SIGNAL 01 · DATA / MARKET

Indian startups raised $66 million last week. That is the lowest weekly number of 2026. The context is the signal.

According to data compiled by YourStory Research, Indian startups raised $66 million across 16 transactions in the week ending May 31. Not a single deal exceeded $15 million. It is the weakest weekly funding total recorded in the Indian startup ecosystem this year.

For comparison: the week of April 28 saw $588 million raised, driven largely by KreditBee’s $280 million Series E. The week of May 5 saw $204 million across 19 deals. Weekly VC inflows have swung between $44 million and $588 million in a six-week window.

That volatility is the more important data point than the $66 million itself.

India’s startup funding does not move in steady flows. It moves in pulses driven by a small number of large deals. When no deal above $15 million closes in a given week, the aggregate number collapses regardless of how many smaller rounds are happening. The 16 transactions that did close last week represent active deployment by early-stage investors. Seed and Series A activity in India has remained relatively consistent through 2026. What has disappeared is the $100 million-plus late-stage round.

Late-stage funding in India fell 38% to $5.6 billion in FY26 from $9.2 billion in FY25, according to Tracxn data. That’s a structural reconfiguration happening. Global capital that previously anchored large Indian rounds is now concentrating in US-based AI infrastructure companies. The rounds above $100 million that defined India’s 2021 and 2022 funding environment are not coming back at the same frequency.

The $66 million week is a symptom of that structural shift, not a signal of ecosystem collapse. Early-stage capital is moving. Deeptech funds are raising. The pipeline of IPO-ready companies is the largest India has seen. What is absent is the late-stage foreign cheque that inflates weekly aggregates.

SIGNAL 02 · VC BEHAVIOUR

Nandan Nilekani just anchored a ₹3,000 crore deeptech fund. The cheque sizes tell you what it is actually targeting.

On May 26, Ashish Kumar launched F2A, or Fundamentum Frontier Advisors, a new investment platform spun out of Fundamentum Partnership. SEBI has approved the core fund, named Fundamentum III, at a target size of ₹2,000 crore. The platform will manage an additional ₹1,000 crore in parallel co-investments, bringing total investment capacity to approximately ₹3,000 crore, or around $314 million.

Nandan Nilekani is the anchor investor. The fund will invest in 12 to 15 companies over three years, writing cheques of ₹40 to ₹90 crore per startup. It is targeting consumer AI, enterprise AI, and physical AI.

The cheque range is the most informative part of the announcement. At ₹40 to ₹90 crore per company, F2A is targeting post-seed to Series B stage startups — companies that have demonstrated product-market fit but are not yet at the scale that justifies a $30 million-plus institutional round. This is precisely the stage where Indian deeptech companies have historically struggled to raise because the capital was either too early (angels) or too late (global growth funds). F2A is explicitly trying to occupy that gap.

The Nilekani anchor matters for a reason beyond brand credibility. Nilekani’s operating record spans Infosys, Aadhaar, UPI, and ONDC. Each of those was an infrastructure project that worked because it was adopted at national scale. Nilekani’s investment thesis has consistently been about backing companies that build infrastructure others build on, not products that compete at the application layer. F2A’s focus on physical AI — robotics, automation, and hardware-software systems — reflects that same preference for infrastructure over application.

Debraj Banerjee, formerly of SIDBI Venture Capital, joins as general partner. SIDBI has been the institutional anchor behind several government-linked venture programs. His presence suggests F2A is positioning to access co-investment from public capital pools alongside private LPs.

SIGNAL 03 · IPO PIPELINE

Acko is preparing to go public at a valuation higher than where it last raised private capital. That is not common in this market.

Acko General Insurance has appointed ICICI Securities, Morgan Stanley, and Kotak Mahindra Capital as book-running lead managers for its planned IPO. According to people familiar with the matter cited by Bloomberg and PTI, the company is targeting a valuation of $2 to $2.5 billion and is expected to raise up to $350 million. It plans to file its DRHP with SEBI via the confidential pre-filing route in H2 2026, with a listing targeted for early 2027.

Acko was last valued at $1.1 billion in its October 2021 Series E round, at which point it had raised $428 million in total. The proposed IPO valuation of $2 to $2.5 billion would represent a meaningful step-up from that last private mark. This is at a time when most Indian startups that raised at 2021 valuations are listing at equal or lower marks.

Their financial trajectory justifies the premium. In FY25, Acko grew revenue 35% to ₹2,837 crore, significantly ahead of the broader Indian insurance sector’s 10% growth. Net losses narrowed 37% to ₹424 crore from ₹670 crore in FY24. The company is not yet profitable, but the loss reduction rate and revenue trajectory have been consistent for two consecutive years.

The business model is the differentiation. Acko built direct-to-consumer insurance from the ground up, eliminating the agent and intermediary distribution layer that traditional insurers including ICICI Lombard and Digit Insurance depend on. It sells motor, health, and travel insurance through its own digital channels and embedded partnerships. The cost of distribution is structurally lower than agent-led models. That advantage becomes more defensible as the company scales.

PolicyBazaar listed in November 2021. Digit Insurance debuted in May 2024. Acko entering the public market in early 2027, with stronger unit economics than either had at listing, positions it at the most mature point any Indian insurtech has reached before going public.

PIVOTAL TAKE

The $66 million week, the ₹3,000 crore fund, and the Acko IPO are the same story told three different ways.

India’s weekly VC number hit its lowest point of 2026 the same week Nandan Nilekani anchored a ₹3,000 crore deeptech fund and a Bengaluru Insurtech company began preparing to list at a valuation higher than its 2021 private mark.

The surface reading is contradiction. The underlying reading is transition.

The capital that is leaving India’s startup ecosystem is late-stage foreign capital that was never permanent. It came when global interest rates were near zero, when every emerging market was being treated as a growth arbitrage, and when India’s consumer internet story was new enough to attract risk appetite from funds that had no particular India conviction. That capital is now concentrated in US AI infrastructure. It is not coming back to India at the same scale or the same stage.

The capital that is growing in India’s ecosystem is domestic, patient, and deeptech-oriented. F2A at ₹3,000 crore, Shastra VC at $100 million, Piper Serica’s Bharat Tech Fund at ₹800 crore, these are not replacements for SoftBank Vision Fund. They are the beginning of a different kind of institutional capital base — one with a longer time horizon, a higher tolerance for technical risk, and a genuine understanding of what it takes to build infrastructure companies in India.

Acko going public at a premium to its 2021 private valuation is the proof point that this transition can produce real outcomes. It is a company that built slowly, improved its financials consistently, and is now approaching the public market from a position of operational credibility rather than narrative momentum.

The $66 million week is a data point about the capital structure India is transitioning away from. The F2A launch and the Acko IPO preparation are data points about the structure it is transitioning toward. Both are happening simultaneously, in the same week. That is the signal.

\ Pivotal Research

Forward this to one investor or founder who would find it useful.

Next issue: Monday, June 8.

The Signal · by Pivotal Research · Bengaluru, India

Pivotal Research (.in) is a boutique market intelligence firm producing decision-grade research for investors, founders, and strategy teams who cannot afford to be wrong. 400-plus engagements delivered. 96% repeat and referral business. We work as a project partner, a monthly diligence retainer, or fully embedded inside an investment committee process. Should your organization require any research services, please drop a note to george [at] pivotalresearch [dot] in.

Three Signals this week. A $63M bet that tells you where global capital is moving in India. A government infrastructure play that changes the cost structure of building AI here. And a new fund whose portfolio reveals a quiet shift in what Indian VCs are actually backing. Lets dive in.

SIGNAL 01 · FUNDING

General Catalyst just doubled Scapia’s valuation in twelve months. The category it chose tells you more than the cheque size.

On May 21, Bengaluru-based Scapia closed a $63 million Series C led by General Catalyst. Existing investors Peak XV and Z47 participated. Total raised to date: $135 million. Post-money valuation: over $500 million, up from approximately $200 million in April 2025.

That valuation move — more than doubling in a year — is the data point worth examining.

Scapia was founded in 2022 by Anil Goteti, formerly of Flipkart. The company issues co-branded credit cards with Federal Bank and BOBCARD, combines them with a travel booking platform, and offers rewards that work across flights, hotels, experiences, and UPI payments. The card is live across 17,500-plus pincodes in India and accepted in over 150 countries.

The product sits at a specific intersection: young Indians who travel regularly and want a financial product built around that behaviour. Every existing travel credit card in India was designed for a different era of the credit card customer.

Scapia is designed from the ground up for the IndusInd-to-Zepto generation.

General Catalyst leading this round is the signal inside the signal. This is a firm that led Stripe, Airbnb, and Snap in their early years. It has been building conviction in India deliberately, not opportunistically. It chose Scapia over every other Indian fintech available to back right now. That choice reflects a thesis: travel-linked financial products for young, digitally native Indians are a large, undercapitalised market and Scapia has the product and distribution to own it.

The competitive risk is real. HDFC’s Infinia, Axis Atlas, and American Express all compete for the travel credit card customer. They have decades of brand equity and points ecosystems. What they do not have is a product experience built for someone who discovered credit through Zepto’s BNPL and books travel through Instagram recommendations. Scapia is betting that distribution and UX matter more than reward rates for this cohort.

SIGNAL 02 · REGULATORY / INFRASTRUCTURE

India is offering GPU compute at Rs 67 per hour. AWS charges roughly Rs 330 for the same hardware. This changes who can build AI here.

The IndiaAI Mission has deployed 38,000 GPUs accessible to Indian startups, researchers, and institutions at a subsidised rate of Rs 65-92 per hour. After applicable project subsidies, the effective rate for approved applications falls below Rs 100 per hour.

AWS charges approximately Rs 330 per hour for H100 access. Azure sits higher. The subsidised IndiaAI rate is three to five times cheaper than the cheapest comparable commercial option.

This compute portal is actually live and not a future commitment. Fourteen service providers are empanelled. Applications are being processed. Sarvam AI was the first startup selected to build India’s sovereign LLM under the mission and has since launched two open-source models.

The government’s original target under the IndiaAI Mission was 10,000 GPUs. It has deployed more than three times that, and announced a further 20,000 units in the pipeline targeting 100,000 GPUs by end of 2026.

The strategic intent is explicit. IT Minister Ashwini Vaishnaw said it plainly: India is using public funds to provide the cheapest compute in the world because in most countries this infrastructure sits with private companies. The implication is sovereignty-first: Indian AI models, trained on Indian data, on Indian infrastructure, without a foreign hyperscaler in the chain at any point.

For a founder building an AI model in India today, the cost calculus has shifted. Training a model that would cost Rs 1 lakh on commercial cloud costs approximately Rs 60,000 through IndiaAI with subsidies. For a research institution or early-stage startup, the difference between viable and unviable is often that margin.

The constraint is not price, but access. Applications require government approval, and the approval process is not self-serve. Founders who want access need to apply early, frame their use case as being in the national interest, and expect a process that moves at government speed, not startup speed.

SIGNAL 03 · VC BEHAVIOUR

Shastra VC just launched a $100M deeptech fund. Its existing portfolio is the most interesting part.

On May 21, Bengaluru-based Shastra VC announced its third fund with a corpus of $100 million. The firm, formerly known as Veda VC, was founded by Vasant Rao, Avijeet Alagathi, and Ashis Nayak. It plans to write cheques of $500K to $3 million in IP-led startups from seed to Series A.

The sectors: space and defence technology, AI, climate sciences, renewable energy, semiconductors, advanced manufacturing, and biotech.

The portfolio tells the more interesting story.

Shastra has backed Simplismart (AI inference infrastructure), Sisir Radar (synthetic aperture radar for defence and earth observation), Alt Carbon (carbon removal), and Avammune (immunology biotech). These are neither consumer apps nor SaaS tools. They are companies building physical infrastructure or proprietary science that takes years to validate and cannot be quickly copied.

This is the third fund from a firm that deployed $55 million across its first two. The fact that it raised $100 million for Fund III — nearly doubling the combined corpus of its first two funds — reflects LP appetite for this category. Something has changed about what limited partners believe India can produce in deeptech.

The context around the launch is relevant. Deeptech was India’s third most-funded startup segment in 2025, after ecommerce and fintech, with approximately $500 million raised across 87 deals. In Q1 2026 alone, the sector pulled in $166 million. Separately, Piper Serica launched its ₹800 crore Bharat Tech Fund this month with a nearly identical sector focus. The government announced its ₹1 lakh crore Research, Development and Innovation scheme on May 13, targeting space, robotics, drones, advanced healthcare, and battery technology.

These moves are not coincidental. They reflect a view that India’s next decade of company-building will be defined less by consumer internet and fintech and more by hardware, science, and sovereign infrastructure.

PIVOTAL TAKE

Three signals. One direction. India is building the infrastructure layer for the next decade, and the capital is starting to follow.

General Catalyst backing Scapia at $500 million. The Indian government deploying 38,000 GPUs at prices that undercut every Western cloud provider. Shastra VC raising $100 million for a fund whose portfolio looks nothing like the India startup playbook of five years ago.

Read separately, these are an interesting funding round, a government scheme, and a new VC fund. Read together, they describe a structural shift in where value is being created and captured in India’s technology economy.

The Scapia round is not just a travel-fintech story. It is evidence that global capital is finding product-market fit in India across categories that were previously considered too niche, too local, or too behavioural to export the investment thesis. General Catalyst is not backing Scapia because India is a large market. It is backing Scapia because Scapia has built a product experience that reflects how a specific generation of Indians actually lives, and that generation is large and growing.

The IndiaAI compute programme is not just a subsidy. It is an infrastructure commitment that changes what it costs to build sovereign AI in India. When training a foundation model costs one-third of what it costs anywhere else, the question of whether India can compete in foundational AI shifts from capability to will.

The Shastra fund is not just another deeptech vehicle. It is a data point about LP conviction in Indian hard technology at a moment when that conviction is still forming. The funds that establish deeptech track records in India in the next three years will have a substantial first-mover advantage when the category matures.

The consumer internet phase of Indian startup building produced extraordinary value. The infrastructure phase, the one now beginning, will likely produce more. It will take longer, require different skills, and reward different investors. The capital moving this week is starting to position for that.

\ Pivotal Research

Forward this to one investor or founder who would find it useful.

For deeper research on any of these signals, write to george [at] pivotalresearch [dot] in.

Next issue: Monday, June 1.

The Signal · by Pivotal Research · Bengaluru, India

SIGNAL 01 · VC BEHAVIOUR

On May 12, Peak XV Partners, Ribbit Capital, and YC Holdings sold a combined 4.7% stake in Groww for ₹5,352 crore through bulk deals on the NSE. The six-month IPO lock-in had just expired.

The stock fell 10%.

That same week, Groww reported profit had more than doubled to ₹686 crore. Revenue grew 81%. Active users reached 16.7 million. It holds 28.5% of India’s retail broking market by active clients, the largest of any platform.

The business did not change on May 12. What changed is that a lock-in period ended.

Peak XV invested in 2018. Ribbit Capital in 2019. These funds held illiquid positions for six to eight years before Groww’s IPO. They manage money on behalf of pension funds, endowments, university foundations. Those LPs expect capital returned after exits. Selling at lock-in expiry is not a signal about Groww’s future. It is a VC fund completing its return cycle, as designed.

The buyers on the other side of the block deal were institutional investors purchasing at a 5-8% discount to market price. They understood what was happening.

There is a separate and legitimate debate about Groww’s valuation at a trailing P/E of 62-115x against competitors trading at 22-32x. That debate existed before May 12 and continues after it. It is a different conversation from the one triggered by the block deal.

WHY IT MATTERS

For angels and early-stage investors: a lock-in expiry sell-down on a strong business is frequently a buying signal, not a selling one. The market will misprice this repeatedly as India produces more IPOs from venture-backed companies.

For founders: the Groww block deal is a preview of what a successful VC exit looks like at scale. Your Series A investor will eventually need liquidity. The mechanics of post-IPO secondary sales are worth understanding now, before your own lock-in clock starts.

SIGNAL 02 · DATA / MARKET

Shadowfax grew its annual profit 17 times in one year. Almost nobody noticed.

Shadowfax listed on Indian exchanges in January 2026. Its Q4 FY26 results, published this week, show revenue of ₹1,237 crore for the quarter, up 74% year on year, and a net profit of ₹56 crore against a loss of ₹10 crore in the same quarter last year.

For the full year FY26: revenue of ₹4,202 crore, up 69%. Net profit of ₹112 crore, up from ₹6.4 crore in FY25.

The market share number is the one worth holding. In FY22, Shadowfax held 8% of India’s third-party logistics express market. By Q4 FY26, that share had reached 27-29%. In four years, it became the second or third largest express logistics company in the country while also turning meaningfully profitable.

Its current market cap is approximately ₹9,822 crore. Delhivery, its most discussed competitor, trades at roughly ₹21,000 crore while still operating at a loss.

Shadowfax is profitable, growing faster, and valued at less than half of Delhivery. That gap either reflects something real about Delhivery’s more diversified enterprise client base, or it is a mispricing that corrects as Shadowfax’s results compound. Possibly both.

The quick commerce exposure is the variable to track. Shadowfax’s hyperlocal segment grew 32% this quarter. If quick commerce consolidates and major platforms pull logistics in-house, Shadowfax’s growth profile changes. That risk is real. It does not change what the FY26 numbers say.

WHY IT MATTERS

For investors: Shadowfax is being ignored because logistics is considered a low-margin category. The P&L disagrees with that framing right now. The profit trajectory, not just the revenue line, is worth a closer look before the market notices.

For founders in ecommerce and quick commerce supply chains: Shadowfax’s ability to serve both express parcel and hyperlocal simultaneously is what drove its market share gains. Single-channel logistics operators are under pressure. The operators running two channels are pulling ahead.

SIGNAL 03 · REGULATORY

Swiggy is removing Accel and SoftBank’s board nomination rights. The reason is more important than the action.

Swiggy filed with stock exchanges this week to clarify changes to its board nomination framework. The company is working toward classification as an Indian Owned and Controlled Company, or IOCC, under FEMA.

To qualify, both ownership and effective control must rest with resident Indian citizens or Indian-controlled entities. Roughly 60% of Swiggy’s shareholding is currently held by foreign investors including Prosus, SoftBank, Accel, and GIC. Swiggy cannot qualify on ownership today.

What it can restructure is governance. The board changes approved on April 10 remove nomination rights previously held by Accel and SoftBank. The founders retain modified rights. The foreign investors retain their shareholding. What changes is who can place a director at the table.

IOCC status under FEMA unlocks easier access to sectors with FDI restrictions and reduces friction in government procurement and regulated industries. The timing is also relevant: Swiggy’s retail shareholder base has been growing since listing, and domestic institutional accumulation brings the 50% resident ownership threshold closer over time.

This filing will be replicated. Several listed Indian tech companies carry similar cap table structures. The ones watching Swiggy’s move most carefully are the ones weighing the same decision.

WHY IT MATTERS

For investors in listed Indian tech: IOCC status is becoming a governance lens, not just a compliance checkbox. It changes a company’s regulatory surface area and its relationship with domestic institutional capital. Watch for similar board amendment filings from other listed tech companies in the next two quarters.

For founders with foreign investors: governance rights and economic rights are separate. Board nomination rights can be amended through Articles of Association changes with shareholder approval, even post-listing. Understanding when that restructuring serves your company’s interests is a founder-level question worth asking early.

PIVOTAL TAKE

Three signals. One pattern. The Indian market is still learning to read its own companies.

Groww’s stock fell 10% the same day it reported its strongest quarterly profit on record. Shadowfax grew annual profit 17 times and trades at half the valuation of its slower competitor. Swiggy restructured its board under a regulatory framework most analysts do not track.

In each case, something real happened that the market did not fully process.

• The Groww reaction is a market literacy problem. Post-IPO block deals by VC funds are routine in any mature public market. In India’s retail-dominated markets, they read as insider selling. That gap between what the signal means and what the market hears will recur every time a venture-backed company’s lock-in expires.

• The Shadowfax gap is a category problem. Logistics companies are not supposed to have interesting earnings. When one grows profit 17 times in a year, the response is measured. The ‘logistics’ label is doing more work than the data.

• The Swiggy filing is a regulatory literacy problem. FEMA’s IOCC framework has existed for years. Most investors in Indian listed tech do not track it. A major governance restructuring that will influence how foreign-backed companies operate in regulated sectors reads as a routine disclosure.

These companies are ahead of the market’s understanding of them. That gap is where the return is.

\ Pivotal Research

If you found one useful thing in this issue, do consider subscribing and forward it to one investor or founder who would appreciate it. That is how The Signal grows.

Want deeper research on any of the signals above? Reach out to us at george [at] pivotalresearch [dot] in.

Next issue: Monday, May 25. See you then.

The Signal · by Pivotal Research
Bengaluru, India

Three things happened in India’s startup ecosystem this week that deserve more attention than they received. One of them looks like a rescue. Look closer.

SIGNAL 01 · ACQUISITION

The upGrad-Unacademy deal is not what it looks like. The number worth studying is ₹900 crore, and it sits on Unacademy’s own balance sheet.

upGrad filed for regulatory approval with the Competition Commission of India (CCI) last week to acquire Unacademy in a 100% share-swap deal valuing the edtech company at approximately ₹2,055 crore (~$218 million). The transaction is expected to close by July 2026, pending CCI clearance.

A company valued at $3.4 billion in 2021 is exiting at a 90% markdown. India’s edtech boom, deflated.

Except the structure of this deal is more interesting than the valuation.

• The Cash Value: At closing, Unacademy is expected to hold ₹900 to ₹950 crore in net cash. This means upGrad is effectively offering ~$118 million of its own equity to acquire $100 million of real money, plus a learner base, a test-prep brand, and a founding team. The cash is the most valuable single line item in this transaction.

• The Investor Roll: SoftBank, Tiger Global, and Temasek are not “exiting.” They are rolling their positions into upGrad, a private entity last valued at $1.9 billion, with no confirmed IPO date. They are trading one illiquid bet for another, essentially “extending and pretending” as their 2021 vintage funds reach their midpoint.

• The Founder Play: Gaurav Munjal stays on as CEO of the unit. Ronnie Screwvala gains a capital-efficient way to consolidate a fragmented market while the target partially funds its own acquisition through its remaining balance sheet.

SIGNAL 02 · SECTOR INFLECTION

Urban Company is losing ₹447 on every InstaHelp order. That number rose last quarter, not fell.

Urban Company reported its Q4 FY26 results on May 8. Revenue grew 43% year on year to ₹426 crore. Net loss came in at ₹161 crore, against ₹2.8 crore in the same quarter a year earlier. Losses expanded 57 times on a year-on-year basis.

The mechanism behind those numbers is a single business line, InstaHelp — Urban Company’s quick home-services vertical that did not exist at the start of FY26.

By March 2026, it was processing over 1.1 million orders a month. In Q4 alone, it handled 2.7 million orders and generated ₹40 crore in net transaction value. The adjusted EBITDA loss for InstaHelp in that quarter was ₹119 crore.

The figure worth examining is ₹447. That is how much Urban Company loses on each InstaHelp order completed. In Q3, the figure was ₹381. The loss per order grew as the order volume grew. CEO Abhiraj Singh Bhal acknowledged this directly in the annual shareholder letter:

losses will remain elevated because the company is prioritising densification, broader micro-market coverage, and partner onboarding. This is intentional burn, not operational failure.

Urban Company (NSE: URBANCO) is a listed company with ₹2,021 crore in cash on hand and an adjusted EBITDA break-even target of Q3 FY28. It can sustain this for longer than most startups can.

This is where the Snabbit’s context from last week’s issue matters.

Two competing theories of the home services market are now running with real capital behind each of them.

Urban Company is buying density through subsidies, paying ₹447 per order to place workers in proximity until the network self-sustains. Snabbit built density structurally, through tight micromarket clustering that produces a 247-metre median inter-job travel distance and a 40,000-orders-a-day throughput from 15,000 workers.

One model purchases density. The other engineered it. The next 18 months will determine which unit economics survive contact with scale.

SIGNAL 03 · INFRASTRUCTURE

Pixxel and Sarvam are building a satellite that runs AI in orbit. India’s infrastructure thesis just moved off the ground.

On May 4, Bengaluru-based Pixxel announced a partnership with Sarvam to build what they describe as India’s first orbital data centre satellite.

Most coverage treated this as a space story. It is an AI infrastructure story.

The satellite, named Pathfinder, weighs approximately 200 kilograms and is scheduled to reach orbit by Q4 2026.

The big difference: Conventional satellites collect imagery and transmit it to ground stations, where it waits for processing. The gap between capture and actionable insight can run to hours. Pathfinder processes data where it is collected. For defence, agriculture, disaster response, and infrastructure monitoring, that difference is real-time.

Sarvam’s models will run on India-built hardware, processing India’s earth observation data, without passing through a foreign cloud (AWS/Google) at any stage.

That’s sovereign AI infrastructure at the hardware layer, and a rare feat of extending the sovereign AI stack into space.

Consider who is doing this. Pixxel has already demonstrated that a Bengaluru company can build and operate hyperspectral imaging satellites that compete with global peers. Sarvam has built a full-stack AI platform for Indian languages. This partnership is the convergence of two companies that each chose a globally competitive infrastructure category and built seriously in it, without waiting for a foreign incumbent to validate the market first.

Pathfinder is a proof-of-concept mission. But it is three quarters from orbit.

PIVOTAL TAKE

Three signals. One fault line. India is pricing two very different kinds of company in two very different ways right now.

The proposed Unacademy outcome, Urban Company’s Q4 results, and the Pixxel-Sarvam satellite. On the surface, three unrelated events from three different sectors. Read together, they describe something specific about where the Indian startup market is in May 2026.

The proposed Unacademy restructuring is a marker for the end of one theory of company building: raise at whatever valuation the market bears, acquire users fast, and assume the exit will arrive before the cash runs out. That theory produced extraordinary paper returns between 2019 and 2021. It is now producing share-swap deals at 90% markdowns, where founders stay on and institutional investors trade one illiquid position for another.

Urban Company’s InstaHelp burn is the same theory, restated for a listed company with deeper pockets. Subsidise orders to buy density, assume unit economics improve at scale, and outlast the competition. It may work. But the loss per order moved in the wrong direction as volume grew last quarter.

And then there is Pixxel-Sarvam. Two companies that chose categories where the defensibility is technical, not financial. No amount of subsidised orders builds a hyperspectral satellite constellation. No marketing budget trains a sovereign AI stack that operates in orbit. These moats are built through engineering depth. Capital can fund them. Capital alone cannot replicate them.

The Indian startup ecosystem in 2026 is bifurcating along a clear line. On one side: companies where capital creates the moat. On the other: companies where technical depth does. The first group is facing its reckoning. The second is being noticed by exactly the kind of acquirer, partner, or investor that moves markets.

If you are deciding where to look for the next three years, that fault line is the most useful frame available right now.

— Pivotal Research

Last week, three things happened in India's startup ecosystem that deserve more attention than they got. One of them — the Portkey acquisition — happened on the last day of April and barely made the weekend news cycle. That's exactly the kind of thing The Signal is here for.

Let's get into it.

SIGNAL 01 · ACQUISITION

Portkey just sold to Palo Alto Networks. Here's what Indian founders should take from it — and what they're probably missing.

On April 30, Palo Alto Networks announced it would acquire Portkey — a Bengaluru-based AI infrastructure startup that most people in India's startup ecosystem had heard of but never properly tracked.

The deal terms weren't disclosed. What was disclosed: Portkey processes trillions of tokens per month. It had raised $18 million total — a $3 million seed from Lightspeed in 2023, and a $15 million Series A from Elevation Capital and Lightspeed just months before the acquisition. It was founded by Rohit Garg and Ayush Gupta, two engineers who spent years building developer tooling before they decided that the real gap in enterprise AI wasn't the models. It was the plumbing around them.

That bet paid off.

Palo Alto Networks, a $120 billion cybersecurity company, has built a reputation for acquiring before the category is obvious — and integrating well after. Portkey becomes the core of Prisma AIRS, PANW's AI security platform, acting as what the company calls "the central nervous system" for enterprise AI agent governance. In plain terms: a control layer that sits between a company's AI agents and everything else — models, APIs, data, external systems — and decides what gets through and what doesn't. The problem Portkey was solving is genuinely important. Autonomous AI agents are increasingly acting as "privileged insiders" inside enterprise systems — making decisions, executing workflows, accessing sensitive data. Most enterprises have almost no visibility into what their agents are doing. Portkey gave them that visibility.

WHY IT MATTERS FOR YOUR PORTFOLIO AND PIPELINE

Two things.

First, this is a clean early exit in a market starved of them. Since 2020, India has seen $2 billion in ESOP buybacks — most of it driven by secondary transactions, not acquisition events. A clean international M&A exit, even at an undisclosed price, is valuable as a data point. It signals that global acquirers are watching Indian AI infrastructure companies seriously — not as cheap engineering talent, but as category-defining product builders.

Second, Portkey's category — AI infrastructure, specifically AI agent governance — is almost completely unaddressed by Indian VCs. There are maybe five companies in India seriously building in this space. Elevation Capital, which co-led the Series A, may have understood this better than most. The question every early-stage investor should be sitting with right now: what other unglamorous infrastructure categories are being quietly built in India that a $100B+ acquirer would pay real money for in 18 months?

The acquisitions that look inevitable in hindsight almost never look inevitable while they're happening.

SIGNAL 02 · FUNDING

A startup founded in 2024 just hit a $360M valuation. The one number that explains how.

This one is about Snabbit. Specifically, it is about a single metric that its founder dropped in an interview this week that deserves to live outside a TechCrunch deal announcement.

The number is 247.

That is the median distance, in metres, that a Snabbit service professional travels between two consecutive jobs. Aayush Agarwal, the company's founder, described it as "the single number that explains why we just closed our Series D."

To understand why 247 metres matters, you need to understand the core problem with India's home services market. Urban Company has been at this for a decade. It is publicly listed. It has more than a million monthly bookings. And it has never been able to solve the unit economics problem at scale — because the underlying model sends professionals across cities, not across streets. Long travel times inflate cost per order, reduce daily job completions, and make the economics of gig work brutal for the workers themselves.

Snabbit's answer was to not operate that way. It clusters supply and demand within tight neighbourhood micromarkets — 140 of them currently, in five cities. A professional takes a job in one building, walks 247 metres, takes the next job. The result: 40,000 orders per day from 15,000 workers — a throughput ratio nearly double what comparable platforms achieve. The company went from 400 daily orders to 40,000 in a single financial year. That is not a rounding error. That is a 100x in 12 months.

The $56 million Series D, co-led by Susquehanna, Mirae Asset's Unicorn Growth Fund, and Bertelsmann, values Snabbit at $360 million. It was worth $180 million six months ago. Total capital raised is now $112 million across four rounds in roughly 15 months — a fundraising pace that communicates one thing clearly: the investors backing this company are not waiting to see if the model works. They already believe it does.

WHY IT MATTERS FOR INVESTORS

The home services race is real, it is happening now, and it has clear implications for capital allocation. Urban Company leads with 6.5 million monthly active users; Pronto is at 2.7 million; Snabbit at 1.2 million. In terms of app downloads in March, the ranking reversed: Pronto at 43%, Urban Company at 31%, Snabbit at 26%. A company with half Pronto's users is capturing nearly the same share of new demand. That gap closes. The question is who consolidates this category.

The more interesting signal, though, is what Snabbit's density model implies for adjacent markets. The same micromarket infrastructure that powers home cleaning can be extended into home cooking — which Snabbit has quietly been piloting for four months. Then home repair. Then elder care. A high-frequency, high-trust relationship with a household is one of the most defensible positions in consumer tech. Whoever earns it at scale in India's cities will be very difficult to displace.

That is what the investors backing Snabbit are paying $360 million for. Not cleaning services. The relationship.

SIGNAL 03 · DATA

India raised $865M in April. The headline is boring. The detail underneath it is not.

April's funding numbers landed this week. Indian startups raised $865 million across 92 deals — down from $948 million in March and comfortably below the $2 billion spike in February (which was flattering because Neysa's $1.2 billion AI round sat inside a single month). Year-on-year, April 2026 was actually better than April 2025 — up from $745 million — but worse than April 2024's $1.03 billion.

These numbers, taken at face value, tell you very little. But the composition tells you a great deal.

Here is what the April data actually shows:

Late-stage funding is hollowing out. The month had just one deal above $100 million — KreditBee's $280 million Series E. Strip that out and growth-stage activity was quiet. That is now a pattern, not an anomaly. Late-stage deal count in India has been declining for three consecutive quarters.

Early-stage is doing the opposite. 69 early-stage deals worth $321 million in a single month. Series A — 23 deals, $220 million. Pre-Series A — 10 deals, $30 million. The base of the pyramid is active. Seed round count was also healthy at 31 deals. This is the market saying: we will back early bets, but we are much more selective about follow-on capital.

The sector split is unambiguous. Fintech dominated with $363 million — 42% of all capital deployed. AI came second with $139 million across 17 deals. The gap between first and second is enormous. India's venture market is not, in aggregate, an AI market yet. The narrative has outrun the capital allocation. In the US, AI has captured more than 40% of all VC deployment. In India in April, it was 16%.

WHY IT MATTERS

For founders: the fundraising environment has bifurcated. If you are at seed or Series A with genuine traction, capital is available — 23 Series A deals in a month is not a quiet market. If you are at Series B or later and your numbers are not clean, you are in the toughest environment since 2023.

For investors: the AI-fintech convergence is the clearest thematic signal in April's data. Oolka — AI-native credit underwriting, $14M Series A, Accel led, Meesho's co-founders investing personally — is the archetype. Companies that sit at the intersection of India's mature fintech infrastructure and genuine AI capability are getting funded quickly. The ones that are simply "AI-enabled" fintech are not. The market is not contracting. It is bifurcating.

PIVOTAL TAKE

The Portkey acquisition and India's April funding data, read together, reveal something about this moment in the Indian startup ecosystem that most analysis misses.

Indian founders have spent the last three years being told that building AI infrastructure is not the right strategy — that India should apply AI to solve Indian problems, not build foundational layers. That thesis is not wrong. But it is incomplete.

Portkey just sold to a $120 billion American company because it built infrastructure — specifically, the control plane for AI agents — that global enterprises needed and that no American startup had built with the same clarity of purpose. The founders did not build a product for the Indian market. They built a product for the world, from India, using the engineering density and cost structure that India makes possible.

This is a different kind of Indian startup story from the one the ecosystem usually tells about itself. It is not "applying global technology to India's unique problems." It is "building global technology from India, using India's structural advantages."

Neysa is another version of this story — $1.2 billion for GPU cloud infrastructure. Sarvam is attempting it with sovereign AI. E2E Networks has been doing it quietly in cloud compute for years. The question worth sitting with: how many more Portkeys are being built right now in Bengaluru, Hyderabad, and Chennai — unglamorous, deeply technical, not trending on LinkedIn — that a global acquirer with a $100 billion market cap would pay serious money for in 2027?

That is where I would be looking.

\ Pivotal Research

If you found one useful thing in this issue, do consider subscribing and forward it to one investor or founder who would appreciate it. That is how The Signal grows.

Want deeper research on any of the signals above? Reach out to us at george [at] pivotalresearch [dot] in.

Next issue: Monday, May 11. See you then.

The Signal · by Pivotal Research

Bengaluru, India

Welcome to the first issue of The Signal — a weekly brief for investors and founders who need the week's most important market moves without spending two hours finding them. The format is simple: three signals from the past week or two, each with context on what to actually do with the information. Then a Pivotal Take — our views, signed. The whole thing is written to be read in six minutes, ideally with coffee before the rest of your inbox demands your attention.

Let's get into it.

SIGNAL 01 · FUNDING

The ex-Blinkit founders who convinced HUL and ITC to hand over their keys

GobbleCube just raised $15 million in a Series A led by Susquehanna Venture Capital — and the deal tells you something specific about where enterprise AI is going in India, and where it is definitely not going.

The company was founded in late 2022 by three former Blinkit executives: Manas Gupta, Srikumar Nair, and Nitesh Jindal. They built an AI platform that helps consumer brands manage the chaos of selling across quick-commerce marketplaces — Blinkit, Zepto, Swiggy Instamart — where demand generation, supply chain, and performance marketing are all moving simultaneously and often against each other.

The numbers are credible: 400+ brands on the platform, embedded inside the systems of 45 large CPG enterprises including HUL, ITC, Tata Consumer, Godrej, and L'Oréal. Ten-times revenue growth in the past year. InfoEdge Ventures and Kae Capital both followed on from their pre-Series A seven months ago — the clearest possible signal that the existing investors like what they're seeing.

The valuation: $68 million post-money on $20 million total raised. That is roughly 4–5x revenue for a company that is growing quickly but has not yet proved international traction.

WHY IT MATTERS FOR INVESTORS

There are two things worth unpacking here.

The first is the Blinkit founder effect. All three co-founders spent years at the centre of India's quick-commerce infrastructure. They didn't guess that brands would struggle to manage hyperlocal digital commerce — they watched it happen from the inside and then built the fix. This is the pattern that Indian enterprise AI is beginning to reliably produce: domain-specific founders, with deep operator experience, solving the specific problem they spent years watching go unsolved. This pattern is worth more attention than the generalist AI platform story.

The second is what the cap table reveals. The three co-founders now hold roughly 9.3% each (a combined 28%), after collectively diluting more than 70% across rounds. For a Series A, that is notable dilution. It suggests either aggressive early fundraising or structured arrangements that weren't fully disclosed. Not necessarily a red flag, but a question worth asking before a co-investment.

FOR FOUNDERS FUNDRAISING IN THE NEXT 90 DAYS

GobbleCube's Series A happened in 18 months from commercial launch, on the back of 10x revenue growth and a roster of brand names that any enterprise salesperson would envy. The deal memo writes itself when the customer list looks like that. Before your next investor meeting, ask yourself: what is the equivalent of "HUL, ITC, and Tata Consumer are live on our platform" for your sector? That's your lead slide.

SIGNAL 02 · CAPITAL MARKETS

Peak XV's $1.3B says the consensus India AI narrative has officially arrived — which means the real opportunity is somewhere else

In February, Peak XV Partners — India's most active institutional VC, formerly Sequoia India — closed $1.3 billion across three new funds: an India Seed fund, an India Venture fund, and an APAC vehicle. This is the firm's first major independent fundraise since its 2023 separation from Sequoia Capital. It now manages over $10 billion across 16 funds.

The headline reads as straightforwardly bullish on India. It is. But a closer reading is more interesting.

The $1.3B raise is roughly half the size of Peak XV's prior $2.85 billion fund from 2021 — a fund that was itself later resized down 16% to $2.4 billion in 2024, when Peak XV cited "elevated public market valuations in India." The new fund is smaller, explicitly focused on AI, and designed to be deployed over two to three years. Managing Director Shailendra Singh told TechCrunch: "We are not trying to match rivals dollar-for-dollar."

That rival is General Catalyst, which announced at the same event — Delhi's AI Impact Summit — that it plans to invest $5 billion in India over five years. That is an escalation. Peak XV's response was, in effect: we know this market better than you do, and we would rather deploy $1.3 billion well than $5 billion badly.

WHY IT MATTERS FOR INVESTORS

Peak XV closed 16 deals in Q1 2026 — more than any other VC in India. That pace, combined with a focused fund of this size, means they will write a lot of cheques in the $5–20 million range over the next 24 months. For angel investors who want to co-invest alongside institutional rounds: the firm with the most deal flow in India just told you it is writing smaller cheques, more frequently, with an AI and fintech focus. That is your map.

It also signals something about where the crowded trade is. When India's largest fund announces a $1.3B AI-focused raise while General Catalyst is committing $5B and OpenAI, Anthropic, and Google are all at the same Delhi summit — the AI consensus trade in India is, as of this week, fully articulated. The companies that benefit most from this capital are probably not the ones you'd expect. It is the infrastructure plays: compute, data, tooling, compliance. The application layer will get crowded fast.

FOR FOUNDERS FUNDRAISING IN THE NEXT 90 DAYS

Peak XV's stated cheque range is $5M–$20M for India Venture, up to $5M for Seed. If you are at revenue and looking for a Series A co-lead or a seed check from a fund that just told its LPs it would be aggressive on AI — and has the capital to prove it — this is the moment to get in front of their team. They deployed into 16 companies in one quarter. That's one new investment every six days.

SIGNAL 03 · INFRASTRUCTURE

The $80 million deal that no VC newsletter covered — and why that's a mistake

On April 16, Polaris Smart Metering announced $80 million in financing from British International Investment — the UK's development finance institution — for its subsidiary Hooghly Smart Metering Private Limited. The capital is earmarked for deploying over 2.2 million smart meters in West Bengal.

You probably did not read about this in your regular startup feeds. Most of them missed it. That is worth examining.

Polaris is backed by I Squared Capital, a global infrastructure private equity firm with $40+ billion in assets. It is not a seed-stage startup. It builds, deploys, and operates smart metering infrastructure under long-term government contracts — currently with a $1.1 billion order book across Uttar Pradesh, West Bengal, and Manipur, covering 7.5 million electric meters and 1.6 million gas meters.

The context matters: India has committed to installing 250 million smart meters by 2027 under its Revamped Distribution Sector Scheme (RDSS). That is a government mandate with a deadline. Polaris has contracts. BII has money. The $80 million is project financing — tied to specific deployment milestones, not a venture bet on future revenue.

WHY IT MATTERS FOR INVESTORS

Two things.

First, BII's India portfolio now includes Ather Energy, KreditBee, Ecofy, and now Polaris. The UK's development finance institution is constructing an India infrastructure playbook — EV, fintech, climate, energy. If you are mapping where patient international capital is moving in India, this is a data point worth marking.

Second: climate and energy infrastructure is staging a quiet re-rating in India. The deals are not getting the coverage that consumer and AI rounds get, partly because the structures are more complex (project finance, debt, government contracts) and partly because the founders are operators, not storytellers. But the capital is moving. $80 million from BII, $80 million from the EV-focused Drivn in Q1, $670 million into Arya.ag for agritech infrastructure — these are not venture outcomes, but they are the kind of capital flows that reshape an ecosystem over five to ten years.

FOR FOUNDERS IN CLIMATE / ENERGY / DEEPTECH

The narrative is shifting. India's government has committed to infrastructure targets that require private capital to hit. The companies that can credibly position themselves inside those government contracts, not pitching to government, but as implementation infrastructure for mandated programmes, have a different fundraising conversation available to them. BII, Norfund, IFC, and other development finance institutions are the natural capital sources for this cohort. They are slower than VCs but they write larger, longer-duration cheques. Worth a conversation if your sector has a policy tailwind.

PIVOTAL TAKE

India's AI quarter is producing two distinct classes of company, and the market has not fully priced the difference.

The first class is the infrastructure layer: compute, data pipelines, compliance tooling, model fine-tuning for Indian languages and regulatory contexts. These companies are unglamorous, capital-intensive, and have limited viral potential. They also have durable moats and the kind of enterprise customers who sign multi-year contracts. Neysa — which raised ₹10,000 crore ($1.2 billion) in India's largest-ever AI round in Q1 — is the clearest example.

The second class is the application layer: AI wrappers over vertical workflows — commerce, HR, legal, finance. GobbleCube sits here. The growth metrics are exciting, the fundraising is competitive, and the customer names read well in a pitch deck. The risk is also higher: switching costs are lower than they appear, and every large incumbent has an AI roadmap that eventually covers the same ground.

Peak XV's fund announcement and General Catalyst's $5 billion commitment both signal that the consensus trade is the application layer. Everyone is looking at the same thing.

The less crowded conversation, the one where you arrive before the headline funds, is at the infrastructure layer, in climate and energy, and in the Tier 2 city deal formation that Q1 2026 data shows is now producing nearly 35% of India's deal volume. The map has quietly redrawn itself. Most investors are still reading last year's version.

That is the bet The Signal is making with its coverage. We will follow the capital before the narrative catches up.

\ Pivotal Research

If you found one useful thing in this issue, do consider subscribing and forward it to one investor or founder who would appreciate it. That is how The Signal grows.

Want deeper research on any of the signals above? Reach out to us at george [at] pivotalresearch [dot] in.

Next issue: Monday, May 4. See you then.

The Signal · by Pivotal Research
Bengaluru, India