Alphabet’s “moonshot factory,” known as X, has long embraced bold and unconventional ideas in its projects. One of the most ambitious was Loon, which aimed to provide internet access using a fleet of high-altitude balloons. Ultimately, Loon was spun off into a separate division within Alphabet, but the company concluded that the business model wasn’t sustainable. By the time Loon ceased operations in 2021, a former engineer on the team had already transitioned to a new initiative focused on an innovative data transmission method—specifically, using laser beams to deliver high-bandwidth internet. Imagine fiber optics, but without the cables.
While this concept isn’t entirely new, the last few years have seen Taara—this project from X—quietly refine its practical applications. Now, Alphabet is introducing a next-generation component for the technology—a chip—that it claims will not only make Taara a feasible option for fast internet delivery but might also herald a new era where light performs many of the roles currently held by radio waves, only at greater speeds.
Mahesh Krishnaswamy, the former Loon engineer heading up Taara, has been captivated by the topic of connectivity ever since he first gained internet access as a student in Chennai, India—traveling to the US embassy to use a computer. “Since that moment, I committed myself to finding ways to connect individuals like myself,” he explains during our conversation at X’s headquarters in Mountain View, California. He eventually moved to the US, worked at Apple for a time, then joined Google in 2013. It was there that he became inspired to explore the potential of light for internet connectivity—not just for links to ground stations but for fast data transfers between balloons. In 2016, Krishnaswamy departed from Loon to establish a dedicated team focused on developing this technology, known as Taara.
I was eager to ask Krishnaswamy about the target audience for this technology. In the 2010s, tech giants like Google and Facebook made significant efforts to connect “the next billion users” through ambitious projects such as Loon and high-altitude drones. (Facebook even contemplated the foundational idea behind Taara—“invisible beams of light… that transmit data 10 times faster than current versions,” as my former colleague Jessi Hempel reported in 2016, but the initiative was quietly terminated in 2018.) However, with various methods now in place, more of the world has access than ever before, which is one factor X cited when discontinuing Loon. Elon Musk’s Starlink has emerged as a global internet provider, with Amazon also introducing its own competitor, Kuiper.
Despite this progress, Krishnaswamy argues that the worldwide connectivity challenge is far from resolved. “Currently, around 3 billion individuals still lack internet access, and it is crucial to connect them,” he notes. Additionally, many people, including those in the US, experience internet speeds that can’t even support basic streaming. Regarding Starlink, he mentions that in densely populated areas, multiple users must share the connection, which results in each person receiving diminished bandwidth and slower speeds. “We can deliver 10 to 100 times more bandwidth to end-users than a typical Starlink antenna, and do it at a fraction of the cost,” he asserts, though it seems his claims are directed towards Taara’s future capabilities rather than its current state.
In recent years, Taara has progressed in implementing its technology effectively. Instead of relying on orbital systems, Taara’s “light bridges”—which are roughly the size of a traffic light—operate on the ground. As Astro Teller, X’s “captain of moonshots,” puts it, “As long as these two apparatuses can see each other, you can achieve 20 gigabits per second, which is on par with a fiber-optic cable, yet without the necessity of burying that fiber.” The light bridges incorporate sophisticated gimbals, mirrors, and lenses to hone in on the target for establishing and maintaining the link. The team has developed methods to mitigate potential interruptions to the line of sight caused by birds, rain, and wind, with fog being the primary challenge. After establishing a high-speed link between light bridges, service providers still need to find traditional means to transfer the data from the bridge to the final device, such as a phone or computer.
Taara has transitioned into a commercial entity, operating in over a dozen countries. One noteworthy success involved spanning the Congo River. One side, Brazzaville, had direct fiber access, while on the other, in Kinshasa, internet costs had been five times higher. A Taara light bridge crossing the 5-kilometer span provided Kinshasa with comparably affordable internet. Additionally, Taara was utilized at the 2024 Coachella music festival, enhancing a cellular network that would have likely been overloaded. Google has also adopted a light bridge to deliver high-speed bandwidth to a new building on its Bayview campus, where extending fiber optics was impractical.
Mohamed-Slim Alouini, a professor at King Abdullah University of Science and Technology and a veteran in optics, describes Taara’s technology as “the Ferrari” of fiber-free optics: “It’s rapid and dependable but significantly expensive.” He noted that he spent around $30,000 on the last light-bridge setup he purchased from Alphabet for testing purposes.
This situation might change with the introduction of Taara’s second-generation product. Engineers at Taara have employed groundbreaking light-augmenting solutions to design a silicon photonic chip that will condense the apparatus used in their light bridges to the size of a fingernail—substituting the mechanical gimbals and expensive mirrors with solid-state circuits—and eventually enabling a single laser transmitter to connect with multiple receptors. Teller emphasizes that Taara’s advancements could instigate a transformation reminiscent of the evolution from tape drives to disks and then to the current solid-state devices.
In the near term, Teller and Krishnaswamy hope to employ Taara technology to furnish high-bandwidth internet in areas lacking fiber connectivity. One scenario might involve delivering premium connectivity to an island community just off the coast or supplying rapid internet after a natural disaster strikes. However, their aspirations extend further; they believe that 6G may become the last generation to employ radio waves. They assert that we are approaching a saturation point on the electromagnetic spectrum. The traditional radio frequency bands are becoming crowded and insufficient to accommodate our increasing need for swift and reliable connectivity. “We are on the threshold of a significant transformation in the global industry,” asserts Teller. He believes that light could be the crucial element that drives the development of 7G technology. (If the anticipation for 5G was intense, the excitement for 7G could be even greater.)
Professor Alouini concurs with this outlook. “Many of us working in the field firmly believe that we will eventually need to depend on optics because the spectrum is becoming congested,” he states. Teller envisions numerous Taara chips forming mesh networks, transmitting light beams in various applications—from smartphones to data centers and self-driving vehicles. “If you accept this perspective, it has the potential to be a monumental advancement,” he concludes.
