Vishal Virani: The Surat Founder Who Rebuilt DhiWise as Rocket.new

Three months after its beta launch, Rocket.new told TechCrunch it had passed 400,000 users and $4.5 million in annual recurring revenue. Those are company claims, and they came from a team in Surat, not Bengaluru. They also came from Vishal Virani, a founder who had spent the previous four years building a different product, DhiWise, and then chose to pivot away from it.

Reading time: about 10 minutes

A Surat engineer, a consultancy and a problem that kept repeating

Accel Atoms, the investor’s early-stage programme, described Virani in December 2022 as a first-generation engineer. By that account, he started an IT consulting firm with his college friend Rahul Shingala while still in his eighth semester.

The firm grew to 150 developers, built software for clients across twelve verticals and generated more than $1 million in revenue, Accel reported. The figures come from that write-up and the founders’ own account. StartupIndiaX has not independently verified them.

The consultancy also showed the pair something. Every new interface meant developers rebuilding similar front-end work from scratch. That was a lot of talent spent on repetition, and the founders decided to automate it.

The result was DhiWise, founded in April 2021.

DhiWise: six months of research before the first product line

Before writing product code, the founders spent six months on research, Accel says. They studied Y Combinator and Accel’s Seed to Scale content to shape a product mindset, and they looked at more than 300 low-code and no-code platforms. Their conclusion: existing tools removed only a small share of a developer’s grunt work.

Two obstacles stand out in Accel’s account. One was an investor mindset that doubted engineers from a city like Surat, without an IIT or Google background, could build a viable product for developers. The other was the switch from running a service company to running a product company.

The service business did help in one practical way. About 60 developers from the old firm, who had seen the problem first hand, helped build DhiWise.

The early signals were encouraging. The founders joined Accel’s GrowthX programme in December 2021. DhiWise then won Product Hunt’s Golden Kitty for Developer Tool of the Year in January 2022 and had 32,000 beta users within a few months, per Accel. Money followed: a $2.5 million seed round in May 2022 and a $7 million Series A in August 2022, with Accel, Together Fund and India Quotient among the backers. Inc42’s data puts DhiWise’s total funding at $9.5 million.

The pivot: why Rocket.new is not DhiWise with a new name

By mid-2025 the company had moved on. TechCrunch described Rocket.new, which launched in beta in June 2025, as a pivot from DhiWise’s developer-workflow business. The old product left something behind: Rocket combines large language models from Anthropic, OpenAI and Google with its own deep learning systems, which TechCrunch says are trained on proprietary datasets from DhiWise.

Virani’s pitch is that many vibe-coding tools, which turn plain-language prompts into apps, win the first demo and struggle afterwards. He told TechCrunch his team is focused on the first-day-versus-second-day gap.

“…the problem of day two”

• Vishal Virani to TechCrunch, September 2025

The company calls its approach “vibe solutioning.” In a blog post covered by Head&Tale, Virani wrote that Rocket remembers decisions about a user’s brand, patterns and customers, so each iteration gets steadier.

“Less theater. More outcomes.”

• Vishal Virani, quoted by Head&Tale, 23 September 2025

The trade-off is speed. TechCrunch reported that generating a first application takes about 25 minutes, against under three for most rivals. In its early testing, the outlet found the result more complete, with all essential modules included.

Read More: Anupam Mittal Shaadi.com Story: From Sagaai.com to a Possible IPO

Sixteen weeks in the open: what Rocket says it has built

Everything in this section is a company claim from the September 2025 funding announcement. Virani said Rocket had more than 400,000 users in 180 countries, over 10,000 of them paying subscribers, and $4.5 million in annual recurring revenue.

He said about 80% of users were building applications he did not class as simple landing pages or branding sites. Around 45% were building mobile apps and 55% websites. E-commerce accounted for roughly 12% of builds and fintech for 10%. TechCrunch also reported users from Meta, PayPal, KPMG, PwC and Times Internet using it for personal projects.

The team stood at 58 people, mostly in Surat, with plans to double engineering and product hiring in India over 12 months.

Virani set targets of $20 million to $25 million in ARR by the end of 2025 and $60 million to $70 million by June 2026. StartupIndiaX found no published figure showing whether either was met, and Rocket did not respond to Inc42’s queries for its June 2026 report. Treat the September 2025 numbers as a snapshot, not a current reading.

From DhiWise to Rocket, year by year

Four years separate DhiWise’s founding from Rocket’s beta. The next step came fast: a $15 million round about three months after launch.

How Rocket makes money

Rocket sells subscriptions priced in AI tokens. At the September 2025 funding announcement, TechCrunch reported a free trial capped at one million tokens, then $25 a month for five million tokens. Inc42’s June 2026 report lists three paid tiers at $25, $50 and $100 a month, with higher tiers adding larger token limits, private projects, code downloads and an option to opt out of data training.

TechCrunch reported that the pricing discourages hobbyists, and that gross margin was 50% to 55%, a figure the company aims to raise to 60% to 70% within months. Both margin figures are company claims as reported.

By the company’s account, the US was the biggest market at 26% of revenue, followed by Europe at 15% to 20% and India at 10%. In April 2026, Inc42 reports, Rocket entered enterprise solutions with Rocket 1.0, which helps businesses build apps and analyse competitors after launch.

Earlier financials belong to the DhiWise era. Entrackr reported FY24 revenue of ₹3.5 crore (about $0.4 million at ₹87.6 per dollar, the rate implied by Entrackr’s own conversion of its round figures) and a loss of ₹11 crore (about $1.3 million), citing standalone filings. Those numbers predate Rocket’s growth and say little about today.

Read More: Zetwerk Founders Story: How Amrit Acharya and Srinath Ramakkrushnan Built Zetwerk

Who backed the bet, and what the filings show

Rocket’s September 2025 round was $15 million. TechCrunch and Head&Tale call it a seed round, led or co-led by Salesforce Ventures with Accel and Together Fund. Entrackr, citing a Registrar of Companies filing, describes an extended Series A: 4,81,718 Series A1 shares at ₹2,728 each, raising ₹131.4 crore (about $15 million, as Entrackr converted it). StartupIndiaX uses “a $15 million round” because outlets disagree on the label.

Per Entrackr, Salesforce Ventures put in ₹74.46 crore (about $8.46 million), Accel ₹49.26 crore (about $5.6 million) and Together Fund ₹7.67 crore. After the allotment, Accel held 18.87%, Salesforce Ventures 13.68% and Together Fund 5.62%, while Virani and Shingala together held 36%.

Entrackr also estimated the valuation at ₹545 crore ($62 million), up from ₹256 crore (about $29 million at the same implied rate) three years earlier. These are estimates, not company disclosures, and StartupIndiaX has not verified them. Founders who want to see how a round like this changes ownership can test their own numbers in the Funding Round Dilution Calculator.

The June 2026 funding talks: what is reported and what is not

This is a developing story. Inc42 reported on 27 June 2026 that Rocket was in talks with Susquehanna International Group (SIG) for a Series A. One source told the outlet the talks were progressing well, and Salesforce Ventures and Accel were expected to take part. Inc42 could not determine the round’s size, and Rocket did not respond to its queries.

A separate report, attributed to the Economic Times by secondary outlets StartupIndiaX reviewed, described talks to raise $40 million to $50 million at about a $500 million valuation, with 360 ONE Asset leading. StartupIndiaX has not seen the original. If that valuation were confirmed, it would be roughly eight times Entrackr’s 2025 estimate (editorial arithmetic).

The two reports name different lead investors. They may describe one round, two, or talks that never concluded. StartupIndiaX found no confirmation of either deal closing as of 27 October 2026.

Rocket against the field

TechCrunch framed Rocket as a challenger to Lovable, Cursor and Bolt. Virani told it that many users build a website on Lovable or Replit, then bring the project to Rocket to generate a native mobile app on top of an existing Supabase backend. That is a company claim about user behaviour.

Another Indian entrant, Emergent, founded by former Dunzo co-founder Mukund Jha, raised $70 million earlier in 2026, per Inc42. Inc42 later published a headline saying Emergent raised $130 million at a $1.5 billion valuation. StartupIndiaX reviewed the headline, not the full article.

On disclosed funding, Rocket is the smaller player. Its bet, in our reading, is depth over speed: a slower first build that holds up through later changes. Whether that is a durable edge or a niche is not yet clear from public data.

Why there is no verified Vishal Virani net worth

No credible source publishes a net worth figure for Vishal Virani. Pages that claim one tend to be scraped profiles, and the only specific listing StartupIndiaX found offered a range too wide to be useful.

What is known is limited. Entrackr reported that he and Shingala together held 36% of the company after the September 2025 allotment, and estimated the company’s valuation at ₹545 crore ($62 million). A stake in a private company is not personal net worth. It cannot be sold freely, and ownership changes with every round. Treat any single net worth number with suspicion.

Five lessons from the DhiWise to Rocket switch

Use the service business as a lab. The consultancy showed Virani the same front-end work being rebuilt again and again. The product idea came from watching his own developers.

Research before building. Six months of study and a review of 300-plus platforms preceded DhiWise’s first product line, per Accel. Founders moving from services to product can copy the habit.

Be willing to set your own product aside. Virani did not abandon DhiWise’s assets. The data stayed and fed Rocket’s models, per TechCrunch.

Pick a different problem from the category leader. While rivals competed on the fastest first demo, Rocket chose the work that comes after it. That choice explains why Rocket accepts a slower first build.

Design around geography, not against it. Accel lists investor doubt about non-IIT engineers from Surat as an early obstacle.

“Zip code does not matter if you build the right product.”

• Vishal Virani, as quoted in a LinkedIn post by Amit Raja Naik, September 2025

What comes next

Rocket’s next chapter depends on two things: whether the June 2026 talks become an announced round, and whether its 2025 growth targets can be shown with real numbers. Until then, the story is a credible pivot with claims still waiting for proof.

If you have built something from a Tier-2 city, or you use Rocket, tell us in the comments. More founder stories are in the StartupIndiaX Founder Stories section.

FAQs