360 ONE Multi-Stage AI Fund: Strategy, Portfolio & Fees

360 ONE Alternates Asset Management Limited manages the 360 ONE Multi-Stage AI Fund, a SEBI Category II Alternative Investment Fund (AIF). The manager describes it as India’s first multi-stage fund dedicated to artificial intelligence. It invests as a concentrated, growth-stage private equity fund with a target corpus of ₹750 crore and a green-shoe option of a further ₹750 crore.

Typically, investors looking at this fund want answers to a few questions. Where does the manager expect AI businesses to make money? Which companies will the fund own? And what does it cost to participate?

To answer them, this review covers the investment thesis, portfolio construction, the proposed day-one portfolio, the team, the fee classes and the fund structure. All figures come from the manager’s distribution partner teaser dated August 2026 and accompanying manager materials, unless stated otherwise.

In short, the 360 ONE Multi-Stage AI Fund, a ₹750 crore Cat II AIF with a fixed term, plans to take 6–8 growth-stage positions of ₹50–150 crore each. Specifically, it concentrates on AI infrastructure and on companies that sell AI as a finished business outcome. The manager has already identified proposed commitments of ₹384 crore, or 51% of the base fund, across three companies.

360 ONE Multi-Stage AI Fund at a glance

Information as of August 2026, per manager materials. The Private Placement Memorandum (PPM) governs these indicative terms.

Key observations

The fund runs a highly concentrated portfolio. For example, manager materials describe 6–8 core positions against a ₹750 crore base, with no more than 25% of investable funds in any one company (internal guideline about 20%). In practice, each holding will carry a large share of the fund’s outcome.

The manager has already earmarked about half of the base fund. The teaser lists ₹384 crore of proposed commitments across three companies before first close. As a result, investors can see much of the portfolio before committing, which is uncommon for a new private equity fund.

Moreover, the fund deploys almost all its capital into initial positions. Follow-on investments in successful companies go through separate co-investment vehicles, so the fund keeps no reserves for them. In addition, the manager expects to offer two co-investment opportunities at first close.

Fees step down with commitment size. The management fee falls from 2.00% at ₹1 crore to 1.25% at ₹75 crore and above, and the additional return falls from 20.0% to 12.5%. However, the 10% hurdle stays the same across all classes.

The thesis is sector-specific. In other words, all capital goes into AI businesses, with a stated tilt towards physical infrastructure (about 55–60% of capital) and outcome-priced AI services (about 40–45%). Investors therefore take on the risks of a single sector theme.

360 ONE AI Fund investment thesis: why value settles at the two ends of the AI stack

The manager’s argument starts with price. For example, manager materials cite Google processing about 3.2 quadrillion tokens a month, roughly 330 times the level of 24 months earlier. Over the same period, the cost of a frontier-model token fell about 15 times in a year. (A token is the unit in which AI models read and write text, roughly three-quarters of a word.)

In plain terms, AI usage is rising very fast while its unit price falls. For this reason, the manager compares AI to electricity. Once a product becomes cheap and widely used, it behaves like a utility, and the manager argues that profits then go to those who own generation and distribution, or who sell finished results. Meanwhile, competition erodes margins in the middle.

The 8-layer map and the fund’s two fronts

The manager splits the AI market into eight layers and groups them into three “fronts”:

Split indicative, per manager materials; subject to the PPM.

The fund stays selective in the middle layers because, as the teaser notes, OpenAI and Anthropic alone absorbed about 43% of global venture funding in the first half of 2026. Meanwhile, open-weight models have cut inference prices more than 100 times in two years. The manager believes Indian capital cannot compete in that layer, and it does not have to. However, the team keeps discretion to invest there if an exceptional Indian company emerges.

“Outcomes-as-a-Service” needs a short definition. Traditional software companies charge per user seat, while IT services firms bill per hour. By contrast, an outcome-priced business charges per job completed: a support ticket resolved, a medical scan read or an app shipped.

Why the manager focuses on India

Manager materials describe India as the world’s second-largest consumer of AI that produces almost none of it domestically. For instance, the teaser cites about 40,000 AI-grade GPUs in India against about 4 million in the US. On this reading, demand has arrived before local supply.

Three further points support the case in the manager’s materials. First, manager materials project that India’s data-centre capacity will grow about six times by 2030. Second, renewable power in India is among the cheapest in the world. Third, India’s $315 billion technology services industry has priced finished work for 30 years, which the manager sees as a natural base for outcome-priced AI businesses.

How the 360 ONE AI Fund builds its portfolio

The fund plans 6–8 core positions built over the investment period, at a pace of 3–4 new companies a year. The manager expects each investment to be ₹50–150 crore, large enough for the fund to lead or co-lead the round. In addition, the target is 5–15% ownership in each company, with a board or observer seat.

The fund will primarily buy unlisted equity and equity-linked securities, such as compulsorily convertible preference shares (CCPS) and compulsorily convertible debentures (CCDs). The fund invests predominantly in India, although SEBI norms allow overseas investments of up to 25% of investable funds.

Manager materials list these guardrails:

• No more than 25% of investable funds in any single company, with an internal guideline of about 20%.
• No leverage at fund level, other than temporary borrowing permitted under the SEBI AIF Regulations.
• Substantially all capital deployed into initial positions to maximise ownership at entry.
• Follow-on capital for successful companies through dedicated co-investment vehicles, where the fund holds the right of first allocation.

The co-investment design matters for investors in higher classes. Specifically, manager materials describe a 1:1 co-investment right, subject to commitment size, with a separate, lower fee schedule (see the fees section).

360 ONE AI Fund day-one portfolio: ₹384 crore in three proposed positions

The teaser identifies three companies for commitment at first close, totalling ₹384 crore, or 51% of the base fund. However, these commitments remain proposals, subject to fund closing, final diligence, documentation and Investment Committee approval. For confidentiality, the teaser uses project names for two of the three.

Neysa

Neysa offers GPU-as-a-Service and a managed AI platform called Velocis. Furthermore, the teaser states a 50–60% total cost of ownership advantage over global hyperscalers. Its founders include Sharad Sanghi, who built Netmagic into a large data-centre company before its sale to NTT, and Anindya Das, Netmagic’s former CTO.

According to the teaser, Neysa has 1,178 GPUs deployed, a plan to scale beyond 20,000, and more than 35 enterprise customers across BFSI, pharma and the public sector. Blackstone leads the company’s financing of up to about ₹11,500 crore, announced in February 2026, in milestone-linked tranches. In addition, 360 ONE Asset-managed funds already hold about 8.5%.

Project Celestial

Project Celestial runs a fibre-optic network built for data-centre-to-data-centre connectivity across metros and long-distance routes. The teaser reports 12,274 route-km operational (own and leased on indefeasible rights of use), a pipeline of 20,350 km by FY28 and more than 50 data centres and AI hubs connected across four metros.

Also, the network carries a 99.99% uptime commitment with a mean time to repair under three hours. About 90% of the signed order book sits on long-tenor contracts. Before this venture, the founders built and sold a telecom platform and have invested about ₹560 crore of their own capital. 360 ONE Asset leads the round.

Project Nova

Project Nova builds full software applications from plain-language instructions and charges by applications shipped. The teaser reports a ₹430 crore revenue run-rate, gross margin above 60%, 16.5 times growth in monthly recurring revenue between September 2025 and June 2026, and more than 1.2 million users in over 65 countries.

Its human-assisted managed delivery service, launched in February 2026, now accounts for more than 90% of revenue. 360 ONE Asset and Avataar Venture Partners co-lead the Series B. Other existing investors include Salesforce Ventures, Accel, Sierra Ventures, Together Fund and India Quotient.

360 ONE AI Fund manager and team

Abhishek Nag is the Fund Manager. He has 19 years across consumer internet and engineering and leads venture capital at 360 ONE. Before that, he served as a Partner at Lightspeed and worked at Netflix, Meta, Uber and Hike.

Meanwhile, Sameer Nath is the Chief Investment Officer, with 28 years across private equity, M&A and capital markets. He previously served as a Managing Director at Citigroup, then co-founded Iron Pillar and founded TrueScale Capital.

The wider team includes Harish Hulyalkar (25+ years in private equity and investment banking, formerly at Arpwood Capital, Iron Pillar and Citigroup), Animesh Pandey, Tessa Thomas and Kshiteej Prasad. Manager materials note that the team is subject to final confirmation and that the PPM sets out the Investment Committee’s composition.

The manager points to three sources of support. First, its sourcing work covers more than 300 AI companies evaluated over two years, with over 60 mapped on its India AI market map. Second, the wider 360 ONE group reported ₹7.77 lakh crore in assets under management as of June 2026. Third, the house previously launched a multi-stage Defence & Space fund of ₹1,500 crore. It also holds existing positions in AI companies such as Qure.ai, where it co-led the Series D with Lightspeed. For other 360 ONE Asset strategies, see our reviews of the 360 ONE Arise Credit Fund and the 360 ONE PIPE Fund Series 2.

360 ONE AI Fund fees, minimum commitment and terms

The fund offers six distribution classes. Therefore, each investor’s class depends on the size of their commitment.

Per manager materials, August 2026. Hurdle is pre-tax and post-expense. A separate class with lower management fees applies to investors who come through a SEBI-registered investment adviser or consultant. The PPM sets out other fees and expenses.

For co-investments, manager materials show a management fee of 1.00% for A3 and A4, 0.75% for A5 and 0.50% for A6. Similarly, the co-investment performance fee is 10% for A3 and A4, 7.5% for A5 and 5% for A6. However, classes A1 and A2 do not carry a co-investment fee line.

How the fee works on ₹1 crore

For example, take an investor who commits ₹1 crore in Class A1. During the commitment period, the manager charges the fee on the full commitment, so it comes to about ₹2 lakh a year, even before the fund draws all the capital. After that, the manager charges the fee only on capital that remains invested, so the fee falls as the fund sells companies.

Also, investors do not pay the full commitment upfront. Instead, the fund calls capital in stages. The launch note indicates an initial drawdown of 40% of commitment, or ₹40 lakh on ₹1 crore, with the rest called as the fund makes investments. Our guide to AIF capital calls and the J-curve explains how drawdowns work over the life of a fund.

The additional return (often called carried interest) applies only if the fund returns more than the 10% hurdle. Once the fund clears the hurdle, the “catch-up” lets the manager receive a larger share of profits until it has 20% of total profits in Class A1. Above that point, investors and the manager share profits in the agreed ratio. Finally, the PPM sets out the exact distribution waterfall.

Category II AIF structure explained

SEBI classifies AIFs into three categories. Category II covers funds that do not use leverage beyond limited temporary borrowing and that typically invest in unlisted companies, such as private equity and private credit funds. Category II funds run for a fixed term, so investors cannot redeem their units before that term ends.

This fund has a term of seven years from first close, extendable by up to two years under the AIF Regulations. The fund returns capital as it exits investments, through listings, sales to strategic buyers or secondary transactions. Investors should therefore expect to keep their money in the fund for most of its life.

Our explainer on private equity through Category II AIFs covers the category in more detail. Also, our guide to reading a Category II AIF PPM explains what to check in the offer documents.

Risks in the 360 ONE AI Fund

Manager materials list six risks and the fund’s response to each:

• Concentration. Large cheques into 6–8 companies, with no fund-level reserves. In response, the manager relies on per-company caps and co-investment vehicles.
• Compute commoditisation or over-build. For instance, GPU capacity may become cheap or excessive. The manager says it backs operators that earn managed-services and software margins on top of GPU rental.
• Immature outcome pricing. The manager favours companies with demonstrated outcome-priced revenue and long-term contracts.
• Incumbents adapting. Large IT companies may build competing services. However, the manager expects them to act as buyers and partners too.
• Policy and execution delays. The manager says it underwrites to private demand and treats government programmes such as IndiaAI as upside.
• Exit timing. In particular, returns depend on listings or sales within the fund’s life.

Beyond these, the usual private equity risks apply: valuations stay on paper until exit (our explainer on DPI, TVPI and IRR shows how to tell paper gains from cash returned), interim values are estimates, and early years often show negative returns while the manager charges fees on commitments. Investors call this pattern the J-curve.

360 ONE AI Fund: frequently asked questions

Key takeaway

Overall, the 360 ONE Multi-Stage AI Fund runs a concentrated, fixed-term Category II AIF portfolio that targets AI infrastructure and outcome-priced AI services in India. With ₹384 crore of its ₹750 crore base already earmarked for three identified companies, investors can assess much of the portfolio upfront. At the same time, the trade-off is single-sector exposure, high concentration and a seven-to-nine-year commitment. Finally, for more on the category, see Kalviro’s AIFs in India page.