step 2. Identify parties and capacity. Each founder identified by full legal name, PAN, address, and DIN. Capacity warranted under Section 11 ICA. Pre-incorporation status flagged.

Step 3. Settle the equity split with a documented rationale. Run the conversation, write the rationale memo, agree the split, document the reason. Two pages, signed alongside the agreement.

Step 4. Build the vesting schedule against the incorporation calendar. Four-year vest, one-year cliff. Reverse vesting on already-issued shares. Acceleration triggers (default to double-trigger in 2026). Schedule attached to the agreement.

Step 5. Draft the IP assignment and DPDP recital. Pre-incorporation IP assigned in. Ongoing IP automatically assigned. Schedule of pre-incorporation IP attached. DPDP data fiduciary recital and inter-founder indemnity included.

Step 6. Draft the non-compete carefully (post-Varun Tyagi). During-term non-compete only; perpetual confidentiality; 12-18 month non-solicit. No blanket post-term non-compete.

Step 7. Build leaver mechanics with a valuation method. Good/bad leaver definitions, vesting consequences, buyback price formula (face value vs FMV), Rule 11UA cross-reference, valuer tie-breaker.

Step 8. Build the dispute resolution waterfall. Mediation first (30-60 days under the Mediation Act, 2023); arbitration second (specify the seat, the rules, the institution); carve-out for interim relief under Section 9 of the Arbitration and Conciliation Act, 1996.

Step 9. Stamp, execute, notarise. Stamp duty paid per the relevant state stamp act (see H2 12 for the matrix). Execute in counterparts; e-signing is admissible per Section 35 of the Bharatiya Sakshya Adhiniyam, 2023 read with the IT Act. Notarise as a prudent step (not a legal requirement; see H3 9.1). Founders should also review the wider drafting framework for service contracts and execution rules in stamp duty rates and execution rules for commercial agreements in India.

Step 10. Cross-reference SHA template for Series A re-papering. The agreement should expressly contemplate that, on the first equity round, the SHA will be drafted to supersede or amend specific clauses (vesting re-papering, transfer restrictions, drag/tag, reserved matters). A short clause to this effect avoids the “is the FA still alive?” debate at Series A close.

Each step has a “common mistake” pointer that’s easy to miss. Step 1: drafting a binding agreement when the cohort isn’t actually committed (you’ve now locked everyone in). Step 3: agreeing the split without writing down the rationale (the next dispute starts with “but you said…”). Step 4: setting vesting from a date earlier than the document execution date (the math doesn’t reconcile at Series A diligence). Step 6: copying a US non-compete template without amendment (it’s void in India). Step 9: forgetting stamp duty (the unstamped document is admissible only after deficit + penalty paid; see H2 12).

9.1. Stamping, notarisation, and registration

A founders’ agreement is not in the list of compulsorily registrable instruments under Section 17 of the Registration Act, 1908. Registration is therefore optional. Notarisation is similarly optional but strongly recommended for evidentiary value.

Stamp duty, however, is mandatory. The applicable rate depends on the state where the agreement is executed (see H2 12 for the matrix). An unstamped agreement is not invalid, but it is inadmissible in evidence under Section 35 of the Bharatiya Sakshya Adhiniyam, 2023 until the deficit duty plus penalty is paid. That can stall an arbitration in early proceedings if the agreement is challenged on stamp grounds.

A subsidiary question: is a notarised co-founder agreement enforceable without registration? Yes. Notarisation strengthens evidentiary value but is not a substitute for stamp duty. An unregistered, notarised, properly stamped agreement is enforceable. An unregistered, notarised, unstamped agreement is enforceable only after stamp duty is paid.

9.2. Pre-incorporation MoU vs binding founders agreement

The MoU vs binding agreement distinction is the most-asked question on Step 1. The MoU records intent: who is exploring what, what each will contribute if the venture proceeds, and confidentiality of the discussions. It is non-binding except for confidentiality and the limited dispute resolution clauses.

The binding agreement is the document this entire guide is about. It records commitments, allocates equity, builds vesting, assigns IP, and creates the leaver and DR machinery.

Founders sometimes try to use the MoU as a substitute for the binding agreement. That fails the moment a real disagreement arises. The MoU was never designed to bear the weight of a leaver clause or an IP assignment. Use the MoU for what it’s for, then sign the binding agreement.

10. Dispute resolution architecture under the Mediation Act, 2023

So which forum should a co-founder agreement specify? In 2026, the answer for most Indian startups is: mediation first, arbitration second, court for residual interim relief. The reasons are practical, not theoretical.

[HISTORICAL] The 2010s and early 2020s in India were the arbitration-default era. Founders’ agreements specified arbitration almost reflexively, with little attention to seat, institution, or procedural sequencing. Then the Mediation Act, 2023 arrived. Mandatory pre-litigation mediation for many commercial disputes became statutory; courts began pushing pre-litigation mediation as a default. The drafting trend in 2026 is to mirror this in the contract: a mediation-first waterfall.

The mediation-first waterfall typically reads: any dispute arising under or in connection with the agreement shall first be referred to mediation under the Mediation Act, 2023; if not resolved within 60 days, the dispute shall be referred to arbitration under Section 7 of the Arbitration and Conciliation Act, 1996; the seat shall be at [city]; the rules shall be the [institution’s] rules. Each leg has its own drafting nuances: the mediation leg needs an institution (DIAC, MCIA, IIAM) and a process; the arbitration leg needs the seat, number of arbitrators, language, and rules.

What about interim relief? The agreement should expressly carve out the right of either party to seek interim relief from a court under Section 9 of the Arbitration and Conciliation Act, 1996 notwithstanding the mediation-first clause. That avoids the awkward situation where a party needs an urgent injunction (e.g., to restrain share transfer or IP misuse) but is contractually barred from approaching the court. The drafting move is a one-line carve-out: “nothing in this clause shall prevent either party from seeking interim relief from a court of competent jurisdiction.” Once arbitration is constituted, Section 17 of the same Act gives the tribunal the same interim relief powers.

[FUTURE] Practitioners expect the mediation-first norm to harden through 2026-2028. Indian courts continue to push pre-litigation mediation as a default; the Mediation Act framework is being operationalised. The cohort of agreements signed in 2026 that don’t have a mediation-first waterfall will look dated by 2028.

10.1. Mediation first, arbitration second

The drafting question worth asking is: how long should the mediation window be? 30 days is too short; 90 days is too long. 60 days is the sweet spot in our view. It gives the parties time to engage seriously, exchange a brief mediation submission, hold one or two sessions, and conclude. If the dispute is not resolvable in 60 days, it likely requires a determinative process.

The institution choice matters. DIAC’s mediation rules are cleaner than ad-hoc mediation; MCIA and IIAM are credible alternatives. Specifying the institution at signing avoids the post-dispute scramble.

A common comparison question: mediation vs arbitration vs court, which forum to choose? Mediation is fast, low-cost, and confidential, but produces a non-binding outcome unless a settlement is reached. Arbitration is slower (12-18 months for a final award), more expensive, but produces a binding award. Court is the slowest and least confidential, but is the residual forum for interim relief and challenges to arbitral awards. The 2026 best-practice waterfall combines the first two and reserves court for residual interim relief.

10.2. Seat selection: India vs Singapore

For early-stage Indian startups, the seat should be in India (Mumbai, Delhi, or Bengaluru). Singapore as a seat is attractive for cross-border disputes and for investors with a global mandate, but the cost is meaningful (institutional fees, foreign counsel, enforcement of award in India under Part II of the Arbitration Act). For a founder dispute, the cost is rarely justified.

If the founders are cross-border (one Indian, one foreign) and the company has a foreign holding structure, Singapore can make sense. The drafting question is whether the seat applies to the founders’ agreement only or also to the SHA layered on top. Aligning the two avoids forum-shopping later.

Should the agreement have an arbitration clause and where should the seat be? Yes, it should. The clause should specify the seat, the rules, the language, and the number of arbitrators. For most Indian founder agreements in 2026, the seat is Mumbai, Delhi, or Bengaluru (consistent with the company’s principal place of business), with one arbitrator for sole-matter disputes and three for material disputes.

10.3. Cox & Kings (2023) and binding affiliated entities

Founders sign at multiple levels: founder personal capacity, holding company, operating company. The arbitration clause needs to reach all relevant entities. Cox & Kings Ltd. v. SAP India Pvt. Ltd., (2024) 4 SCC 1, the December 2023 Constitution Bench, recognised the group-of-companies doctrine in Indian arbitration: non-signatory affiliates can be bound to an arbitration agreement based on consent inferred from conduct and group context.

The drafting move is to include all relevant entities as signatories (rather than relying on the doctrine to reach them post-dispute). Where that is impractical, the clause should expressly state that affiliated entities are bound by the arbitration agreement to the extent permitted by law, and that the doctrine applies. This anchors the architecture and reduces the risk of a founder challenging the arbitration jurisdiction on the basis that they signed personally but the dispute is between corporate entities.

11. Co-founder agreement vs SHA, MoU, term sheet, employment agreement, LLP agreement

A co-founder agreement is one document in a chain. To use it well, you need to know what each adjacent document does and how they interact. Here is the five-way comparison matrix:

Each document does a distinct job. The FA is the inter-founder commercial bedrock. The SHA layers investor protections on top at the first round. The MoU is the pre-commitment placeholder. The employment agreement is for the founder’s salary, benefits, and day-job obligations (separate from their shareholder rights). The LLP agreement is for venture structures using the LLP form rather than a Pvt Ltd.

A common comparison question: founders agreement vs founders MoU vs term sheet vs LoI, which one binds? The MoU and LoI are typically non-binding except for confidentiality and exclusivity. The term sheet is mostly non-binding except for specific clauses (confidentiality, exclusivity, dispute resolution, costs). The founders’ agreement is binding. The SHA is binding.

Another comparison: co-founder agreement vs employment agreement, should a founder also have an employment contract? Yes, in most cases. The FA covers shareholder rights (equity, vesting, transfer restrictions, leaver). The employment agreement covers salary, benefits, statutory deductions (PF, ESI, gratuity), and day-job obligations. Founders who skip the employment agreement create a gap: salary payments to a founder without an employment agreement raise questions in tax assessments, PF compliance, and labour law. Have both.

Co-founder agreement vs partnership deed vs LLP agreement, which structure for which stage? For a pre-incorporation phase (no structure yet), an MoU is appropriate. For a Pvt Ltd, the FA + AoA + (later) SHA is the stack. For an LLP, the LLP agreement is the primary document and the FA principles get integrated into it. The choice of Pvt Ltd vs LLP is a separate strategic decision driven by tax, fundraise readiness, and liability considerations.

11.1. From co-founder agreement to shareholders agreement at the first round

[SECOND-ORDER] The most-missed downstream effect of the founders’ agreement is what happens to it at Series A. The SHA is the document the investor’s lawyer drafts. It supersedes the FA on every subject it covers (vesting, transfer restrictions, drag/tag, reserved matters, dispute resolution). Founders who don’t anticipate this find themselves negotiating the same clauses twice: once at incorporation, once at Series A close.

Why do investors insist on re-doing the FA / SHA at the seed stage? Two reasons. First, the FA was drafted before any external party existed; investors need their interests reflected. Second, the FA may have clauses that are out of step with current investor norms (single-trigger acceleration, vague reserved matters, weak transfer restrictions). The investor’s lawyer rebuilds the framework.

Can a founders agreement bind future investors who join later? Not directly. The FA is between founders. Future investors enter via the SHA. The drafting move is to draft the FA so it can be cleanly amended or superseded by the SHA at first round, with the founders’ commercial bedrock (vesting, IP, confidentiality) flowing through.

Why does the term sheet sometimes override the founders agreement? The term sheet’s binding clauses (exclusivity, confidentiality, dispute resolution) bind the founders during the diligence and negotiation period. Once the SHA is signed, the SHA supersedes both the term sheet and the FA on overlapping subjects. The drafting hierarchy is: FA at day one, term sheet during fundraise (limited binding), SHA at close (full supersession).

12. Stamp duty, execution, and enforceability state by state

Stamp duty is where founders’ agreements quietly fall down. The rate depends on the state where the agreement is executed, and on the nature of the document. A founders’ agreement is typically classified as a “general agreement” under the relevant state stamp act, but some states (Maharashtra, in particular) have specific entries for shareholders’ agreements that may attract a higher rate. The Indian Stamp Act, 1899 provides the base; each state has its own schedule.

Here is an indicative state-wise matrix. Values may change with state amendments; verify the current rate before execution:

The values above are indicative and based on the general-agreement entry in the relevant state schedule. Where the agreement covers transfer of shares or future share allotment in specific terms, the higher SHA-style rate may apply (especially in Maharashtra). Always verify the current schedule and consult the state’s stamp authority before execution.

What if you signed in one state and the company is registered in another? The agreement is liable to stamp duty in the state of execution. If the agreement is then used as evidence in a different state, the stamp duty differential may need to be paid up. The cleaner move is to execute in the state where the company is registered.

A common pitfall: founders execute the agreement, miss the stamp duty, and only discover the gap at arbitration. Section 35 of the Bharatiya Sakshya Adhiniyam, 2023 (successor to Section 35 of the Indian Stamp Act for evidence purposes) provides that an instrument not duly stamped is inadmissible in evidence until the deficit duty plus penalty is paid. The penalty can be up to ten times the deficit. That delays the proceeding, costs money, and signals sloppy paperwork to the tribunal.

Another pitfall: company restructuring from LLP to Pvt Ltd. The original LLP agreement does not automatically convert to a founders’ agreement of the new Pvt Ltd. The founders need to execute a fresh founders’ agreement (or a deed of novation) on conversion, properly stamped under the relevant state act. Without it, the new Pvt Ltd has no inter-founder agreement to enforce.

In practice, what experienced practitioners do is draft the agreement to be executed in the state where the company is being incorporated, pay the higher of the relevant stamp duty rates (treating the agreement as an SHA-style document), and notarise as a prudent step. The cost is small. The downside protection is meaningful.

13. Tax and angel tax interaction with vesting (Section 56(2)(viib))

A subtle but important issue. When shares are issued at vesting (as in reverse vesting, where the company allotts shares progressively to a founder), the issue price needs to be benchmarked against the fair market value computed under Rule 11UA of the Income-tax Rules. If the issue price is below FMV, the differential becomes income in the company’s hands under Section 56(2)(viib) of the Income-tax Act, 1961: the “angel tax” provision.

In practice, founders’ shares are typically issued at face value at incorporation, when there is no external valuation and the FMV is essentially face value. The Section 56(2)(viib) issue arises when a vesting tranche is issued well after incorporation, at a time when the FMV (as computed under Rule 11UA) is materially higher than face value. If the company allotts that tranche at face value, the differential is taxable.

[FUTURE] DPIIT-recognised startups have historically benefitted from a Section 56(2)(viib) carve-out for share issuances above FMV (the angel tax exemption for eligible startups). The eligibility criteria and the windows for Section 80-IAC tax holiday under the Income-tax Act, 1961 have been amended several times. Practitioners expect further amendments through 2026-2027 as the government continues to refine the startup tax framework.

The drafting move that minimises angel tax risk is to (a) issue all founder shares at incorporation (rather than progressively at vesting), (b) layer reverse vesting on top so the company has the right to repurchase unvested shares, and (c) where DPIIT recognition is available, claim the carve-out at the appropriate stage. Cap-table SaaS tools (EquityList, Qapita) typically flag Rule 11UA issues at allotment.

The pitfall is treating Section 56(2)(viib) as the company’s tax problem rather than the founders’ problem. A mistimed allotment at face value, when FMV is higher, creates a tax liability that comes out of the company’s runway. Founders should sign off on each allotment with the tax position documented.

14. IPO-readiness: SEBI promoter lock-in and what your co-founder agreement should anticipate today

Why does this matter today, when an IPO might be five years away? Because the SEBI ICDR lock-in framework retroactively limits founder liquidity once the company files for a public issue. If the founders’ agreement doesn’t anticipate this, the founders find themselves locked in at IPO without a pre-negotiated buyback or partial-exit option.

Under the SEBI (Issue of Capital and Disclosure Requirements) Regulations, 2018, the minimum promoter contribution (Regulation 14) is locked in for three years from the date of allotment in the public issue. Holdings in excess of the minimum promoter contribution are locked in for shorter periods (Regulation 16: 18 months under post-2025 amendments, six months in some structures). Practitioners track each ICDR amendment cycle closely.

[FUTURE] The 2026 SEBI ICDR amendment cycle is expected to refine the lock-in framework further. The 2025 amendments tightened reporting on minimum promoter contribution. Practitioners expect lock-in periods to be refined again through 2026-2028 as SEBI calibrates against listing-volume trends and post-listing volatility. Founders’ agreements signed today should anticipate this trajectory rather than assume static rules. The full 2025 amendment context is in the 2025 SEBI Takeover Code amendments.

What should the founders’ agreement include? Three things. First, an acknowledgment that, on a public issue, the SEBI ICDR lock-in will apply to founder shares. Second, a partial-exit window provision: the founders’ agreement should permit a pre-IPO sale of a portion of vested founder shares (typically 10-20% of the founder’s holding) in the months before the DRHP filing, subject to investor consent and applicable law. Third, a buyback option: the company retains the right to repurchase a portion of the founder’s shares pre-IPO at FMV, providing partial liquidity without breaching the lock-in.

The pitfall is silence. Founders’ agreements that don’t address IPO-readiness create awkward conversations in the months before the DRHP filing, when the founders discover that all their shares will be locked in for three years post-listing.

14.1. Differential Voting Rights (DVR) for founders

The 2019 SEBI DVR norms allow unlisted companies to issue superior-voting-rights (SVR) shares to founders. The structure can lock founder control at low equity dilution. It is useful when the founders want to maintain control through a high-velocity fundraise without ceding voting majority.

Should every founders’ agreement include DVR? No. DVR is a niche structure that works for companies with strong founder-led conviction and investor receptivity. Most early-stage Indian startups do not need it. Where it is being considered, the founders’ agreement should reserve the right to issue SVR shares to founders at a future date (subject to shareholder approval, Companies Act compliance, and SEBI ICDR alignment if listing is anticipated).

The comparison question: founders’ DVR vs ordinary shares, when to use DVR for founders? Use DVR when founder control is the strategic priority and the cap table will dilute aggressively across multiple rounds. Use ordinary shares (the default) when the founder cohort is comfortable with proportional dilution. DVR creates listing complications under SEBI’s framework and should be deployed deliberately.

15. Cross-border founders: FEMA, FDI, and structural choices

What happens when one founder is Indian and another is foreign? FEMA and the FDI framework change the architecture of the agreement.

Under the Foreign Exchange Management Act, 1999 and the FEMA (Non-Debt Instruments) Rules, 2019, foreign investment into an Indian company is subject to sectoral caps and entry routes (automatic vs government). For most sectors relevant to startups (technology, e-commerce, fintech subject to specific RBI rules), the automatic route applies. The foreign founder’s shares are foreign direct investment; they are reported under FDI rules; and any subsequent transfer to or from the foreign founder triggers FEMA reporting.

The drafting question is whether the agreement should be governed by Indian law (the default for an Indian operating company) or by a foreign law (where a foreign holding company is interposed). For a pure Indian structure, Indian law governs. For a US-India founder pair with a Delaware or Singapore holding company on top of an Indian operating company, the holding-co founders’ agreement may be governed by Delaware or Singapore law, with a separate Indian operating-company arrangement governed by Indian law.

When does it make sense to flip to a Singapore or Delaware holding-co structure? When the cap table includes US investors with a strong preference for those jurisdictions, when a future listing is anticipated outside India, or when the founders need a tax structure that requires the foreign holding. The flip itself triggers tax implications and requires careful structuring.

The pitfall is signing a founders’ agreement governed by Indian law but executed by a non-resident founder without aligning it with FEMA reporting. Section 47 FEMA reporting risk attaches: the company has to file a Form FC-GPR on issue of shares to the foreign founder, and any transfer requires Form FC-TRS. Missing these filings exposes the company to FEMA penalty and the deal to compliance risk.

In practice, what cross-border founder cohorts do is execute two agreements: a primary founders’ agreement governed by Indian law (for the operating company) and a parallel arrangement at the holding-co level (governed by the holding-co jurisdiction). Both agreements cross-reference each other. The overall architecture is more complex but cleaner under audit.

16. Frequently asked questions

1. What is a co-founder agreement under Indian law?
A co-founder agreement is a private written contract between two or more founders of a startup, signed at or before incorporation. It records each founder’s equity, vesting, role, IP assignment to the company, confidentiality obligations, leaver mechanics, and dispute resolution forum. It is governed by the Indian Contract Act, 1872 and operates alongside the company’s Articles of Association.

2. Is a co-founder agreement legally binding in India?
Yes. Under Section 10 of the Indian Contract Act, 1872, an agreement made by competent parties for lawful consideration with a lawful object is enforceable. A co-founder agreement signed by adults, supported by mutual consideration (equity, time, IP), satisfies that test. The fact that no specific statute requires it does not affect enforceability.

3. Are oral co-founder agreements valid in India?
Section 10 ICA does not require writing for most contracts, so oral agreements can be valid. But evidentiary value is much weaker. An oral agreement produces a contested chronology in court; a written agreement produces a single document the parties can read from. Always sign in writing.

4. Is a co-founders agreement mandatory in India?
There is no statute that mandates a co-founder agreement. It is a commercial best practice, not a legal requirement. But the absence of one creates risks: partnership inference under the Indian Partnership Act, 1932, vague claims at exit, IP title gaps at Series A, and disputes that take years to resolve. Treat it as effectively mandatory.

5. When should a co-founder agreement be signed?
At or before incorporation. Signing after the company exists creates a gap during which liabilities may attach without coverage. Signing too early (before the founders are committed) creates a different problem: locking in cohorts that haven’t yet decided. The right moment is when the cohort is committed and the company is being incorporated.

6. Do you need to register a founders agreement in India?
No. Section 17 of the Registration Act, 1908 lists compulsorily registrable documents; a founders’ agreement is not in that list. Registration is optional. Notarisation is similarly optional but recommended for evidentiary value.

7. Is stamp duty payable on a founders agreement, and at what rate?
Yes. Stamp duty is mandatory and the rate depends on the state of execution (typically Rs. 100-500 for a general agreement, with higher rates in states like Maharashtra for SHA-style documents). An unstamped agreement is admissible in evidence only after the deficit duty plus penalty is paid under Section 35 of the Bharatiya Sakshya Adhiniyam, 2023.

8. How much does it cost to draft a founders agreement in India?
The professional fees for a properly drafted founders’ agreement range from Rs. 1 lakh to Rs. 2 lakh for a competent commercial lawyer in tier-1 cities, depending on complexity. Add stamp duty (Rs. 100-500), notarisation costs (Rs. 200-1,000), and any cap-table SaaS subscription. The total is typically Rs. 1.2-2.5 lakh for the document plus first-year cap-table tooling.

9. What is a founder vesting schedule, and what does “4-year vest, 1-year cliff” mean?
Vesting is the contractual mechanism that limits a founder’s right to keep equity if they leave early. “4-year vest, 1-year cliff” means: nothing vests for the first 12 months, then 25% vests at month 12 in a single tranche (the cliff), then the remaining 75% vests in equal monthly tranches over the next 36 months. At month 48, the founder is fully vested.

10. How is intellectual property assigned to the company by founders?
Through an IP assignment clause in the founders’ agreement. The clause should cover (a) pre-incorporation IP that needs to be assigned in, (b) ongoing IP created during the engagement, automatically assigned, and (c) post-departure handover obligations. The lawful object test under Section 23 ICA is satisfied by the equity consideration. Without a positive assignment, pre-incorporation IP defaults to the founder personally under Section 17 of the Copyright Act, 1957.

11. Can a founders agreement be amended later?
Yes. The agreement should include an amendment clause requiring written consent of all founders (or a specified supermajority). At Series A, the SHA typically supersedes the founders’ agreement on overlapping subjects, which is a form of amendment by supersession. Material amendments should be documented in writing and signed by all founders.

12. What is the difference between a founders agreement and a shareholders agreement?
A founders’ agreement is between founders only and is typically signed at or before incorporation. A shareholders’ agreement is between founders, investors, and the company, and is typically signed at the first equity round (seed or Series A). The SHA supersedes the FA on overlapping subjects (vesting, transfer restrictions, drag/tag, reserved matters). Both are governed by the Indian Contract Act, 1872.

13. What is a “good leaver” vs “bad leaver” clause?
Good leaver covers situations where the founder leaves through no fault of their own (death, permanent disability, removal without cause, mutual termination). Bad leaver covers fault-based departures (voluntary resignation before vest, termination for cause, material breach). Good leavers typically retain vested shares and may receive accelerated vesting; bad leavers forfeit unvested shares and may be required to sell vested shares back at face value.

14. Are post-employment non-compete clauses enforceable in India?
Generally no. Section 27 of the Indian Contract Act, 1872 voids any agreement restraining a person from exercising a lawful profession, trade, or business, with a narrow exception for sale of goodwill in connection with sale of business. The Delhi High Court reaffirmed this position in May 2025. What survives is during-term non-compete (per the Gujarat Bottling line), perpetual confidentiality, and 12-18 month customer or employee non-solicitation.

15. Does the DPDP Act 2023 affect a co-founder agreement?
Yes, materially. Each founder is a data fiduciary under the DPDP Act, 2023, with personal liability exposure. The founders’ agreement should include (a) a data fiduciary recital acknowledging Section 4 and Section 9 obligations, (b) an inter-founder indemnity covering Section 33 penalties (up to Rs. 250 crore), and (c) a leaver-clause carve-out for ongoing data exposure post-departure.

16. What happens to a founder’s shares if they leave the company early?
It depends on the leaver category and the vesting status. For a bad leaver before vesting matures, unvested shares are forfeited (the company repurchases at face value); vested shares may also be repurchased depending on the breach. For a good leaver, vested shares are retained and unvested shares may be accelerated. Without a leaver clause, the departing founder retains all shares regardless.

17. Can a founder be removed from the company under a co-founder agreement?
Yes. The founders’ agreement and the AoA should mirror a removal clause specifying grounds (cause vs without cause), procedure (notice, board resolution, opportunity to be heard), and consequences (good leaver vs bad leaver pricing). The Supreme Court in the Tata Sons ruling confirmed that removal of a director by board majority does not, by itself, constitute oppression under Section 241/242 of the Companies Act, 2013, where the AoA permits removal.

18. Does a founders agreement need to anticipate a future IPO and SEBI promoter lock-in?
Yes. SEBI ICDR Regulations lock in minimum promoter contribution for three years from the date of allotment in a public issue (Regulation 14), with shorter periods for excess holdings (Regulation 16). A founders’ agreement signed today should anticipate this by including (a) a partial pre-IPO exit window, (b) a buyback option providing limited liquidity, and (c) acknowledgement of the lock-in framework. Founders who don’t anticipate find themselves locked in retroactively.

17. References

Case Law

1. Cox & Kings Ltd. v. SAP India Pvt. Ltd., (2024) 4 SCC 1; 2023 INSC 1051 (5-judge Constitution Bench, 6 December 2023)
2. Desiccant Rotors International Pvt. Ltd. v. Bappaditya Sarkar, Delhi HC, CS(OS) No. 337/2008 (interlocutory order, 14 July 2009)
3. Gujarat Bottling Co. Ltd. v. Coca Cola Co., (1995) 5 SCC 545; AIR 1995 SC 2372 (citation verified across 577+ citing judgments; primary judgment URL on Indian Kanoon could not be confirmed within the verification window; Monitor to revisit)
4. Niranjan Shankar Golikari v. The Century Spinning and Manufacturing Co. Ltd., AIR 1967 SC 1098; 1967 SCR (2) 378
5. Percept D’Mark (India) Pvt. Ltd. v. Zaheer Khan, (2006) 4 SCC 227; Indian Kanoon document is the December 2003 Bombay HC predecessor (case caption reversed at HC stage); the SC affirming decision (2006) 4 SCC 227 is the controlling authority
6. Superintendence Co. of India (P) Ltd. v. Krishan Murgai, (1981) 2 SCC 246; AIR 1980 SC 1717
7. Tata Consultancy Services Ltd. v. Cyrus Investments Pvt. Ltd., (2021) 9 SCC 449; AIROnline 2021 SC 179
8. Varun Tyagi v. Daffodil Software Pvt. Ltd., Delhi HC, FAO 167/2025, 28 May 2025
9. Western Maharashtra Development Corporation Ltd. v. Bajaj Auto Ltd., (2010) 154 Comp Cas 593 (Bom); Arbitration Petition No. 174 of 2006

Statutes

1. Indian Contract Act, 1872: sections cited 10, 11, 23, 27
2. Indian Stamp Act, 1899: read with state stamp acts (Maharashtra Stamp Act, 1958; Karnataka Stamp Act, 1957; Tamil Nadu Stamp Act; Telangana Stamp Act; Gujarat Stamp Act, 1958; and Indian Stamp Act, 1899 as applicable to Delhi and West Bengal with state amendments)
3. Registration Act, 1908: section cited 17
4. Indian Partnership Act, 1932: section cited 4
5. Copyright Act, 1957: section cited 17
6. Income-tax Act, 1961: sections cited 56(2)(viib), 80-IAC; Rule 11UA of the Income-tax Rules valuation framework
7. Specific Relief Act, 1963: sections cited 14, 41
8. Arbitration and Conciliation Act, 1996: sections cited 7, 9, 17
9. Foreign Exchange Management Act, 1999: read with FEMA (Non-Debt Instruments) Rules, 2019
10. Limited Liability Partnership Act, 2008
11. Companies Act, 2013: sections cited 58, 67, 68, 72, 169, 241, 242
12. SEBI (Issue of Capital and Disclosure Requirements) Regulations, 2018: Regulation 14 (minimum promoter contribution lock-in) and Regulation 16 (lock-in for excess holdings)
13. Bharatiya Sakshya Adhiniyam, 2023: section cited 35
14. Digital Personal Data Protection Act, 2023: sections cited 4, 9, 33
15. Mediation Act, 2023

Secondary sources (commentary)

1. Mondaq (KT Advisors), “Founder’s Agreement – A Primer” (21 May 2024): https://www.mondaq.com/india/shareholders/1468380/founders-agreement-a-primer
2. Mondaq, Commentary on Varun Tyagi v. Daffodil Software (April 2025): https://www.mondaq.com/india/employee-rights-labour-relations/1643996/

Legal Disclaimer

This article is for informational and educational purposes only and does not constitute legal advice. Co-founder agreements involve commercial, tax, and regulatory considerations specific to each venture. For advice on drafting, executing, or enforcing a co-founder agreement in your particular circumstances, consult a qualified Indian commercial lawyer.

Founders often bring in outside help by engaging an advisor on a consultancy agreement rather than as a co-founder or employee.