That’s the conundrum Zafer Ahmed and Nan Zhong faced in 2012, when they realized their popular walkie-talkie app, TiKL, had maxed out. It was a hit in the U.S. and in South Korea, with some 27 million downloads, but it had little adoption anywhere else in the world. The app thrived because the mobile carriers like Sprint and Nextel stopped providing built-in walkie-talkie apps on their phones. Most people didn’t need walkie-talkie in their daily lives, though, and the app’s adoption slowed to a crawl. “We took all [the carrier’s] users and realized why they shut it off: The user base is 25 million,” Ahmed says. So he and Zhong set on building something else. (TiKL is still available for download, but it is infrequently updated.)

The company’s next project could have been a flop. It’s difficult to build one app with millions of users. Could they repeat their success?

Ahmed and Zhong, who hail from India and China, respectively, noticed that each country’s wireless infrastructure was struggling to keep up with demand. Most of India’s networks are still using the 2.5G wireless standard, for example; developed countries have 4G or better, which allows for faster and richer data exchange. For India, that means many messaging and phone call apps available in the country, including Skype, don’t work very well. The TiKL team saw an opportunity to take market share.

They launched Talkray, an app for free phone calls, group calls, text messages, photos, and videos. The app, they say, has been engineered to reduce the bandwidth, packet rate, and latency required to place voice calls in an effort to ensure that it would function with a slow connection. Talkray has been downloaded by 12 million users and millions of them are active on the app each day, Zhong says. The app has spread largely by word of mouth: TiKL has spent no money on marketing the app and has received little press coverage about it.

Talkray is especially popular in India, Southeast Asia, Tanzania, and in island countries such as the Maldives and Jamaica. Growth is accelerating: It took Talkray over a year to get to five million users, but the next five million joined in just four months. Last month, Talkray added two million users. The number of users who open the app every day is growing by 20% each month, the company says.

Today the company, still doing business as TiKL, announces it has raised $5.3 million in a Series B round of funding from investors including Lerer-Hippeau, SV-Angel, Zhen Fund, and General Catalyst. The additional capital will go toward expanding its team, which has until now operated with just five employees.

The company, which graduated from the Y Combinator startup accelerator in 2012, previously raised $2.1 million from General Catalyst, Andreessen Horowitz, Lerer Ventures, and SV Angel.

Talkray hasn’t made any attempts to monetize thus far. Ahmed says it could become profitable tomorrow using popular models like selling stickers, money transfers, or minutes. For now, the company plans to stay lean and, as is custom at Silicon Valley startups, wait until it has hundreds of millions of users before monetizing them. Ahmed believes that, like WhatsApp (which had just 55 employees but sold to Facebook in February for $19 billion), Talkray could scale to more than 100 million users with a “low double-digit” number of employees. The WhatsApp deal put a spotlight on messaging apps, including Talkray. TiKL has held acquisition discussions with a few large technology companies, but decided to stay independent.

There’s one catch. Talkray’s competitive advantage, reliability on a slow connection, won’t be much of an advantage once the mobile infrastructure improves in developing countries. Ahmed says that there will always be a demand for a messaging app that works, even as bandwidth increases in places like India. “No matter how big you make the bandwidth, you always need more,” he says. “It’s never enough.” TiKL plans to add more features to the app, including video messaging. But its core selling point remains the ability to reduce the amount of data used, Ahmed says. “Our secret sauce will always be that.”