Blog

Pronto is a home services app that sends trained, background-verified workers to homes for chores like cleaning, laundry, utensil washing, meal prep, car washing, and gardening.

The latest Pronto home services funding update is a big one: the startup has extended its Series B to $45 million with a fresh $20 million investment from Lachy Groom, pushing its valuation to $200 million. The pitch to customers is simple, but the problem underneath it isn’t. Household labor in India still runs on informal referrals, uneven quality, and the constant risk of no-shows. Founded in 2025 by Anjali Sardana, Pronto is trying to turn that mess into something more predictable, faster, and easier to book.

What is Pronto home services and how does it work?

Pronto works like quick commerce for household chores. A customer opens the app, picks from a menu of 18 services, adds one or more tasks to a cart, chooses whether they want help now, later, or on a recurring basis, and pays in-app. For instant bookings, a trained worker can arrive in about 15 minutes in supported micromarkets.

The service catalog is wider than a basic “maid on demand” pitch suggests. It includes hourly help, bathroom cleaning, fridge cleaning, dusting and wiping, sweeping and mopping, utensils, kitchen prep, ironing and folding, laundry, balcony cleaning, fan cleaning, packing or unpacking, and express cleans before or after a party. That matters. The app isn’t just replacing one domestic worker relationship — it’s unbundling chores into bookable tasks.

Pronto also removes a lot of the awkward manual coordination that comes with direct hiring. Users can stack multiple tasks into one visit and set a daily or weekly cadence. They can also pause or reschedule from the app and see pricing upfront instead of negotiating each job separately. If a worker can’t make it, the system automatically reassigns the booking. That’s a very different experience from waiting around for a person who may or may not show up.

And that’s really the product. Not just “cleaning.” It’s reliability software wrapped around labor operations. The app handles booking logic and assignment. Scheduling and payments are baked in. Pronto handles training and verification on the supply side. In a category that has historically run on WhatsApp messages, cash, and neighbor recommendations, that shift is bigger than it sounds.

Who founded Pronto and why are investors backing it?

The founding story

Pronto was founded in 2025 by Anjali Sardana, who serves as founder and CEO. The company started with a blunt thesis: households need dependable help for repetitive chores, and workers need structured shifts and steadier income instead of informal arrangements that can vanish overnight. The startup was initially built out of Gurugram and later shifted its headquarters to Bengaluru to strengthen product and engineering as the instant home-services race heated up.

The move to Bengaluru is telling. Sardana isn’t treating Pronto like a local services marketplace with a bit of software on top. She’s building it more like an operations-heavy consumer tech company. Engineering, product, and data science play a bigger role, while support and some operations remain in Gurugram. That’s the kind of setup you’d expect if speed, matching, and utilization are core to the business.

Why Anjali Sardana had market fit

Before starting Pronto, Sardana worked at Bain Capital and at 8VC. That doesn’t make someone automatically good at running city-by-city labor operations — nothing does. But it does mean she came in with exposure to how high-growth companies are financed, evaluated, and scaled. Lachy Groom’s bet also seems to have been heavily founder-driven; he moved quickly after meeting Sardana and focused on execution quality as much as category size.

She’s also been unusually explicit about the ambition. “Organizing informal labor is going to be one of the defining shifts of the next decade in services. The longer-term vision of Pronto is to be the world’s largest labor organization platform,” Sardana said. That’s a huge claim. It’s also a cleaner description of the business than “Uber for maids,” which frankly undersells what Pronto is trying to build.

Traction, fundraising, and competition

The operating numbers are moving fast. Daily bookings have climbed to 26,000 from around 18,000 since the first close of the Series B, which works out to nearly 780,000 monthly bookings. Pronto’s professional workforce has grown from 1,440 in January to 6,500 over the past 4 months. The platform is running at more than 65% utilization. That’s a useful signal in a labor marketplace where idle supply can wreck economics.

This latest round came in 2 parts. Pronto first closed a $25 million tranche of its Series B and then added Groom’s fresh $20 million, taking the round to $45 million and the valuation to $200 million — roughly double in a short span. Existing backers include General Catalyst, Bain Capital Ventures, Glade Brook Capital Partners, Epiq Capital, and Groom, and total capital raised now stands at about $60 million.

Competition is getting aggressive, fast. Urban Company said its InstaHelp service crossed 1 million bookings in March, while Snabbit also reported 1 million orders around the same period and then raised a $56 million Series D last month. Pronto’s edge, at least on paper, is its push into instant fulfillment and dense micromarkets. It’s also built around repeat household usage rather than occasional repairs or beauty appointments. The real incumbent, though, isn’t another app. It’s the offline habit of hiring through personal networks and dealing with all the mess that comes with it.

Why does the Pronto funding round matter?

A funding round like this isn’t just about adding cash to the balance sheet. For Pronto, it buys time to deepen city density — which matters more than flashy expansion maps in a business where response times, worker utilization, and repeat behavior are everything. The company will spend the next 6 months going deeper in existing cities, not just spraying itself across new ones.

It also gives Pronto room to widen the service basket without losing focus on the chores people book most often. The startup has already expanded into categories like car washing, gardening, and home cooks in select Bengaluru micromarkets. If that works, Pronto gets a bigger share of household spend from the same user instead of constantly paying to reacquire customers for one narrow service.

There’s also a signal here for investors. Groom didn’t back Pronto because home cleaning is a sexy category. He backed it because this kind of operationally ugly business can become very defensible if a startup gets supply organization, quality control, and fulfillment density right. That’s why the valuation jump matters. It says investors think Pronto is executing faster than most people expected.

How big is the market for instant home services in India?

It’s a big market, and it’s still barely digitized. Redseer pegged India’s home services market at ₹5.1 trillion to ₹5.2 trillion in FY2025, with the sector projected to grow to ₹8.4 trillion to ₹8.6 trillion by FY2030 at a 10% to 11% CAGR. Online penetration is still below 1%. That tells you how much of the category remains stuck in offline, fragmented workflows.

That’s why startups like Pronto, Snabbit, and Urban Company are attracting so much capital. This isn’t just another convenience app trend. It sits at the intersection of rising urban incomes and more dual-income households. It also reflects demand for predictable service quality and quick-commerce-style expectations around speed. Put differently: consumers have gotten used to tapping a button and getting groceries in minutes. Now they expect the same thing from household labor.

What to watch next for Pronto home services

Pronto home services has gone from a very young startup to a heavily watched category player in less than a year. The next test isn’t whether it can raise more money — it probably can. It’s whether the company can keep service quality tight while adding workers, deepening cities, and rolling more chores into the same app.

Read how Alphadroid raised ₹36 Cr in a pre-Series A round led by Alkemi Growth Capital to scale AI-powered service robots and RaaS automation for hotels, restaurants, healthcare, and logistics businesses across India.

FAQ

– What is the latest funding round for Pronto? 

 Pronto has extended its Series B round to $45 million after taking a fresh $20 million investment from Lachy Groom. That extension values the company at $200 million and brings total funding raised to about $60 million so far.

– How does the Pronto app work for home services? 

 Pronto lets users book chores through an app by selecting tasks, choosing instant, scheduled, or recurring slots, and paying digitally. The service menu spans 18 categories, and users can bundle multiple jobs into one visit instead of coordinating separate workers for each task.

– Who is Pronto founder Anjali Sardana? 

 Anjali Sardana is the founder and CEO of Pronto, which she launched in 2025. Before starting the company, she worked at Bain Capital and 8VC, giving her early exposure to venture investing and high-growth startup execution.

– Is Pronto part of quick commerce or the home services market? 

 It’s really both. Pronto operates in the home services market, but its operating model borrows heavily from quick commerce by trying to deliver trained household help in 10 to 15 minutes inside dense micromarkets.

Ethos is a London startup building an AI expert network that helps companies find and vet specialists faster than LinkedIn or old-school expert-call firms. On May 6, 2026, it raised a $22.75 million Series A led by a16z. The problem it’s chasing is obvious once you see it: job titles are a lousy proxy for real capability when a client needs someone with weirdly specific experience. Ethos was founded in 2024 by James Lo and Daniel Mankowitz, and its bet is that voice data can capture expertise a résumé never will.

What is Ethos and how does the AI expert network work?

Here’s the basic workflow. A customer writes a natural-language brief describing the exact kind of person they need. Ethos searches across 500 million-plus profiles and reaches prospects through a network with more than 1 million warm connections. It screens people with an AI voice agent, then turns interview transcripts into research outputs using public filings, commercial datasets, and more than 5 million academic papers.

The expert side is where Ethos tries to look different. Instead of asking people to fill out a static form and hope their title says enough, the platform runs voice-based onboarding to pull out sub-specialties and adjacent experience. It also captures domain nuance that usually gets lost. That’s why the company says it can handle odd but realistic requests — like operators from finance-automation startups backed by top-tier investors, or doctors in a narrow field who also publish research and understand drug development.

A lot of manual work disappears if that works. Traditional expert networks still rely on recruiters doing profile searches and sending outreach. They also screen for fit, book calls, check compliance, and package notes for the client. Ethos is trying to compress all of that into one workflow so the buyer can move from “I need this exact kind of person” to a usable conversation and written output without juggling five tools and a pile of coordinators.

There’s more. Ethos says its AI interviewer can run long-form conversations of up to 60 minutes, and some experts can join without the usual scheduling back-and-forth. That makes the product feel less like a directory and more like a research system with expert sourcing built in.

Who founded Ethos, and how is the AI expert network growing?

The founding story

Ethos was founded in London in 2024 by CEO James Lo and CTO Daniel J. Mankowitz. Lo came at the idea from the labor side — he has said he wanted to create better economic and work opportunities for people. Mankowitz saw the economy more like a knowledge graph of people, companies, and products, where better matching could create more value. Put differently, both founders thought the market cared too much about titles and not enough about actual capability.

Why the founders fit this market

Lo’s background is unusually practical for this kind of company. Before Ethos, he worked at McKinsey and later at SoftBank, where he was involved in transformation work around companies including WeWork and Arm. That matters because expert networks are really a service business disguised as software. You need to understand buyers, operations, and high-stakes research workflows.

Mankowitz brings the harder technical edge. He worked as an AI researcher at DeepMind on YouTube video compression, Gemini, and AlphaDev. AlphaDev is the project that discovered faster sorting algorithms, and Mankowitz was one of the named researchers on that work. That doesn’t automatically make someone good at enterprise software. But it does explain why Ethos treats interviews, matching, and expert data as machine-readable problems instead of staffing problems.

Early traction and economics

Ethos isn’t naming customers yet, but the product is already used by top hedge funds, private equity firms, leading foundational AI labs, and enterprise consulting clients. The network is growing fast too, with roughly 35,000 people joining each week through an invite-led model. Ethos takes 30% or more as a per-project fee, depending on the work, and is already tracking toward eight-figure annualized revenue. The whole team is still just 8 people. Probably both.

Funding details

The new round is a $22.75 million Series A led by a16z, with General Catalyst, XTX Markets, Evantic Capital, and Common Magic also participating. The raise was announced on May 6, 2026. Ethos hasn’t publicly laid out a detailed capital plan, but the timing is clear enough: it now has the money to keep building product while scaling the network behind it.

Where Ethos sits against incumbents and newer tools

The incumbents are easy to name: LinkedIn for discovery, then expert networks like GLG, Third Bridge, and AlphaSights for paid access. Those businesses still do plenty of volume, but the old model depends heavily on titles, recruiter judgment, and manual coordination. Ethos’s pitch is that these are shallow signals, and that a structured voice interview produces a much richer profile of what someone actually knows.

The newer comparison set is messier. Listen Labs and Outset use AI moderators to run interviews at scale, mostly for customer or user research rather than expert-network sourcing. NewtonX mixes AI with custom recruiting for B2B research and expert work. Tegus leans hard into expert-call transcripts and investor research. Ethos is trying to splice pieces of all those categories together. It combines expert discovery and AI-led screening. It also folds in AI-led interviewing and synthesized research outputs in one stack. It layers in public signals from blogs, academic papers, and social links, which gives it a broader matching graph than a résumé-only system.

Why are investors betting on AI expert networks now?

a16z’s thesis here is pretty direct. The firm argues that legacy platforms expose only thin signals, while voice interactions can surface sub-specializations people rarely write down well. That matters even more in regulated or high-consequence settings, where accuracy, context, and compliance all matter.

Ethos also isn’t just selling to firms that want a better expert call tomorrow morning. It’s showing up at a moment when AI labs are spending real money to map human expertise, collect feedback from professionals, and build products for law, health, finance, and management work. That makes Ethos interesting to investors for a bigger reason: it could become part expert network, part data-collection layer, part research workflow for AI-native companies.

How big is the expert network market in 2026?

This isn’t some tiny corner of finance anymore. Inex One pegged the global expert-network industry at more than $2.5 billion in 2024 and around $3 billion in 2025, after years of steady expansion. That growth has come from the same basic pattern: more firms want primary research fast, and they’re willing to pay for real practitioners instead of relying only on published reports or generic survey panels.

The structural shift is buyer behavior. Expert networks used to be associated mostly with investors and consultants. Now the demand base is wider — product teams, enterprise strategy groups, and AI labs all want direct access to people who’ve actually done the work. Voice AI also changes the math. If screening and interviewing can happen faster, with better compliance controls and lower coordination costs, the old expert-call workflow starts to look pretty clunky. That’s the opening Ethos is chasing.

Final take on Ethos

Ethos is betting that the future of the AI expert network won’t look like a better directory. It’ll look like a full research engine built around voice, structured expertise, and fast synthesis. The funding gives the company room to prove that idea at a bigger scale, but the real test is still ahead: whether it can keep quality high and compliance tight as it expands into more specialized, regulated categories.

Read how BigEndian Semiconductors raised $6M in pre-Series A funding led by IAN Alpha Fund to commercialise its secure system-on-chips for surveillance, IoT, telecom, and enterprise hardware, targeting growing demand for specialised edge AI silicon with built-in security and production-ready design.

FAQ

– What funding did Ethos raise? 

 Ethos raised a $22.75 million Series A on May 6, 2026. a16z led the round, and General Catalyst, XTX Markets, Evantic Capital, and Common Magic also joined.

– How does Ethos work as an AI expert network? 

 Ethos lets companies describe the expertise they need in plain language, then uses AI to find, contact, screen, and interview matching experts. The platform searches more than 500 million profiles, uses over 1 million warm connections for outreach, and can turn interviews into research reports after the call.

– Who are the founders of Ethos? 

 Ethos was founded in 2024 by James Lo and Daniel J. Mankowitz. Lo previously worked at McKinsey and SoftBank, while Mankowitz came from DeepMind, where he worked on projects including Gemini and AlphaDev.

– Why is Ethos part of the expert network market? 

 Because its core business is still matching companies with qualified people for paid insight, research, and interviews. What makes it different is the AI layer: Ethos uses voice onboarding and automated interviews to modernize a market that industry estimates place at roughly $3 billion in 2025.

BigEndian Semiconductors designs secure system-on-chips for surveillance, telecom, IoT, and enterprise hardware. The Bengaluru startup has now raised $6 million in pre-Series A funding led by IAN Alpha Fund, a round that matters because buyers in these markets don’t just want more compute. They want silicon built for specific workloads, tighter security, and a realistic path from design to production. Founded in 2024, the company is led by co-founder and CEO Sunil Kumar and a team of chip veterans with experience across semiconductor design, embedded systems, and global business development.

The fresh capital will go into commercialising BigEndian’s first SoC and adding more product engineering muscle. It will also deepen ties with foundries, IP partners, and OEMs. That’s a practical use of money, not a vanity round. In semiconductors, the gap between “we designed it” and “customers can buy it” is where a lot of startups get stuck.

What does BigEndian Semiconductors actually build?

BigEndian Semiconductors is building fabless chips for vision-heavy and security-sensitive devices, starting with surveillance silicon and expanding toward secure Vision Edge AI architectures. In plain English, it designs the brains that can sit inside cameras and related edge devices, process data locally, and bake security into the hardware-software stack instead of bolting it on later.

Its product direction is unusually specific for such a young company. BigEndian’s first chipset is aimed at enterprise and consumer surveillance use cases, and the company has linked that work to “Project VASU,” a program tied to its Cadence partnership for CCTV and surveillance SoCs. Job descriptions and company material also point to adjacent vision applications like CCTVs and dash cams. That fits the broader push toward edge inference rather than shipping every video stream back to the cloud.

What does that mean for a customer? An OEM building cameras or other connected devices doesn’t just need a chip. It needs silicon and embedded software. It also needs security controls and a path to manufacturing that won’t blow up the schedule. BigEndian’s pitch is that it can handle more of that chain in one place, from architecture and system engineering to tape-out and downstream enablement, while working with outside fabrication partners because it’s a fabless company.

There’s one more point here. The startup has already completed tape-out of its first commercial chip. That doesn’t mean mass adoption is guaranteed. But it does mean BigEndian is past the concept stage and into the part of the business where product, manufacturing, and customer qualification get very real.

Who founded BigEndian Semiconductors?

BigEndian was founded in 2024 by Sunil Kumar, Renuka Prasad, Harpreet Wadhawan, Dinesh Annayya, Kanagaraju Ponnusamy, and Jansen Cheng. On the operating side, Kumar serves as CEO, Renuka Prasad leads system engineering, Kanak P heads engineering, Dinesh Annayya leads silicon engineering, Harpreet Wadhawan is CFO, and Jansen Cheng handles global business development. That’s a loaded founding bench for a company this early.

Why this team has market fit

Kumar’s background is a big part of the story. His profile ties him to Tata Tele, Airtel, Broadcom, Intel, ARM, Zenoti, and Forward Slash, and describes experience across telecom, IP switching, wireless broadband, and computer vision. He also has IIT Madras degrees and an MBA from IIM Bangalore. That mix matters because BigEndian isn’t building a pure research project. It’s trying to turn chip design into a sellable product business.

Renuka Prasad brings a different kind of credibility. He has more than 30 years of experience spanning semiconductor package design, testing, low-power electronics, FPGA-based design, embedded systems, and verification, with earlier roles at Logics, Cypress, Beceem, and Broadcom. If Kumar is the market-facing operator, Renuka looks like one of the people who helps keep the engineering honest.

Wadhawan rounds out the founding mix from the capital side. He has 28 years in investment banking across the US and India, including work at Barclays Capital and enterprise, software, and IP-heavy businesses. For a semiconductor startup, that’s not cosmetic. Hardware companies burn money differently, raise money differently, and negotiate partnerships differently.

Early execution signals

BigEndian’s clearest signal so far is execution speed. The company has taped out its first commercial chip, and Vertex’s Ben Mathias said the team got there in “record time.” Back in September 2024, the startup had about 18 employees; its LinkedIn page now lists a company size of 11-50. Still small.

But it’s enough to suggest a focused engineering build rather than a bloated one.

Fundraising details and where the money goes

IAN Alpha Fund led this new pre-Series A round of $6 million, with participation from Vertex Ventures SEA & India, IvyCap Ventures, and angel investors. Before this, BigEndian raised a $3 million seed round in September 2024 led by Vertex Ventures SEA & India. That takes total disclosed funding to $9 million in under 2 years.

Kumar put it bluntly: “Raising capital in semiconductors is never about the money alone.” He’s right. The round is meant to push the company from tape-out into commercialisation, expand product engineering, and strengthen relationships with foundries, IP providers, and OEMs. In chip startups, those partnerships are the product roadmap.

Competition and where BigEndian sits

BigEndian isn’t alone in India’s fabless chip wave. Vervesemi raised $10 million in February 2026, and C2i Semiconductors raised $15 million in a round led by Peak XV Partners earlier this year. But those companies are better seen as peers in the same financing trend than as exact one-to-one product rivals. BigEndian is more narrowly identified with secure surveillance and vision-edge silicon.

Its real competition also includes legacy imported surveillance chipsets and the messy workaround many OEMs know too well: using general-purpose silicon plus external security layers and custom integration work. BigEndian’s differentiation is the “security-by-design” pitch and its camera-first starting point. It has also already crossed tape-out. That’s likely what investors are paying for: not just design talent, but a team that can get from RTL to a manufacturable chip and then into customer programs.

Why BigEndian Semiconductors’ $6M round matters

A pre-Series A round after tape-out changes the conversation. Earlier capital is often about faith in founders, architecture, and timing. This round looks more like a bet on commercial execution. That’s a tougher test. And a more meaningful one.

For customers, this funding should make BigEndian a more credible supplier. OEMs and enterprise buyers don’t want a startup that can demo silicon but can’t support product engineering and software integration. They also need support through long manufacturing cycles. More capital gives the company room to do the boring work that actually matters: validation, partner management, and getting the chip into real designs.

For investors, the thesis is pretty clear. IAN Alpha Fund’s Rajnish Kapur said the chip business is shifting from “scale to specialisation,” and that fits BigEndian’s angle almost perfectly. This isn’t a broad consumer processor play. It’s a wager that domain-specific, secure silicon for edge workloads will matter more as AI spills out of data centres and into devices.

Is the India semiconductor market ready for more fabless chip startups?

Globally, the answer looks like yes. Deloitte expects semiconductor industry sales to reach $975 billion in 2026, with AI infrastructure doing much of the heavy lifting. It also says generative AI chips could account for roughly $500 billion of revenue in 2026, even while PCs, smartphones, and other non-AI markets stay softer. That split is exactly why more specialised chip design startups are getting a closer look.

India’s policy backdrop is getting stronger too. The Union Budget for FY2026-27 set aside ₹1,000 crore for India Semiconductor Mission 2.0. Under the Design Linked Incentive scheme, India was supporting 24 semiconductor design startups as of January 2026, and those startups had attracted nearly ₹430 crore in venture funding. The same programme says startups had completed 16 tape-outs. Of those, 6 chips were fabricated at advanced foundry nodes including 12 nm.

The manufacturing pipeline is no longer theoretical. India had already approved 10 semiconductor projects worth about ₹1.6 lakh crore, with pilot production underway at 4 units, and two more projects approved in early May 2026 pushed total approved investment to roughly ₹1.64 lakh crore across 12 units. That doesn’t suddenly make India self-sufficient in chips. But it does make the country a more believable home for startups that want to design silicon and then actually ship it.

Final take on BigEndian Semiconductors

BigEndian Semiconductors still has a lot to prove. Tape-out isn’t volume production. Funding isn’t customer adoption. Secure edge silicon is a hard business with long cycles and zero patience for mistakes.

But this round does mark a real step up. BigEndian now has capital, an experienced founding team, and a chip that has reached tape-out. It also has a market window that makes sense. The next thing to watch isn’t another announcement. It’s whether the company can turn this first SoC into design wins and then into shipments.

Read how Ecofy raised $15M from Mirova to expand rooftop solar and EV financing across India, helping households and small businesses access tailored green loans for clean energy and electric mobility assets.

FAQ

– What funding did BigEndian Semiconductors raise? 

 BigEndian Semiconductors raised $6 million in a pre-Series A round led by IAN Alpha Fund. Vertex Ventures SEA & India, IvyCap Ventures, and angel investors also joined, and the company had previously raised $3 million in seed funding in September 2024.

– How does BigEndian Semiconductors’ product work? 

 It designs secure SoCs for surveillance and other edge devices, with hardware and software built together from the start. The first chip is aimed at camera-heavy use cases, and the company is also working on Vision Edge AI designs so devices can run more AI tasks locally instead of depending entirely on the cloud.

– Who are the founders of BigEndian Semiconductors? 

 The startup was founded in 2024 by Sunil Kumar, Renuka Prasad, Harpreet Wadhawan, Dinesh Annayya, Kanagaraju Ponnusamy, and Jansen Cheng. Kumar brings experience from Intel, ARM, Broadcom, and Zenoti, while Renuka Prasad’s background includes Cypress, Beceem, and Broadcom.

– Why are investors backing Indian fabless semiconductor startups now? 

 Because the market is shifting toward specialised chips, and India is finally building both policy support and execution depth. In 2026, Deloitte projected global chip sales at $975 billion, while India’s DLI programme was already supporting 24 design startups and recording multiple startup tape-outs.

Ecofy is a Mumbai-based lender that finances rooftop solar systems and electric vehicles instead of plain-vanilla consumer credit. That matters because households, fleet operators, and small businesses buying these assets still run into a basic problem: the hardware is getting cheaper, but access to tailored financing is still patchy. Now Ecofy NBFC has raised $15 million from Mirova to expand that lending engine, with the money earmarked for residential and commercial rooftop solar plus electric 2- and 3-wheelers. Founded in 2022 by Rajashree Nambiar and Govind Sankaranarayanan, the company now serves more than 130,000 customers across 26 states and 500-plus cities.

What is Ecofy NBFC and how does it work?

Ecofy is a green-only NBFC that underwrites loans around the asset being bought — a rooftop solar system, an electric 2-wheeler, or an electric 3-wheeler — rather than forcing borrowers into generic personal-credit products. For a solar customer, the pitch is practical: finance the installation, structure the EMI to sit below the current electricity bill, and make the economics work from month 1 instead of after a long payback wait.

The rooftop solar product is more than a loan stub. Ecofy also wraps in insurance options and a partner-backed annual maintenance contract. The company says uptime can stay above 99%. It lends to homes and to commercial and industrial users, which matters because a kirana owner, a small workshop, and a housing society don’t buy solar in the same way or repay on the same rhythm.

There’s also a subsidy-aware layer to the experience. Ecofy’s solar financing can sit alongside the PM Surya Ghar process, so a household isn’t left juggling installer paperwork, subsidy steps, and a separate retail loan on its own. In Indian solar sales, paperwork is often where adoption slows down.

On the EV side, Ecofy finances electric 2-wheelers and 3-wheelers — the workhorse categories for last-mile mobility and small commercial income. Its products span roughly ₹1 lakh to ₹1.5 crore with tenures from 6 months to 5 years. This isn’t a one-size-fits-all app loan. It’s a credit stack built for everything from a single vehicle purchase to larger clean-energy assets.

Who founded Ecofy NBFC and what’s the backstory?

Built by career lenders

Ecofy was started in 2022 by Rajashree Nambiar and Govind Sankaranarayanan, who now serve as MD and CEO, and COO and whole-time director, respectively. The company calls itself India’s first green-only NBFC. That specialization is the whole thesis: build a lender whose balance sheet is dedicated to sustainable assets instead of treating solar and EV loans as side categories inside a much broader retail book.

Why the founders fit this market

Rajashree Nambiar didn’t come out of a climate lab or a startup incubator. She spent decades inside mainstream financial services, including as MD and CEO of Fullerton India Credit, after earlier leadership roles at IIFL Finance and a 22-year run at Standard Chartered, where she was head of retail products for India and South Asia. In plain English: she knows mass-market lending, distribution, risk, and operations. She also knows how to build credit businesses at scale.

Govind Sankaranarayanan brings the other half of the puzzle. Before Ecofy, he spent 27 years with Tata and served as group COO and CFO at Tata Capital, where he helped build a business that reached ₹65,000 crore in assets under management and ₹7,000 crore in revenue. That kind of experience matters when your startup isn’t chasing clicks. It’s trying to build a lending institution that can raise capital, price risk, and survive cycles.

Early traction, funding history, and the rivals getting closer

Ecofy’s latest update says it serves more than 130,000 customers across 26 states and over 500 cities. Separate company-profile data places its team in the 151-to-250 range and its headquarters in Worli, Mumbai. In March 2026, just weeks before the Mirova deal, Ecofy also raised ₹380.5 crore, or about $42 million, in an equity round led by British International Investment and Finnfund, with FMO and Eversource Capital participating.

That earlier round came with more signals about execution. At the time, Ecofy said it had 1.25 lakh customers and ₹1,400 crore in AUM. It also had over 23 bank and financial-institution partners, plus more than 100 OEM relationships. The new $15 million from Mirova — an affiliate of Natixis Investment Managers and making its fourth India deal under this mandate — is being used for onward lending into rooftop solar and electric mobility.

The direct challengers are small but real. Solfin has built a solar-focused lending model and said it crossed ₹100 crore in solar-loan disbursals within 9 months. Three Wheels United offers tech-enabled EV finance with up to 100% funding on 2- and 3-wheelers. Then come the bigger threats: Tata Capital and Mahindra Finance on the lending side, IREDA and REC as institutional price-setters downstream, and solar manufacturers like Tata Power Solar and Waaree moving closer to the customer by bundling finance with installation.

Ecofy’s edge, if it keeps one, is focus. A green-only lender can build underwriting, servicing, and partnerships around asset classes that generic lenders still treat as exceptions. That’s also what Mirova is backing when it talks about platforms with “scale, local reach, and measurable impact” — not just a startup story, but a distribution vehicle for climate capital.

Why does this Ecofy NBFC funding round matter?

This round matters because lending businesses don’t scale on branding alone. They scale on access to capital that matches the tenor of the loans they want to write. Ecofy said the fresh money will go into onward lending for rooftop solar and EVs. So this isn’t cosmetic balance-sheet money. It’s fuel for the loan book.

It also sharpens Ecofy’s identity at a moment when generalist lenders are drifting into green credit. If you’re an installer, OEM, or small business owner, the real question isn’t whether climate finance sounds noble. It’s whether the lender at the point of sale can approve fast, price sensibly, and understand the asset. Ecofy’s treasury head, Vivek Khandelwal, framed the Mirova partnership around making finance more “accessible” for households and small businesses.

There’s a second signal here. Mirova isn’t writing this check as a casual fintech bet. Its India activity sits inside a broader emerging-markets energy-transition mandate, so the investment says Ecofy has become useful as a local pipe between global climate capital and end borrowers who would never raise that money directly. That’s a more durable thesis than another consumer-lending app.

How big is the market for rooftop solar and EV finance in India?

The timing isn’t random. India’s rooftop solar market was valued at $2.52 billion in 2025 and is projected to reach $4.27 billion by 2034. This isn’t just consultant optimism. The operating data is moving too. India installed 7.1 GW of rooftop solar in 2025, up 122% from 2024, with residential users accounting for 76% of additions.

Policy is doing a lot of the heavy lifting. By December 2025, more than 2.08 million rooftop solar systems had been installed under PM Surya Ghar: Muft Bijli Yojana. That kind of subsidy-led demand surge is exactly why financing specialists matter: the bottleneck shifts from customer interest to credit availability, documentation, and disbursal speed.

The EV side is even more capital hungry. India’s electric transport sector attracted about ₹2.23 lakh crore, or $25.6 billion, in investment from 2020 to 2025 — but that was only 18% of the total funding estimated to be needed by 2030. So even with all the headlines around EV adoption, there’s still a giant financing gap sitting between policy ambition and actual vehicle ownership.

What should Ecofy NBFC prove next?

Ecofy NBFC now has money, momentum, and a clean story investors like. What it still has to prove is harder: can a specialist green lender keep underwriting discipline as competition tightens from banks, NBFCs, and embedded-finance players that already control customer acquisition?

The next test isn’t another headline round. It’s loan performance, repeat distribution partnerships, and whether Ecofy stays the preferred finance layer when a customer is actually ready to buy solar or an EV.

Read how General Autonomy raised ₹32 Cr to expand its humanoid robots and robot dogs, betting on autonomous machines built for labour-heavy industrial work and difficult real-world environments where traditional automation struggles.

FAQ

– What funding did Ecofy raise from Mirova? 

 Ecofy raised $15 million from Mirova in May 2026. The capital is meant for onward lending in residential and commercial rooftop solar, along with electric mobility categories including 2-wheelers and 3-wheelers.

– How does Ecofy finance rooftop solar and EV purchases? 

 Ecofy offers asset-linked loans instead of generic consumer credit. For rooftop solar, it structures EMIs around expected savings and offers insurance and maintenance support. It can also work alongside the PM Surya Ghar subsidy process. For EVs, it finances vehicles such as electric 2-wheelers and 3-wheelers with product terms tailored to the use case.

– Who founded Ecofy and what experience do they bring? 

 Ecofy was founded in 2022 by Rajashree Nambiar and Govind Sankaranarayanan. Nambiar previously led Fullerton India Credit and held senior roles at IIFL Finance and Standard Chartered, while Sankaranarayanan spent 27 years at Tata and helped build Tata Capital at scale.

– Is Ecofy a bank or an NBFC in India’s climate finance market? 

 Ecofy is an NBFC, not a bank, and it focuses only on green assets. That makes it unusual in India because its lending book is built around categories like rooftop solar and electric mobility rather than broad retail or SME credit.

General Autonomy is a Bengaluru robotics startup building humanoid robots and robot dogs for labour-heavy physical work. The latest General Autonomy funding round brings in about ₹32 crore, giving the company more room to keep building in a market where hardware is expensive, timelines are long, and India’s robotics supply chain still isn’t mature. Founded in 2023 by former ShareChat cofounders Farid Ahsan and Bhanu Pratap Singh, the company is still pre-revenue. This round matters because it’s less about scale today and more about buying time to turn ambitious prototypes into machines people will actually deploy.

What does General Autonomy build?

General Autonomy is building two kinds of legged robots under one broader robotics stack. The first is a humanoid platform called Atom 01 — a made-in-India, autonomous, untethered robot that the startup has shown publicly as a life-sized walking machine. The second is Param, a quadruped robot dog built for rougher environments where wheels struggle and fixed automation just won’t work.

The pitch is pretty clear: these aren’t one-trick robots built for a single motion on a factory line. General Autonomy is working on systems that can handle navigation and manipulation. Cleaning and tool use are part of the plan. That matters because a lot of work doesn’t happen in neat, structured cells. It happens in messy spaces with stairs, narrow passages, clutter, people, and changing conditions.

Param gives the clearest view of how the product works in practice. The robot dog weighs 35 kg and can carry a 20 kg payload. It runs at up to 3 m/s, climbs stairs, handles 45-degree slopes, and uses onboard sensors for autonomous navigation and tracking. Obstacle detection and fall recovery are built in. That makes it useful for inspection rounds, power plants, oil rigs, search-and-rescue work, and public safety sites where sending a person first is risky and sending a wheeled bot is pointless.

The humanoid side is the harder bet — and the more interesting one. A robot shaped for human environments can, in theory, work with existing tools, doors, workstations, and cleaning tasks without forcing customers to rebuild everything around the machine. That’s the logic behind Atom 01. It’s also why General Autonomy is trying to build a full robotics stack instead of just a flashy demo bot.

Who founded General Autonomy and what’s its track record?

The founding story

General Autonomy was founded in 2023 by Farid Ahsan and Bhanu Pratap Singh, both former ShareChat cofounders and IIT Kanpur alumni. They registered the company in May 2023, a few months after stepping back from day-to-day roles at ShareChat.

The shift from social media to robotics looks abrupt on paper. It’s not quite that random. Ahsan has said the company wants to automate labour-intensive workflows across industries and everyday settings, and the name itself points to a broader ambition: building physical systems that can operate across different tasks, not just one narrow workflow.

Why Ahsan and Singh fit this bet

Ahsan and Singh aren’t first-time founders learning the basics of company-building on the fly. They helped create ShareChat, one of India’s best-known consumer internet companies, after meeting at IIT Kanpur. That experience counts. Hiring, shipping, fundraising, surviving bad phases — they’ve done all of it before.

There’s also a more direct robotics thread here. Ahsan studied materials science at IIT Kanpur and had been interested in autonomous navigation for years. Singh came from the robotics club side of campus and later became the technical brain at ShareChat as CTO. So this isn’t just “successful founders pivot to deeptech because it’s trendy.” There’s some real founder-market fit underneath it.

Still, consumer internet and robotics are wildly different businesses. Hardware burns money faster. Failure is slower and more visible. And a robot doesn’t let you patch bugs as casually as an app. That’s what makes this second act bold rather than safe.

From ShareChat to a harder category

Before General Autonomy, the pair were best known for building ShareChat with Ankush Sachdeva after graduating in 2014. They’d already tried a bunch of product ideas as students before landing on the regional-language social platform that finally worked. That history matters because it shows pattern recognition and persistence, not just one lucky hit.

General Autonomy is a much tougher operating environment. The company isn’t generating revenue yet. It’s building hardware in India, where Ahsan has said the local supply chain for this kind of robotics work barely exists. That means more has to be built from scratch. Every prototype cycle gets costlier.

Traction, fundraising, and where it sits against rivals

The latest General Autonomy funding round totals about ₹32 crore, or roughly $3.3 million, and existing backers Elevation Capital and India Quotient led it. Blue Asva Varenya Fund, FBC Venture Partners, Spearhead Capital, and GIVA cofounder Ishendra Agarwal also joined the round. In February, the startup issued 1,776 Series Seed 2 CCPS at ₹1.68 lakh apiece, and both Ahsan and Singh put in ₹99.6 lakh each.

This comes after a ₹25 crore pre-seed round in 2023. The new round values the company at around ₹280 crore, up from roughly ₹200 crore in the previous financing. It’s a step up for a startup that’s still in heavy R&D mode and not yet commercial.

Ahsan hasn’t tried to sugarcoat the economics. “It’s a costly business to build robots,” he said, adding that the company is “just doubling down” on robot dogs and humanoid robots. The money will go into R&D and into speeding up the company’s robotics stack.

Early signals are small but real. General Autonomy has about 19 team members. In January, it was one of 10 deeptech startups picked by Startup India to showcase at National Startup Day 2026.

Competition is getting crowded. Perceptyne is pushing AI-driven semi-humanoid robots for industrial automation. Addverb is the giant domestic automation name and has moved beyond warehouse systems into quadrupeds and humanoids. iHub Robotics is pitching semi-humanoid service machines, while Invento Robotics and Sirena Technologies are part of the broader Indian robotics mix. General Autonomy’s angle is different: legged autonomy and a made-in-India build story. It’s also betting that affordable robots for messy environments will matter more than another fixed automation product.

Why does the General Autonomy funding round matter?

This round is basically a vote for technical ambition before commercial proof.

That sounds obvious. It isn’t. A lot of investors love AI software because it scales fast and breaks cheaply. Robotics is the opposite. It takes more capital, more time, and more patience. When existing investors like Elevation Capital and India Quotient lead again, they’re saying the team has shown enough progress to justify another long, expensive stretch of product development.

It also matters for the product roadmap. General Autonomy isn’t using the cash to grow sales teams or chase short-term revenue. It’s using it to keep building robot dogs and humanoids. The software-hardware stack underneath them is part of that. For customers, that raises the odds of getting a domestic alternative to imported robots that are often expensive, closed, or badly suited to Indian operating conditions.

There’s a deeper signal too. Investors aren’t just backing a founder brand. They’re backing the idea that physical AI in India may finally be investable before the market is fully mature.

Is India ready for humanoid and robot dog startups?

India isn’t a robotics powerhouse yet. But it’s moving faster than it used to.

The clearest sign is industrial robot adoption. India installed 8,510 industrial robots in 2023, up 59% year on year — the fastest growth rate in the world that year — and ranked 7th globally in annual installations. The country’s installed base reached 44,958 units in 2023. And the momentum carried into 2024, when installations moved past 9,000 units.

That’s the good news.

The tougher truth is scale. China installed 276,288 robots in 2023, about 32 times India’s number. So India is still early. Very early. But that’s exactly why founders and investors care. The headroom is huge, especially in manufacturing, logistics, ecommerce, infrastructure inspection, and industrial safety.

The humanoid category alone is starting to look meaningful. India’s humanoid robotics market is expected to reach about $149.4 million by 2030, growing at roughly 19.5% a year. That’s not a massive market yet, but it’s big enough to support serious early players if they can survive the R&D grind.

General Autonomy funding: what happens next?

The next test for General Autonomy funding won’t be another headline. It’ll be execution.

Can a 19-person team turn robot demos into durable products and can it build around India’s thin robotics supply chain instead of getting stuck because a critical component or manufacturing process isn’t local? Can it ship a machine that’s reliable enough for real customers, not just cool enough for social media?

That’s what to watch now.

Read how Moment Energy raised $40M in Series B funding to scale second-life EV battery storage systems across North America, helping commercial and industrial customers turn retired electric vehicle batteries into lower-cost energy storage infrastructure.

FAQ

– What is the latest General Autonomy funding round?  

 General Autonomy has raised about ₹32 crore in a seed round. Elevation Capital and India Quotient led the round, with participation from Blue Asva Varenya Fund, FBC Venture Partners, Spearhead Capital, and GIVA cofounder Ishendra Agarwal. The round values the startup at roughly ₹280 crore and follows a ₹25 crore pre-seed raise in 2023.

– What does General Autonomy actually build?  

 General Autonomy builds legged robots, not standard warehouse bots or fixed robotic arms. Its public prototypes include the humanoid Atom 01 and the quadruped Param, and its robotics stack is being developed for navigation and manipulation. Cleaning and tool handling are part of the plan across industrial and everyday environments.

– Who are the founders of General Autonomy?  

 General Autonomy was founded by Farid Ahsan and Bhanu Pratap Singh in 2023. Both are IIT Kanpur alumni and former ShareChat cofounders, which gives them a rare mix of startup-building experience and technical depth. Ahsan led operations at ShareChat, while Singh handled core technology as CTO.

– Is General Autonomy a humanoid robotics company or an industrial automation startup?  

 It’s really both. The company is building humanoid robots and robot dogs, but the commercial aim is industrial and labour automation rather than consumer gadgets. That puts it in the overlap between deeptech, robotics infrastructure, and physical AI.

Moment Energy repurposes retired EV battery packs into commercial storage systems. The company just raised a $40 million Series B to expand across North America. The problem is simple: power demand keeps climbing while grid infrastructure and fresh battery supply aren’t keeping pace. Co-founder and CEO Edward Chiang started the company in 2019 with Sumreen Rattan, Gabriel Soares, and Gurmesh Sidhu. The team believed old EV batteries could provide cheaper domestic energy storage instead of heading straight to recycling. That bet just got a lot bigger.

What does Moment Energy do with second-life EV batteries?

Moment Energy takes healthy retired EV battery packs — usually lithium-ion packs with roughly 70% or more state of health — and turns them into battery energy storage systems for commercial and industrial customers. The workflow is pretty direct: the company reviews pack data and handles transport. It runs intake testing at its UL 1974-certified facility, removes the automaker’s original battery management system, installs its own controls, and rebuilds those battery modules into stationary storage products. That last step matters. Moment’s whole pitch is that a reused EV pack shouldn’t pretend it’s still inside a car.

Its flagship product today is Luna BESS, a modular system aimed at businesses that need backup power, peak shaving, EV charging support, or better use of on-site solar and wind. Moment sells it in a 400 kWh “Half Luna” format and a 1 MWh “Full Luna” format, both with 2–4 hour discharge duration, 480 VAC output, plug-and-play deployment, and 24/7 remote monitoring. Luna is also scalable to 10 MWh for larger commercial and industrial sites.

The feature set is less flashy than a lot of startup decks. That’s probably the point. Moment emphasizes a custom battery management system and continuous monitoring. It also highlights electrical and thermal safeguards, plus fire-risk mitigation validated through UL 9540 and UL 9540A testing. In October 2025, Luna became the first repurposed battery system built from second-life lithium-ion batteries to clear UL 1973, UL 9540, and UL 9540A safety milestones together.

For customers, the before-and-after is pretty practical. Instead of just eating ugly peak-demand charges or overbuilding electrical service, a site can charge the system when power is cheaper and discharge during expensive spikes. The same platform can also support fleet charging and add backup power. It can also reduce downtime during grid disruptions. Its newer pack-swapping architecture is meant to make failed modules easier to replace and let customers benefit from newer battery chemistries over time instead of scrapping the whole unit.

How Moment Energy built its second-life EV battery business

Four engineers, one garage, and a battery problem

Moment Energy started in November 2019 in a home garage in Surrey, British Columbia. The four co-founders — Edward Chiang, Sumreen Rattan, Gabriel Soares, and Gurmesh Sidhu — were all engineers, and they’d already worked together building electric race vehicles before starting the company. That’s a relevant origin story for a startup whose core job is figuring out what’s left inside an aging battery pack and how to use it safely.

Chiang is the public face of the company, and his background lines up with the business. He studied mechatronics engineering at Simon Fraser University, later earned a Forbes 30 Under 30 nod, and was named Fasken’s Climate Tech Founder of the Year in 2025. Public details on the rest of the founding team are thinner, but their operating roles are clear: Rattan is COO, Soares is CTO, and Sidhu is CPO.

Traction got there before the hype did

This isn’t a pre-product battery startup. Moment has deployed 10 projects across EV charging, grid, and defence applications, and TechCrunch reported the company now has about 72 employees. It signed battery supply relationships with Nissan North America in 2020 and Mercedes-Benz Energy in 2022. That’s a big deal for any business that depends on predictable access to retired packs.

The early deployments are the kind that make sense for second-life systems. Quadra Island was an off-grid proof point. God’s Pocket Resort used Moment’s Flora BESS to cut diesel use, and a later profile said the site had logged 1,000 cycles with a 66% reduction in diesel usage. The company also deployed storage integrated with solar and wind for relocatable camps under the Canadian Department of National Defence.

The funding stack is getting serious

The new round is a $40 million Series B. Evok Innovations led it, with Liberty Mutual Investments, W23 Global Fund, and Acario joining existing backers including Amazon’s Climate Pledge Fund, Voyager Ventures, and In-Q-Tel. Moment says total capital raised is now above $100 million. Earlier rounds included a CAD 3.5 million seed led by Version One Ventures in 2021 and a $15 million Series A co-led by Amazon’s Climate Pledge Fund and Voyager Ventures.

There’s also government-backed capital in the mix. TechCrunch reported the company secured a $20 million loan from the U.S. Department of Energy, and Moment separately announced a $20.3 million DOE award tied to its first certified EV battery repurposing facility in the U.S. That helps explain why the company is building a gigawatt-scale factory in Austin, Texas rather than staying a niche project developer.

Where it sits against competitors

Moment isn’t the only company chasing second-life battery storage. B2U Storage Solutions has built large-scale storage around second-life EV packs using a model that keeps original packs intact with its EV Pack Storage approach. Smartville has pushed a multi-pack platform built around its own battery management hardware and qualification software. That’s real competition, but it also shows the category still hasn’t settled on one standard architecture.

Moment’s differentiation is more conservative than sexy. It strips out the automaker battery management system instead of leaving it in place. It writes its own software and has made certification the center of the pitch. Chiang told TechCrunch that some rivals test against UL standards without actually getting certified, while Moment’s view is simpler: “We got it.” Pair that with a modular design and domestic manufacturing story, and you can see why investors like the setup.

Why this second-life EV batteries round matters

This round matters because it moves Moment closer to being a manufacturer, not just a clever systems integrator. The capital will expand its North American manufacturing footprint, scale specialist teams, and increase production capacity across the U.S. and Canada for utilities, industrial users, and data centers. That’s a different phase of company-building. And it’s a much harder one.

It also gives Moment a shot at turning safety into a commercial advantage instead of just a compliance line item. Chiang’s argument is that certification affects insurance, permitting, and whether serious customers will even touch reused batteries. Liberty Mutual’s investment arm joining the round doesn’t prove the business model works at scale, but it reinforces the idea that Moment is trying to build something underwritable, not just interesting.

There’s another signal here. Chiang told TechCrunch that data center companies have been calling, but he also sounded wary of signing far-off deals just to raise the next round. That’s refreshing, frankly. Battery startups have a habit of selling the 2030 vision before they’ve nailed the 2026 factory. Moment still has a lot to prove.

Why second-life EV batteries are getting real demand

The macro setup is pretty obvious now. In the U.S., cumulative utility-scale battery storage capacity topped 26 GW in 2024 after 10.4 GW of new additions, and EIA said operators planned another 19.6 GW for 2025. That kind of buildout creates room for more than one battery supply model, especially if reuse can lower costs or shorten lead times for commercial projects.

Globally, battery storage is expanding even faster. The IEA said 108 GW of new battery storage capacity was deployed in 2025, up 40% from 2024, with installed capacity now 11 times higher than in 2021. Around 80% of those additions were utility-scale, and battery-based UPS systems used heavily in data centers grew 30% to 45 GW. So yes, Chiang’s “infinite” demand line is an exaggeration. But it’s not a random one.

There’s also a supply-chain angle investors clearly care about. TechCrunch cited BNEF data showing Chinese companies control about 72% of the global market, which turns battery storage into more than a cost issue. It becomes an industrial policy issue too. That’s a big reason second-life systems built from batteries already circulating in North America look less like an environmental side project and more like domestic infrastructure.

The real test for Moment Energy

Moment Energy has a credible story because it isn’t just saying reused batteries are cheaper or greener. It’s saying second-life EV batteries can be safe, insurable, modular, and manufacturable at scale. That’s a much tougher claim.

Now comes the part that matters most: getting the Austin factory built and shipping more certified systems. Then it has to prove this model works outside pilot-scale enthusiasm.

Read how CopilotKit raised $27M in Series A funding to build an agentic frontend stack that lets AI agents interact directly inside apps through actions, state changes, and generative interfaces, helping enterprises move beyond standalone chatbots toward deeply integrated AI workflows.

FAQ

– What did Moment Energy raise in 2026? 

 Moment Energy announced a $40 million Series B on May 5, 2026. Evok Innovations led the round, and the investor list included Liberty Mutual Investments, W23 Global Fund, and Acario alongside earlier backers such as Amazon’s Climate Pledge Fund, Voyager Ventures, and In-Q-Tel.

– How does Moment Energy turn EV batteries into energy storage systems? 

 It buys healthy retired EV packs, tests and grades them, removes the automaker battery controls, and rebuilds the modules into stationary battery energy storage systems for commercial and industrial sites. The resulting products can handle things like demand charge reduction and backup power. They can also support on-site renewable integration, with Luna BESS sold in 400 kWh and 1 MWh formats.

– Who founded Moment Energy? 

 Moment Energy was founded in 2019 by Edward Chiang, Sumreen Rattan, Gabriel Soares, and Gurmesh Sidhu. The team started the company in Surrey, British Columbia after previously working together on electric race vehicles, and Chiang came into the business with a mechatronics engineering background from Simon Fraser University.

– Is Moment Energy a battery recycling company or a battery storage company? 

 It’s closer to a battery storage company with a repurposing engine at the center. Instead of shredding EV batteries for raw materials right away, Moment tries to extend their useful life in stationary storage first, which puts it in the second-life battery segment rather than pure recycling or conventional new-battery BESS manufacturing.

CopilotKit builds software that lets AI agents live inside apps and respond with actions, state changes, and interactive interfaces instead of a plain chat box. In May 2026, the Seattle startup announced a $27 million Series A. The round matters because a lot of companies are done with bolt-on chatbots that feel detached from the product itself. Founded in 2023 by brothers Atai Barkai and Uli Barkai, CopilotKit is trying to turn AG-UI — its open protocol for agent-to-interface communication — into a standard layer for enterprise software teams building agentic apps.

What is CopilotKit and how do its AI agents work?

CopilotKit is an agentic frontend stack. A developer connects an agent backend or direct model to a React app and wires in CopilotKit’s client libraries. That gets them chat and generative UI. It also adds in-app actions and human-in-the-loop workflows that stay synced with the application itself. Under the hood, AG-UI acts as the event layer between frontend and backend, using web-native transport like HTTP and WebSockets to stream state, actions, and interface updates back and forth.

A lot of teams usually build that part by hand — badly. CopilotKit’s framework includes real-time context awareness and bidirectional app-agent state sync. It also offers headless UI options for custom design systems, plus observability hooks so teams can inspect what the agent did and why. There’s also an Inspector tool for debugging interactions in real time and historically. That matters once a demo becomes a production workflow.

The most interesting bit is the UI layer. Through A2UI and MCP-style generative UI patterns, an agent can render structured components instead of dumping text into a chat bubble. CopilotKit supports both a dynamic schema approach — where a secondary model generates the interface schema and data — and a fixed schema approach, where teams predefine the component tree and let the agent stream data into it. That’s how Barkai gets from a vague prompt like “show me revenue by category” to a chart, a card, or some other interactive element that fits the app.

On top of the open-source core, the company is now pushing CopilotKit Enterprise Intelligence. That product adds persistent threads and analytics. It also includes continuous learning from human feedback, self-hosted deployment, air-gapped support, SSO, and role-based access control. In plain English: it’s the stuff a big company needs after the prototype works and security, compliance, and continuity show up to ruin the party.

Who founded CopilotKit and why did they start it?

The founding story

CopilotKit was founded in 2023 and publicly launched in 2024. Atai Barkai is CEO, and Uli Barkai runs growth. Their starting point was simple: chat-only AI inside software feels clunky, especially when the real job isn’t conversation but interaction — approving something, editing something, taking an action, or moving through a workflow with the UI reacting in real time.

That view shows up all through the company’s pitch. Atai’s core argument is that agents shouldn’t answer only with text; they should use interfaces defined by the product team. He described that shift as “not just with blocks of text, but with interactive UIs,” which is basically the whole thesis in 10 words.

Why the founders fit this market

Atai Barkai looks like the technical half of the pair. Before CopilotKit, he worked at Meta on media SDK and infrastructure problems. He also spent time at Doximity as a senior software engineer, helped build FermiCloud, and later founded tawkitAI, which produced PodcastGPT. He studied physics at the University of Pennsylvania. That doesn’t make a founder special by itself, but he’s spent years close to developer tooling, infrastructure, and productized AI.

Uli Barkai’s background is more on the distribution side. Before co-founding CopilotKit, he led marketing at tawkit and worked on PodcastGPT’s go-to-market story. His academic background runs through financial economics at Columbia and philosophy at Tel Aviv University. It’s a slightly unusual mix, but useful if your job is translating technical ideas into a market people will actually buy from.

Traction and fundraising

The early signals are strong, even if startup traction claims deserve a raised eyebrow. CopilotKit’s libraries are now reaching more than 4 million weekly downloads. Its open-source projects have more than 40,000 GitHub stars and 150 contributors. Its tooling also powers millions of agent-user interactions in production. The company says customers include DocuSign, S&P Global, Cisco, and Deutsche Telekom, while its team sits at about 25 employees.

The new money is a $27 million Series A. Glilot Capital, NFX, and SignalFire led the round. CopilotKit says the capital will help expand the enterprise toolkit around AG-UI, support self-hosted deployments, and grow the team. Open standards don’t pay the bills by themselves.

How CopilotKit compares with Vercel AI SDK and OpenAI alternatives

CopilotKit isn’t alone. Vercel’s AI SDK helps developers build AI apps. Assistant-ui focuses on interface components for chat-style assistants. OpenAI’s Apps SDK gives developers richer UI options — but only inside ChatGPT. The older alternative is even less elegant: teams bolt a chatbot onto an app, then custom-build the plumbing to connect it to backend actions and frontend state.

Its pitch is that it’s more horizontal than those stacks. CopilotKit wants to sit across whatever model provider, cloud, or agent framework an enterprise already uses. It also offers self-hosting and stricter control over how much the agent can change the UI. Atai summed that up cleanly: “enterprises want optionality and they want self-hosting.”

Why does the CopilotKit funding round matter?

Because this isn’t just a bigger balance sheet. It’s the point where an open-source protocol bet turns into an enterprise software company.

AG-UI already has adoption across major vendors and frameworks, and CopilotKit now wants to capture the paid layer on top of that usage: support and hardened deployment. It also wants persistence, analytics, and governance. If that works, the company doesn’t need to beat every model provider. It just needs to become the default frontend and orchestration layer teams reach for when they want agents inside real products.

It also matters for customers. A lot of enterprise AI pilots stall when teams hit security reviews, audit requirements, or the messy reality of multi-session workflows. CopilotKit’s new enterprise push is aimed right at that gap. That’s why features like air-gapped deployment, persistent threads, and role-based access control matter more than another flashy demo.

How big is the enterprise AI agents market?

Pretty big already, and growing fast enough to explain why investors are leaning in. Grand View Research estimates the global enterprise AI agents market hit $3.67 billion in 2025 and could reach $83.4 billion by 2033, a projected 48.4% CAGR. North America was the largest revenue-generating region in 2025, which lines up with where most enterprise software budgets and AI infrastructure vendors are concentrated.

The timing also makes sense. Developer tooling for agents is fragmenting into protocols and interface layers — AG-UI for frontend-backend communication and A2UI for declarative UI rendering. There are also MCP-style app patterns for tool use, plus a growing list of supported frameworks from LangGraph to Google ADK and Microsoft Agent Framework. That sounds wonky. It is. But it’s also the kind of plumbing layer that becomes valuable once teams stop treating AI as a novelty and start embedding it into core software flows.

Final take on CopilotKit funding

The easiest way to read CopilotKit funding is as a bet on the layer between the model and the screen. Not the model. Not the chatbot shell. The connective tissue that lets an agent understand what a user is doing, take action inside the app, and present something usable back. The next thing to watch is whether CopilotKit can turn protocol adoption into durable enterprise standardization before larger platforms absorb the same territory.

Read how a16z crypto raised a $2.2B crypto-only fund to back blockchain and web3 startups, doubling down on long-term crypto infrastructure and digital asset investing even as AI pulls venture capital away from the sector.

FAQ

– What is the CopilotKit funding round? CopilotKit raised $27 million in a Series A announced on May 5, 2026. Glilot Capital, NFX, and SignalFire led the round, and the company says the money will go toward expanding its enterprise product and hiring beyond its current team of about 25 people.

– How does CopilotKit work inside an app? CopilotKit gives developers a frontend stack that connects an AI agent to the app’s own state, actions, and UI components. Instead of returning only text, the agent can trigger workflows and stream updates. It can also render structured elements like cards or charts through AG-UI and A2UI-style generative UI.

– Who founded CopilotKit? CopilotKit was founded in 2023 by brothers Atai Barkai and Uli Barkai. Atai came in with engineering experience from Meta, Doximity, FermiCloud, and tawkitAI, while Uli had already worked on marketing and growth at tawkit before helping build CopilotKit’s commercial side.

– Is CopilotKit part of the enterprise AI agents market? Yes — it sits in the enterprise AI agent infrastructure layer, specifically around agentic frontends, generative UI, and production deployment tooling. That’s a category Grand View Research sized at $3.67 billion in 2025, which helps explain why investors are backing platforms that make agents usable inside business software instead of just chat windows.

a16z crypto is Andreessen Horowitz’s venture arm for crypto and web3 startups. The new a16z crypto fund is a $2.2 billion vehicle announced on May 5, 2026, and it lands at an awkward moment: crypto trading is soft, venture funding has cooled, and AI is pulling a lot of investors off-course. Chris Dixon has led Andreessen Horowitz’s crypto push since its early bets in 2013, and the firm launched its first dedicated crypto fund in 2018. That history matters. This latest raise says a16z still wants to be the biggest specialist checkwriter in crypto even when the cycle feels quiet.

The headline number is huge. But the bigger message is sharper: a16z says this fund won’t chase the hotter AI trade and will stay “dedicated 100% to crypto entrepreneurs.” The firm is also promoting CTO Eddy Lazzarin to general partner. That expands the core GP investing roster to 4 people alongside Dixon, Ali Yahya, and Guy Wuollet.

What does the a16z crypto fund actually do?

The simplest answer is this: a16z crypto writes venture checks into crypto and web3 startups from early stage through growth, then surrounds those founders with a much bigger operating stack than most specialist funds can offer. Its crypto team backs companies across infrastructure and applications. It also backs network-based products. Founders get access to research, engineering, security, legal and regulatory support, recruiting, go-to-market help, governance work, media, and its CSX accelerator. That’s not just branding. It’s the product.

For a founder, the before-and-after is pretty different. Before taking money from a platform-heavy firm, a startup might have to patch together outside auditors, token design advice, policy counsel, recruiting intros, and distribution help on its own. Afterward, a16z wants that work to feel more centralized and faster, especially for projects that need help with protocol design, token mechanics, or regulatory questions that a normal SaaS fund can’t answer.

That technical angle is why Lazzarin’s promotion matters. He joined a16z crypto in 2019 as a data scientist and investing partner, became CTO in 2023, and has been leading the engineering, data science, research, and security teams. Dixon said Lazzarin also helped drive Jolt, an open-source zkVM project, and has become one of the firm’s clearest thinkers on token design and classification. In plain English, a16z is elevating someone who can talk about code, incentives, and market structure in the same meeting.

The Fund 5 thesis makes that even clearer. The team is pitching crypto as infrastructure for instant global money movement, stablecoin savings, tokenized assets, prediction markets, onchain lending, and even AI-agent commerce. It also frames crypto as a counterweight to software systems that are becoming more centralized and, in the firm’s words, more opaque, especially in AI.

How did a16z crypto become so big?

Chris Dixon’s long bet on crypto

Andreessen Horowitz had already been investing in crypto before the dedicated brand existed, and Dixon has been the public face of that push for years. Before joining a16z in 2012, he cofounded SiteAdvisor and Hunch, which gave him founder credibility before he became one of venture’s loudest web3 advocates. The first dedicated a16z crypto fund launched in 2018 at $300 million. Fund 5 is now the firm’s fifth crypto vehicle, bringing total capital raised for the strategy to $9.8 billion.

The bench behind the brand

This isn’t a one-person franchise anymore. Ali Yahya came out of Google X and Google Brain before becoming a general partner focused on crypto. Lazzarin’s background runs through Facebook Messenger analytics and Netflix data systems, which helps explain why a16z values him as both an investor and a technical operator. Wuollet was already on the GP team. Now the firm has 4 general partners steering crypto investing.

Track record, traction, and the new fund

a16z crypto isn’t selling a blank slate. Its portfolio includes Coinbase, Kalshi, and Solana Foundation. Those names signal very different parts of the market, from exchange infrastructure to prediction markets to core blockchain ecosystems. That spread matters because Fund 5 is being pitched around turning crypto infrastructure into products people actually use every day, not just placing another broad macro bet on token prices.

Competition and market positioning

The direct competition is obvious. Paradigm is raising as much as $1.5 billion for a new fund that would stretch beyond crypto into AI and robotics. Katie Haun’s firm announced $1 billion in new funds on May 4, 2026, and while it remains crypto-focused, it’s also talking openly about AI agents where they intersect with crypto, blockchain, and fintech. Even Y Combinator’s current Requests for Startups page leans toward AI-native workflows and other categories without an explicit crypto callout.

a16z wants to look different here. Legacy alternatives for founders still include generalist venture firms and hedge funds that show up in bull markets. There are also direct token investors that don’t offer much operating help. a16z is betting that specialist depth still wins if the market gets more technical, more regulated, and more selective. Unlike rivals flirting with adjacent themes, this fund is being sold as crypto-only.

Why does this $2.2 billion a16z crypto fund matter now?

Because the timing is almost contrarian.

On the same day a16z unveiled Fund 5, Coinbase said it would cut about 700 employees, or 14% of staff, as part of a restructuring. That’s not the backdrop you pick if you’re trying to surf euphoria. It’s the backdrop you pick if you believe the best deals come when everyone else is distracted, tired, or chasing the next thing.

The promotion of Lazzarin sharpens that signal. a16z could’ve treated this as a pure fundraising announcement. Instead it paired the raise with a bet on a more technical investor profile, someone tied to token design, zero-knowledge tooling, and deeper support for portfolio companies. That suggests Fund 5 isn’t just bigger capital. It’s supposed to be more hands-on capital.

For founders, clarity matters. If other major crypto funds are spending time on robotics, AI agents, or broader fintech, a dedicated pool of $2.2 billion aimed only at crypto startups becomes a real recruiting tool. Not a guarantee. But definitely a recruiting tool.

What is the crypto venture market in 2026?

Cold by recent standards. Not dead.

CoinGecko said spot trading volume on the top 10 centralized exchanges fell 39.1% in Q1 2026 to $2.7 trillion, and March alone dropped to $0.8 trillion—the weakest month since November 2023. That helps explain why a mega-fund announcement felt a little weird this week. The market isn’t in panic mode, but it also isn’t anywhere near the old frenzy.

Venture activity tells the same story. DL News, citing DefiLlama, said crypto startups raised nearly $5 billion in Q1 2026, down from closer to $6 billion in the year-ago quarter. That’s still a lot of money. It’s just less forgiving money, and founders are feeling the difference as AI valuations reset investor expectations.

Long term, the prize is still enormous. Grand View Research pegs the global blockchain technology market at $31.28 billion in 2024 and projects it could reach $1.43 trillion by 2030. Its U.S. outlook alone points to roughly $402 billion in revenue by 2030. The short-term mood is sluggish. The long-term market story is still big enough to justify specialist funds with real patience.

What to watch after the a16z crypto fund raise

This raise doesn’t prove crypto venture is back. It proves the largest specialist firms still think the reset is survivable.

What matters next is deployment. Watch whether Fund 5 pours into stablecoin infrastructure, onchain financial apps, tokenized assets, and crypto tools for AI agents, the exact areas a16z flagged in its thesis. If those bets turn into breakout companies while rivals keep drifting toward broader AI themes, the a16z crypto fund could end up looking less like a late-cycle flex and more like one of the few clean conviction calls of 2026.

Read how Altara AI raised $7M in seed funding led by Greylock to build an AI-powered intelligence layer for industrial and scientific teams, helping companies across batteries, semiconductors, and medical devices unify fragmented engineering data and accelerate failure analysis without replacing existing systems.

FAQ

– What is the new a16z crypto fund? 

 It’s a $2.2 billion fifth crypto fund from Andreessen Horowitz’s a16z crypto unit, announced on May 5, 2026. The raise brings the strategy’s total capital to $9.8 billion and is aimed at backing crypto startups building products on top of blockchain infrastructure, not just speculative token plays.

– How does a16z crypto actually help founders? 

 It does more than invest. a16z crypto gives portfolio companies access to research, engineering, security, legal and policy help, recruiting, go-to-market support, governance work, media, and the CSX accelerator. That makes it closer to a full operating platform than a simple checkbook.

– Who runs a16z crypto now? 

 Chris Dixon still leads the franchise, and the GP investing team now has 4 people: Dixon, Ali Yahya, Guy Wuollet, and Eddy Lazzarin. Lazzarin was promoted from CTO to general partner in the same announcement as Fund 5. That says a lot about how central technical depth has become to the firm’s investing style.

– Is crypto venture capital still a big market in 2026? 

 Yes, but it’s slower and more selective than a year ago. Crypto startups still pulled in nearly $5 billion in Q1 2026, even as trading volumes weakened and many investors chased AI, while the broader blockchain technology market is still forecast to expand dramatically through 2030.

Altara AI builds software that turns scattered engineering and scientific data into one usable system for companies working on batteries, semiconductors, medical devices, and other physical-world products. The San Francisco startup has raised a $7 million seed round led by Greylock, a bet that the mess of spreadsheets, legacy tools, and disconnected lab systems has become painful enough to support a real software business. Altara was founded in 2025 by Eva Tuecke and Catherine Yeo, who met while studying computer science at Harvard and want to shrink a failure-analysis process that can take weeks into something much closer to minutes.

What is Altara AI and how does it work?

Altara AI is basically an intelligence layer for industrial and scientific teams. Instead of asking a battery maker or chip company to rip out old systems, Altara connects to the data where it already lives. It ingests that data and builds a shared context that its agents can reason over. The pitch is simple: no giant migration project first, no forcing a new system of record, just software that sits on top of the mess and makes it useful.

The product is designed for ugly data. Altara can work across semiconductor wafer maps, SEM inspection images, instrument time-series data, spreadsheets, technical reports, PowerPoints, and legacy domain software. That matters because failure analysis in physical sciences usually isn’t one neat dashboard problem. It’s a hunt across text, numbers, images, and historical records that were never meant to talk to each other.

Once that context is in place, customers can use prebuilt workflows or create their own agents without needing deep AI expertise. Altara lists experimental design and yield analysis among the workflows it supports. Information synthesis, anomaly detection, and failure analysis are in the mix too. It’s also trying hard to look enterprise-ready from day 1, with SOC 2 Type II compliance, SSO support, audit logs, role-based permissions, and deployment options that include self-hosted VPC setups or single-tenant cloud infrastructure.

That combination shows what Altara wants to be. Not a chatbot for scientists. More like observability software for hardware R&D and manufacturing — software that traces what happened, why it happened, and what to do next. Greylock’s Corinne Riley made that comparison directly when she likened Altara’s role in physical science to an SRE stack in software.

Who founded Altara AI and why now?

The founding story

Altara was started in 2025 by Eva Tuecke and Catherine Yeo after they saw how much time technical teams still burn on manual data triage. Yeo described the current workflow as a “scavenger hunt,” which feels right if you’ve ever watched an engineering team chase a failure across sensor logs, moisture readings, temperature data, and old reports. Their timing isn’t random either. Both founders are coming at this just as agentic AI is getting good enough to reason across messy, multimodal inputs instead of only tidy business documents.

Why the founders fit this job

Tuecke’s background is unusually on-point for a company selling into frontier hardware. She did computational particle physics research at Fermilab, worked on Starlink software at SpaceX, and spent time in machine learning research at the MIT-IBM Watson AI Lab. Greylock sees her as someone who has already worked at the edge of hard science and complex engineering systems.

Yeo brings the AI product side. Before Altara, she built AI and collaboration features at Warp for a large developer user base. Earlier, she did AI research at Disney, IBM, MIT, and Harvard. Greylock also points to a less obvious edge: she grew up in a family of 5 electrical engineers working in semiconductors, which helps explain why Altara sounds less like a generic AI wrapper and more like software built by people who understand technical workflows.

There isn’t a big prior startup exit here. But there is a pattern. Tuecke had already built smaller projects before Altara, and Yeo had a mix of product shipping, research, and venture experience, including time as a Contrary venture partner and prior work across Apple and IBM. For a seed-stage company selling to exacting industrial buyers, that mix of technical credibility and product sense probably matters more than résumé theater.

Early signals, the seed round, and where Altara sits

Altara is just coming out of stealth, with demo requests open and an in-person San Francisco team that already includes alumni from Applied Intuition, SpaceX, Warp, Jane Street, and Microsoft. Greylock says customers are already trusting the company with critical workflows, and the careers page shows Altara hiring across software engineering, research engineering, design engineering, and chief of staff roles. It’s not proof of huge traction yet. But it does suggest the company is past the napkin stage.

The seed round totals $7 million. Greylock led it, with Neo, BoxGroup, Liquid 2 Ventures, and Jeff Dean participating; Altara also says leaders from OpenAI and AMD joined as angels. That investor mix is a loud signal that the company is being framed as deep technical infrastructure, not just another AI app with a science theme.

Competition and market positioning

Altara’s most useful comparison isn’t another lab-automation startup. It’s Resolve, the Greylock-backed company that uses AI to diagnose software failures and was valued at $1.5 billion in TechCrunch’s comparison. Altara wants to do the hardware version of that job. It aims to figure out why a battery cell, semiconductor process, or device test failed, using fragmented technical evidence instead of logs from cloud infrastructure.

That puts it in a different lane from startups like Periodic Labs and Radical AI. Periodic is chasing AI scientists and autonomous labs, and it came out of stealth with a massive $300 million seed round. Radical AI raised $55 million to build an integrated materials-discovery system that spans design, lab work, and manufacturing. Altara looks much more pragmatic: plug into existing systems, fix the information problem first, and sell software before trying to reinvent the scientific method. That’s a lot less glamorous. It may also be a lot easier to commercialize.

Why are investors backing Altara AI now?

Because the company is attacking a painful problem with a product that looks deployable right now. Altara isn’t promising fully autonomous discovery from day 1. It’s promising faster failure analysis and better experiment design. Less manual triage too. That kind of wedge tends to sell better than moonshots do.

The investor case also rests on trust. These are high-stakes environments where a bad answer can waste months of R&D or send a manufacturing line in the wrong direction. Altara’s emphasis on auditability, enterprise controls, and data privacy — including the promise that customer data isn’t used for model training — suggests the company understands that buyers in semiconductors and medical devices won’t tolerate black-box software theater.

There’s a category argument here too. Riley called AI for physical science the “next big frontier,” and Greylock’s own memo makes clear it thinks scientific and industrial data is one of the most valuable underused assets in tech. Altara now has to prove it can turn that thesis into repeatable workflows inside a few verticals, not just broad interest from investors.

How big is the market for AI in physical sciences?

It’s not a tiny niche. Grand View Research estimates the global AI in manufacturing market was worth $5.32 billion in 2024 and projects it will hit $47.88 billion by 2030, a 46.5% compound annual growth rate. That’s broad manufacturing, not just Altara’s slice of it, but it shows why investors are starting to care about software built for industrial and scientific workflows instead of only office productivity.

The demand signals are already there. McKinsey’s 2025 COO100 survey found that 46% of manufacturing executives reported limitations in their data or IT/OT systems, with outdated infrastructure and poor data quality among the biggest blockers to AI adoption. That sounds a lot like the exact hole Altara is trying to fill.

At the same time, adoption is still messy. Deloitte found that 87% of manufacturers had started a GenAI pilot, but only 24% had deployed a use case in at least 1 facility and just 10% had implemented GenAI across broader networks. The gap is wide. That creates room for companies that can make AI trustworthy inside operational environments, not just impressive in a demo.

What happens next for Altara AI?

Altara AI isn’t trying to build the lab of the future from scratch. It’s trying to make today’s lab and manufacturing stack less chaotic.

That may be the smarter opening move. The company can win a lot of value if it becomes the system engineers trust when a wafer run goes sideways or a battery test fails and nobody knows why. The next thing to watch is whether Altara can turn this seed round into named customer wins and deeper adoption in the industries it already targets — especially semiconductors, batteries, and medical devices.

Read how Jurisphere.ai raised $2.2M from Info Edge Ventures, Flourish Ventures, and others to expand its AI-powered legal workspace, helping lawyers automate document review, legal research, drafting, and matter management through a unified platform built for high-volume legal workflows.

FAQ

– What funding did Altara raise? 

 Altara raised a $7 million seed round announced on May 5, 2026. Greylock led the financing, with Neo, BoxGroup, Liquid 2 Ventures, and Jeff Dean participating, and Altara also says leaders from OpenAI and AMD joined as angels.

– How does Altara AI work for hardware and manufacturing teams? 

 Altara AI connects to existing scientific and engineering data sources and unifies that information in context. It then runs agents on top of it for workflows like failure analysis and anomaly detection. Yield analysis and experimental design are part of the pitch too. The key point is that it works across mixed data types — from spreadsheets and reports to wafer maps, images, and instrument logs — without forcing a full system migration first.

– Who founded Altara AI? 

 Altara was founded in 2025 by Eva Tuecke and Catherine Yeo after the two met while studying computer science at Harvard. Tuecke previously worked at Fermilab and SpaceX, while Yeo worked as an AI engineer at Warp and did research at Disney, IBM, MIT, and Harvard.

– Is Altara AI a materials science startup or an enterprise software company? 

 It looks more like enterprise software for physical-science organizations than a pure materials-discovery startup. Unlike companies such as Periodic Labs and Radical AI that are building autonomous labs or end-to-end scientific discovery systems, Altara sells an intelligence layer that plugs into existing R&D and manufacturing environments.