Amazon Thinks the Future of Data Centers Depends on a Technical Problem It Just Solved | WIREDSkip to main contentMenuSECURITYPOLITICSTHE BIG STORYBUSINESSSCIENCECULTUREREVIEWSMenuAccountAccountNewslettersSecurityPoliticsThe Big StoryBusinessScienceCultureReviewsChevronMoreExpandThe Big InterviewMagazineEventsWIRED InsiderWIRED ConsultingNewslettersPodcastsVideoLivestreamsMerchSearchSearchLauren GoodeBusinessMay 28, 2026 6:30 AMAmazon Thinks the Future of Data Centers Depends on a Technical Problem It Just SolvedThe tech giant says a breakthrough in data-center networking has dramatically accelerated the flow of information through its massive cloud infrastructure.Courtesy of AmazonCommentLoaderSave StorySave this storyCommentLoaderSave StorySave this storyAmazon says it recently achieved a major breakthrough in networking design—and has been quietly deploying the new technology in its data centers since late last year. The company claims it has significantly increased data speeds while reducing energy use, potentially giving the tech giant an edge as companies race to build ever-faster systems in the cloud.The new technology hinges on a “quasi-random” design that combines elements of traditional, structured data networks with the performance advantages of more random architectures. Researchers have explored random networks for decades, but the technology has never been successfully scaled. Now, Amazon thinks it has cracked the code.The fact that Amazon is using this in the real world is “remarkable,” says Brighten Godfrey, a computer science professor at the University of Illinois at Urbana-Champaign and an expert in networking, who was not involved in Amazon’s research. Godfrey co-authored a seminal 2012 paper on random network graphs, which he says are a “mind-bending problem to solve, in general.”A team of engineers and researchers at Amazon Web Services, including several recruited from academia, has been working on the random networking problem since 2023. Amazon also designed a new piece of data center equipment it dubbed the ShuffleBox, which automatically shuffles the cables required for this kind of networking.“By essentially flattening the network, we eliminated the bottlenecks that come with traditional networking designs,” Matt Rehder, vice president of AWS Network Engineering, said in an exclusive interview with WIRED. “We think we’re the only ones who have done this at scale.”Courtesy of AmazonNetwork EffectsAmazon detailed its new networking design in a paper published last month titled “RNG: Flat Datacenter Networks at Scale.” RNG stands for “resilient network graphs,” which are neither entirely structured nor entirely random.Interestingly, the Amazon team behind RNG isn’t making this networking pitch around generative AI. This is about making the company’s everyday data center architecture more efficient. “RNG is a great fit for our core demands, but AI training data patterns are far more coordinated and centrally orchestrated, so they don’t approximate a random graph,” Rehder says.Since the mid-1980's, communications networks—from telecom to to data centers—have been predominantly designed with a “fat-tree” topology, which includes two or three vertical layers of switches and routers. These are connected by “fat” nodes at the top of the structure, where there are multiple routers of the same type, and thinner branches towards the bottom. Put very simply, in a fat-tree network, data moves up and down the stack. The increased bandwidth near the top of the structure, where the data bisects, helps eliminate bottlenecks.Over time, the tech industry has developed and deployed variations on the fat-tree architecture. But the design has room for improvement. It’s generally reliable, but also rigid, inefficient, and requires complex cabling. As in, actual physical cables.If you’ve ever been in a data center or an office building’s server room, you’ve likely seen nests of colorful cables spilling out of metal racks. Cabling is one of the greatest costs in networking, Rehder says, and Amazon’s global data centers are currently connected with 20 million kilometers of fiber optic cables. That’s roughly the distance it would take to travel from Earth to the moon and back 25 times.In 2012, as the demand for cloud computing services was exploding, a group of researchers at University of Illinois Urbana-Champaign, including Godfrey, introduced a concept known as Jellyfish. Fixed network designs in use at the time were struggling to meet growing demand, so the researchers proposed a “high-capacity network interconnect which, by adopting a random graph topology, yields itself naturally to incremental expansion.” They believed this random approach could be more efficient and scalable than networks built using the fat-tree architecture.“We gave it the name Jellyfish because it’s fluid,” Godfrey says. “You can connect the routers and switches randomly and it becomes this flexible pool of network capacity, which is very efficient.”However, Jellyfish also introduced new challenges in layout, data routing, and cabling. Routing in random graphs is trickier, Godfrey says, because there are many more and diversified paths that data can take from its source to its destination. Cabling is harder because the endpoints of the cables are chosen randomly.A couple of years later, Google began toying with another solution: It started integrating optical circuit switching, or OCS, into its network designs. This approach uses tiny mirrors to reflect light from an input port to an output port, which lets Google refigure optical cabling in real-time. But, again: This adds a certain amount of engineering complexity, as well as cost.Courtesy of AmazonCourtesy of AmazonSo RandomAmazon, meanwhile, was searching for the “holy grail,” says Giacomo Bernardi, who is one of the lead authors on the new paper, along with Amazon Scholars Ratul Mahajan and C.S. Seshandhri. In an ideal world, a data network would be flat and efficient, resilient to hardware failures, random enough to maximize performance, and scalable enough to grow without becoming unwieldy. It would also rely on simpler, streamlined cabling rather than increasingly complex fiber-optic systems.When he and his colleagues began trying to build such a network, Bernardi says he had already become obsessed with Penrose tiling, a kind of aperiodic tiling named after the British physicist Roger Penrose. (Other researchers have been so inspired by Penrose tilings that they’ve tried to translate the patterns into error-correcting code in quantum computers.) Bernardi wondered if Amazon could use a similar construction and create a flat “mesh” by following a repeating pattern. He and his team tried building a simulation of what that might look like.It didn’t work. Penrose tiling was promising on paper, Bernardi says, but the simulated data network was unreliable, and the researchers didn’t achieve the efficiency gains they had hoped for. They achieved better outcomes, he realized, when they replaced the more structured parts of the network design with randomness. “We ‘embraced the chaos,’ and adopted a quasi-random approach,” Bernardi explains.Courtesy of AmazonA critical component of this design is the ShuffleBox, a new optical device Amazon developed that mixes connections between routers internally. During a short tour of one of Amazon’s networking labs in Cupertino, WIRED was able to view the chaotic bundle of cables running between routers in a traditional fat-tree structure, and compare them side-by-side with the tidy waves of cables being run through ShuffleBoxes in the new design.Rehder says Amazon’s RNG design has made the company’s data centers both more efficient and more resilient. Compared to traditional networks, he claims it uses 69 percent fewer routers and switches, delivers 33 percent higher data throughput, cuts network power consumption by 40 percent, and lowers operating costs by 27 percent.The first instance of RNG was unleashed in a Dublin data center in 2024. Amazon then expanded the technology to data centers in Germany and Spain. The company says that now, most newly-built data centers are being outfitted with the RNG networking protocol.CommentsBack to topTriangleYou Might Also LikeHow to find us: Add WIRED.com to your preferred sources in GoogleThese women are trying to optimize their vaginasBig Story: AI gig work is the new waiting tables—and it's soul-crushingThis summer, the American water crisis becomes realEvent: How to adapt, compete, and win in the next era of businessLauren Goode is a senior correspondent at WIRED covering all things Silicon Valley, including artificial intelligence, semiconductors, venture capital, startups, workplace culture, and tech's most interesting people and trends. Previously she worked at The Verge, Recode, and The Wall Street Journal. Please send story tips (no PR pitches) to ChaoticGoode.12 ... Read MoreSenior CorrespondentXblueskyTopicsAmazonAWSdata centersCloud ComputingresearchmathRead MoreHuawei's ‘Chip Queen’ Throws Down the GauntletThe Chinese company is adapting to the demise of Moore’s Law, which guides chip production. It could complicate US chip dominance.Will KnightAnthropic Gets in Bed With SpaceX as the AI Race Turns WeirdIn an unexpected turn, the two companies signed a deal for Anthropic to use computing resources from Elon Musk’s xAI.Lauren GoodeWelcome to the Great American Satellite AgeA new generation of satellite startups in San Francisco is racing to capitalize on recent technological breakthroughs in space-based data collection and communications.Paresh DaveThe Gulf’s AI Boom Has an Undersea Cable ProblemHyperscalers are pushing the Gulf to rethink internet infrastructure as AI raises the stakes of cable disruptions.Chris Hamill-StewartMira Murati Wants Her AI to ‘Keep Humans in the Loop’The Thinking Machines Lab founder and former CTO of OpenAI tells WIRED she isn’t interested in automating people out of jobs. Instead, she’s building AI that can collaborate.Will KnightMeet Rassvet, Russia’s Answer to StarlinkWith the launch of the first 16 satellites, Russia begins construction of a network for satellite internet that aims to cover the entire country by 2030. But getting there won’t be easy.Lucia BellinelloGreg Brockman Officially Takes Control of OpenAI’s Products in Latest Shake-UpOpenAI is once again reorganizing its executive ranks as part of its effort to unify ChatGPT and Codex into one core product experience.Maxwell ZeffDiabetes Detection Needs Better Tools. They’re on the WayBlood glucose levels can be a reliable indicator of diabetes risk. But in some populations, it's not enough to catch the disease early.Anna McKieTesting for ‘Bad Cholesterol’ Doesn’t Tell the Whole StoryThere’s a more accurate way of measuring who’s at risk for cholesterol-related health issues. So why don’t more doctors use it?Anna McKieDemis Hassabis Thinks AI Job Cuts Are DumbThe CEO of Google DeepMind tells WIRED that companies should use the productivity gains of AI to do more, not lay people off.Will KnightHow Your Modem and Router Work Together to Get Your Devices OnlineYour modem and router are the dynamic duo you need to get online, so long as you don't mistake one for the other.Simon HillWhat Microsoft Executives Really Thought About OpenAI in 2018Leaders at the tech giant were skeptical of OpenAI—but wary of pushing it into the arms of Amazon, according to evidence revealed during the Musk v. Altman trial.Maxwell ZeffWIRED is obsessed with what comes next. Through rigorous investigations and game-changing reporting, we tell stories that don’t just reflect the moment—they help create it. When you look back in 10, 20, even 50 years, WIRED will be the publication that led the story of the present, mapped the people, products, and ideas defining it, and explained how those forces forged the future. WIRED: For Future Reference.More From WIREDSubscribeNewslettersLivestreamsTravelFAQWIRED StaffWIRED EducationEditorial StandardsArchiveRSSSite MapAccessibility HelpReviews and GuidesReviewsBuying GuidesStreaming GuidesWearablesCouponsGift GuidesAdvertiseContact UsManage AccountJobsPress CenterCondé Nast StoreUser AgreementPrivacy PolicyYour California Privacy Rights© 2026 Condé Nast. All rights reserved. WIRED may earn a portion of sales from products that are purchased through our site as part of our Affiliate Partnerships with retailers. The material on this site may not be reproduced, distributed, transmitted, cached or otherwise used, except with the prior written permission of Condé Nast. Ad ChoicesSelect international siteUnited StatesLargeChevronItaliaJapónCzech Republic & SlovakiaFacebookXPinterestYouTubeInstagramTiktok

Amazon has reportedly achieved a significant breakthrough in data center networking by solving a complex technical problem, leading to dramatic improvements in information flow within its cloud infrastructure. The core of this innovation lies in a networking design based on a "quasi-random" architecture, which successfully merges the reliability of traditional, structured data networks with the performance benefits offered by more random organizational structures. This advancement allows the company to increase data speeds while simultaneously reducing energy consumption, potentially providing a competitive advantage in the rapidly evolving cloud computing race.

The development of this technology stems from research into random networks, a field explored for decades, although scaling these concepts has previously proven challenging. Amazon’s approach introduced a new design called RNG, which stands for Resilient Network Graphs at Scale. This approach is neither purely structured nor purely random, aiming to create a highly efficient and resilient network topology. The practical implementation involves a novel piece of data center equipment called the ShuffleBox, which automatically manages and shuffles the necessary cabling connections between routers. This mechanism effectively flattens the network structure by eliminating the bottlenecks inherent in traditional networking designs.

Historically, communications networks, including those in data centers, have largely utilized a fat-tree topology, characterized by multiple vertical layers of switches and routers connected by "fat" nodes at the top. While this structure manages data flow by routing it up and down the stack, it is often rigid, inefficient, and requires complicated physical cabling, which represents a major cost in networking for entities like Amazon, which currently utilizes twenty million kilometers of fiber optic cables globally.

Early efforts to challenge this architecture focused on random graph topologies. In 2012, researchers, including Brighten Godfrey, proposed a concept called Jellyfish, which advocated for a random graph topology to allow for natural, incremental expansion. However, these initial random network designs introduced complexities concerning data routing and physical cabling endpoints. Subsequent attempts to introduce dynamic reconfiguration, such as Google's use of optical circuit switching, added further engineering complexity and cost.

The pursuit of an ideal data network involves achieving a state that is flat, highly efficient, resilient to hardware failures, random enough to maximize performance, and scalable without reliance on overly complex fiber-optic systems. Researchers attempting to model this ideal structure explored concepts like Penrose tiling to create repeating, aperiodic patterns, but simulations did not yield the expected efficiency gains. Instead, the researchers found that embracing chaos and adopting a quasi-random approach led to superior outcomes when replacing structured sections of the network design with randomness.

Amazon’s RNG design, facilitated by the ShuffleBox technology, demonstrates substantial operational improvements when compared to traditional networks. According to information provided by Amazon’s leadership, this new networking protocol results in 69 percent fewer routers and switches, delivers 33 percent higher data throughput, reduces network power consumption by 40 percent, and lowers overall operating costs by 27 percent. This technology was first implemented in a Dublin data center in 2024 and has since been expanded to data centers in Germany and Spain, with the goal of equipping most new facilities with the RNG protocol.