Twenty years ago, web-savvy folks were focused on solving the Internet’s “last-mile” problem. Today, by contrast, one of the biggest bottlenecks to expanding Internet access is rather around a “middle-mile” problem—crossing cities and tough terrain, not just driveways and country roads. The Google X spinoff Taara is promoting a simple alternative to fiber-optic cables: Free-space optical lasers. Using over-the-air infrared C-band lasers, Taara is rolling out tech that the company says reliably delivers 20 gigabit-per-second bandwidth across distances up to 20 kilometers.However, what happens to open-air laser signals on a rainy or foggy day? What about a flock of birds or stray tree branch blocking a tower’s signal? Plus, C-band communications tech is decades old. So why haven’t other innovators tried Taara’s approach before?IEEE Spectrum spoke with Taara’s CEO Mahesh Krishnaswamy about the company’s Google X pedigree (and its Google Fiber and Google Project Loon alumni) as well as upcoming new technologies, set to roll out in 2026, that’ll expand Taara towers’ bandwidth and range. Plus, the fledgling company’s wagering, their industry footprint might get a tiny boost too. What does Taara do, and what problem or problems is the company working to solve? Mahesh Krishnaswamy, CEO of Taara, says the Internet’s “middle-mile” problem presents an outsized opportunity. TaaraMahesh Krishnaswamy: Taara is a project that incubated over the last 7 years at [Google/Alphabet] X Development, and we recently graduated. We’re now an independent company. It is a technology that uses eye-safe lasers to connect between two line-of-sight points, using beams of light, without having to dig trench fiber. The problem we are really solving is that of global connectivity. Today, as we speak, close to 3 billion people are still not on the Internet. And even the five billion that are connected are running into challenges associated with speed, affordability, or reliability. It’s really a global problem that affects not just millions but billions of people.So Taara is addressing the digital divide problem? Krishnaswamy: Some of the ways our customers and partners have deployed [Taara's tech] is they use it for redundancy or to cross difficult terrain. A river, a railroad crossing, a mountain, anywhere the land is difficult to dig and traverse through, we are able to reach. One example is the Congo River, which is the world’s deepest river and one of the fastest flowing rivers. It separates Brazzaville [in the Republic of the Congo] and Kinshasa [in the Democratic Republic of the Congo]. Two separate countries on either side. But they’ve not been able to run fiber optic cables underneath the river. Because the Congo River is very fast-flowing. And so the only alternative is to go about 400 km, to where they’re able to safely navigate it. But we were able to connect these two countries very easily, and as a result bring bandwidth parity. One side had five times higher bandwidth cost than the other side. The Road to New Free Space Optical Internet TechWhat is Taara doing today that couldn’t have been done five or 10 years ago? Krishnaswamy: We’ve been slowly but steadily building up the improvements to this technology. This started with improvements in the optics, electronics, software algorithms, as well as pointing and tracking. We have enough margin to tackle most of the challenges that typically were limiting this technology up until recently, and we are one of the world’s largest manufacturers of terrestrial, free-space optics. We are live right now in more than 12 countries around the world—and growing every day.What is your company’s main technological product? Krishnaswamy: Today the technology that we have is called Taara Lightbridge. This is our first-generation product, which is capable of doing 20 Gbps, bi-directionally, at up to 20 km distance. It’s roughly the size of a traffic light and weighs about 13 kilograms. Taara’s traffic light–sized Lightbridge terminal serves as the hub for the company’s free space Internet tech—with thumbnail-sized components being promised for 2026. TaaraBut we are now about to embark on a significant sea change in our technology. We are going to take some of the core photonics and electronics components and shrink it down to the size of my fingernail. And it will be able to point, track, send, and receive light at tens of gigabits per second. We have this Taara chip in a prototype form, which is already communicating indoors at 60 meters as well as outdoors at 1 km. That is a big reveal, and this is going to be the platform by which we’re going to be building future generations of products.When will you be launching that?Krishnaswamy: It’ll be the end of 2026. The Internet’s Middle-Mile and Last-Mile Problems How does all of this relate to the tech being “middle mile” rather than what used to be called “last mile”? How much distinction is there between the two? Krishnaswamy: If you were to follow the path of data all the way from a subsea fiber, where you have Internet landing points, there’s this very vast capacity fiber that’s bringing it all the way from the edge of the coast into some main city. That’s a longhaul fiber. These are the national backbones, usually laid by the countries. But once you bring it to the town, then the operators, the data centers, start to take it and distribute the bandwidth from there. They start down what we call the middle mile. That’s anywhere from a few kilometers to 20 kilometers of fiber. Now in some cases they will be passing very close to a home. In some cases, they’re a little bit further out. That’s the last mile. Which is not necessarily a mile. In some cases, it’s as short as 50 meters. Does Taara cover the whole length of the middle mile?Krishnaswamy: Today Taara operates where we are able to bridge connections from a few kilometers to up to 20 km. That’s the middle mile that we operate in. And almost 50 percent of the world today is within 25 km of a fiber point of presence. So it’s very much accessible for us to reach most of those communities.Now the next generation technology that I’m talking about, the photonics chip, will allow us to go even shorter distances and will allow us to close the gap on the last mile as well. So today we are mostly operating in the middle mile, and in some cases we can connect the last mile. But with the next generation chip, we’ll be operating both in the middle mile as well as the last mile.What about the Google X background? Do you have people from Project Loon or from Google Fiber now working at Taara?Krishnaswamy: Yes. I was personally working on Project Loon, and I was leading up the manufacturing, the supply chain, and some of the operational aspects of it. But my passion was always to solve the connectivity problem. And at X we always say, fall in love with the problem, not the solution per se. So you started using Project Loon’s open-air signaling tech that connects one Internet balloon to another, but you just did it between fixed stations on the ground? Krishnaswamy: Yes, the idea was very simple. What if we were to bring the technology connecting balloons in the stratosphere down to the ground, and start connecting people quickly?It was a quick and dirty way of getting started on connecting and closing out the digital gap. And little did I know that across the street, Google Access was also working on similar technology to cross freeways. So I pulled together a team from Google Access and then from Project Loon. And today the Taara team includes people from various parts of Google who worked on this technology and other connectivity projects. So it’s a team that is really passionate about connectivity globally. The Challenges Ahead for Free Space Optical TechOK, so what about foggy days? What about rain and snow? How does Taara technology send over-the-air infrared data traffic through inclement weather? Krishnaswamy: Our biggest challenge is weather, particularly particulates in weather that disperse light. Fog is our biggest nemesis. And we try to avoid deploying in foggy areas. So we built a planning tool that allows us to actually predict the anticipated availability. As long as it’s light rain, and it doesn’t disperse [optical signals], then it’s fine.A simple rule of thumb is if you can see the other side, then you should be able to close the link. We’re also exploring some smart rerouting algorithms, using mesh. Ultimately, we are subject to some environmental degradations. And it’s really how you overcome that is what we’ve been focusing on.Why 20 km? Is Taara trying to extend that to greater distances today? Krishnaswamy: The honest truth is it started out with one of our first customers in rural India who said, “I have many of these access points which are up to 20 km away.” And as we started to dig deeper, we realized we can connect a vast majority of the unconnected places within 20 km of a fiber point of presence. So that ended up becoming our initial specification. How about pointing? If you’re beaming a laser out over 20 km, that’s a tiny target to aim at. Krishnaswamy: When we deployed first in India, we ran into a lot of monkeys that we had to deal with who are territorial. There would be like 20 or 30 of these monkeys jumping and shaking the tower, and our link would always oscillate. So we can’t physically drive them away. But we could actually improve our pointing and tracking, which is exactly what we did. So we have gyroscopes and accelerometers built in. We are constantly monitoring the other side. There’s also a camera inside the terminal. So if you are really out of alignment, we can always repoint it again. But basically we have made a significant amount of improvements in our pointing and tracking. That’s one of our secret sauces.What are the near-term hurdles for the company? Near-term ambitions? Krishnaswamy: I used to work at Apple, so I brought some of the best practices from there as well to make this technology manufacturable. We want physics to be the upper bound of what is capable, and we don’t want any compromises. And the last thing I’ll say is we are really pioneering the light generation. This is a complete relook at how light can be used for communication purposes, which is where we’re starting out. When you have something this small, that could deliver such high speeds at such low latencies, you could put it into robots and into self-driving cars. And it could change the landscape of communications. But if you were to not just use it for communication, it could go into LIDAR or biomedical devices that scan and sense. You could do a lot more using the underlying technology of phased arrays in a silicon photonics chip. There’s so much more to be done.

View Entire Post