Amazon status: access issues and outage reports

Problems in the last 24 hours

The graph below depicts the number of Amazon reports received over the last 24 hours by time of day. When the number of reports exceeds the baseline, represented by the red line, an outage is determined.

Most Reported Problems

The following are the most recent problems reported by Amazon users through our website.

• Website Down (46%)
• Errors (29%)
• Sign in (25%)

Live Outage Map

The most recent Amazon outage reports came from the following cities:

City	Problem Type	Report Time
Marquette	Website Down	1 hour ago
Boston	Sign in	6 hours ago
Bordeaux	Errors	13 hours ago
Gonesse	Errors	16 hours ago
Mexico City	Sign in	16 hours ago
Marquette	Website Down	19 hours ago

Full Outage Map

Community Discussion

Tips? Frustrations? Share them here. Useful comments include a description of the problem, city and postal code.

Beware of "support numbers" or "recovery" accounts that might be posted below. Make sure to report and downvote those comments. Avoid posting your personal information.

Amazon Issues Reports

Latest outage, problems and issue reports in social media:

• 
  Perspez (@perspez) reported 39 seconds ago

  $SLNH THE AI INFRASTRUCTURE BUILDOUT – Part 2/8 The Hyperscalers Raise the Stakes The central question is no longer whether Microsoft, Meta, Amazon and Google can afford to keep building. It is whether the economic returns can develop fast enough to support another round of even larger investments. Their latest reports show that demand continues to run ahead of available capacity. Microsoft offered the clearest example. Azure grew 43%, and its annual revenue exceeded $100 billion for the first time. Even after adding substantial infrastructure, Microsoft said customer demand remained above the capacity it could supply. When software and engineering improvements released additional CPU or GPU resources, that capacity was absorbed almost immediately. Microsoft is not expanding solely in anticipation of a distant market. Much of the demand is already waiting for the infrastructure to arrive. Google shows how the same investment can support several businesses at once. Cloud revenue rose 82%, operating profit expanded sharply and backlog reached $514 billion. The company is also deploying its models across Search, advertising and consumer products, spreading the potential return across a much broader platform. Alphabet nevertheless remained supply constrained and raised its capital-spending outlook to accelerate capacity delivery. Amazon provides the clearest view of how far in advance customers are securing compute. AWS revenue increased 37%, while backlog climbed to $496 billion. Amazon expects capital spending to reach $220 billion in 2026. Management still does not expect to satisfy all anticipated demand in 2026 or 2027. A large share of future compute capacity has already been reserved, with some commitments extending into 2028. Meta represents a different economic model. Microsoft, Amazon and Google can rent scarce compute directly to customers. Meta must earn its return indirectly through stronger advertising, higher engagement, better recommendations and new products. Revenue rose 28%, while the company maintained a capital-expenditure outlook of $130–145 billion despite rising depreciation and pressure on near-term cash generation. The monetization models differ, but the investment decision is broadly the same. All four companies are willing to absorb substantial costs today because failing to secure enough capacity could limit growth across their most important businesses. These commitments now extend far beyond servers. The hyperscalers are reserving chips, memory and networking equipment while also securing land, energy, buildings and development capacity years before many sites begin producing revenue. That delay changes how the spending appears in financial statements. A data center can consume capital for years before opening. Once operational, it may support revenue for decades. Free cash flow can therefore weaken during the construction phase even while underlying demand remains strong. The more serious risk is not necessarily a sudden collapse in demand. It is a mismatch between the infrastructure being built and the economics available when it opens. A campus delivered late, at excessive cost or around an architecture that has already changed can produce disappointing returns even in a growing market. Not every announced project will justify the capital invested in it. For now, the reports describe four companies trying to avoid a more immediate problem: allowing limited capacity to restrict growth during a major technology transition. The hyperscalers are not retreating from the buildout. They are increasing their commitments and drawing the rest of the infrastructure chain forward with them. ■ Part 3 – GPUs, CPUs and the Expanding Compute Layer: How denser and more powerful systems reshape everything built around them.

• 
  afmb (@afmb) reported 5 minutes ago

  @AmazonHelp I’m being made to wait for 48 hours before it’s sorted. And your supplier must have known the goods were damaged as the broken bits don’t match up.

• 
  Snot (@Spanky_unltd) reported 5 minutes ago

  Never once have had an issue in my entire life with FedEx, UPS, or Amazon.

• 
  angie nichols (@shabbasgal) reported 6 minutes ago

  @CPFC this is a long shot but I’ve been let down my someone on Amazon. It’s my son’s 13th birthday tomorrow and as a present I had brought a team signed print of the team to frame . Well this should have been delivered today and it hasn’t. The sender isn’t responding . He’s a big

• 
  B85! (@CZseventyfive) reported 9 minutes ago

  @Basiofon @CultureCrave The problem with going streaming. Ppl already have HBOMAX, Supergirl wouldn’t have brought new paying subscribers to that platform. Getting it in Amazon would mean 1 purchase for an entire family instead of 2-4 individual ticket purchases. It would have lost WB even more money

• 
  spara10 (@spara10twitch) reported 16 minutes ago

  @AmazonHelp The problem is that the beast of reincarnation Game says it’s arriving today but it hasn’t even changed to being like dispatched or shipped or anything. It’s just saying the order has been placed in it and the bar hasn’t moved so I don’t know if the Game is actually turning up.

• 
  SirSaintTylan (@SIRSAINTTYLAN) reported 18 minutes ago

  @amazon I'm using Amazon years and I have never had an issue like this where they give a tracking number and it shows no information I got as much information I need to know on it that employees are stealing at the warehouse

• 
  Alexander Morris (@alexmorrisfm) reported 20 minutes ago

  If you’re looking for guidance on SpaceX profit outlook, look no further than OG Rocket Man himself Elton John: “I think it’s gonna be a long, long time.”  SpaceX on Tuesday reported its first quarter as a public company and blew past estimates. Revenue was $7.8 billion, up 92%, and about $1 billion ahead of forecasts. On an adjusted EBITDA basis, SpaceX is profitable both in its Starlink and AI businesses (not so much on the space segment); on a GAAP basis SpaceX lost $541 million in the quarter.  Thanks to its record-breaking $86 billion June IPO, SpaceX has $100 billion in cash, equivalents, and marketable securities. (Also contributing to the total: $38 billion in long-term debt.) That’s important, because SpaceX is spending at a prodigious rate. And for that you can credit Elon’s rivalry with Anthropic, OpenAI, Microsoft, Google, and others in frontier AI, where he sees a $26.5 trillion total addressable market, according to the SpaceX prospectus.  For the June quarter, SpaceX reported CapEx of $18.4 billion, or more than 235% of revenue. The 92% revenue jump was impressive but paled versus a 550% increase in CapEx. Most of that went to – surprise! – AI data centers. Q2 AI CapEx was $15.8 billion, six times its AI revenue. CFO Bret Johnsen expects that pace to hold for a couple more quarters. Meanwhile, Elon’s ambitions keep growing. On the call he declared Starlink’s designs on the U.S. mobile telephony market now dominated by AT&T, Verizon, and T-Mobile. That’s going to take billions more in capital – maybe hundreds of billions. Bring it on I say: more than two bars around town is a fantasy for DC folks.  Elon is an astonishingly skilled entrepreneur. He built a car company (as a later-stage investor). He built a rocket company (from scratch). He famously sold flamethrowers (yep). He wants to build a colony on Mars (many yeps). He plans a network of 1 million orbital data centers (so far he has 10,000 satellites in orbit, never mind the space junk risk).  Elon’s business pushes the laws of physics to the brink, but he can’t defy them. And gravity is pulling on SpaceX shares. Launched at $135, peaking shortly after the offering at $225.64; the stock was down more than 10% Wednesday, at $112 and change.   Even now, SpaceX sports a market cap of $1.5 trillion, the 10th most valuable in the world. And on Thursday, more than 900 million SpaceX shares which had been subject to post-IPO lockup agreements became free trading – more than doubling the float.  Look, Elon believes. He thinks SpaceX will reach a $100 billion revenue run rate by year end. He sees $1 trillion in annual revenue by 2030. The highest revenue companies in the world right now? Amazon and Walmart, both north of $700 billion. A trillion, really? Not impossible, but I’m with Elton: I think it’s gonna be a long, long time.  -- This post is for informational purposes only and does not constitute investment advice, a recommendation, or an offer to buy or sell any security.

• 
  Jagat Kumar (@jagatprusty_) reported 26 minutes ago

  @AmazonHelp The issue has resolved. Please make sure sellers don’t cheat people using your platform.

• 
  loupl 🏴‍☠️ (@cortexbt) reported 29 minutes ago

  JEFF BEZOS REVEALED THE ONE DECISION RULE THAT KEPT A TRILLION-DOLLAR COMPANY FROM MOVING LIKE A GIANT. Most companies slow down as they grow. His name is Jeff Bezos. He built Amazon by treating decisions as two different kinds of doors. Some are one-way doors. They are hard to reverse, so they deserve time. But most decisions are two-way doors. You can test them, reverse them, and try again. Bezos said those decisions should be made quickly, often with about 70% of the information you wish you had. The mistake is using permanent-decision caution on reversible decisions. Watch Bezos explain the framework that kept Amazon fast while it became one of the largest companies on earth.

• 
  Miles Insights | Amazon Influencer (@Miles__Insights) reported 31 minutes ago

  @FrontRowContent pricing by insecurity costs the most long term, the number you say out loud the first time tends to stick as your rate for a year. the amazon version of polishing forever is holding a video back to fix something nobody checks in review, spec compliance matters more

• 
  Home Shield (@HomeShieldiOS) reported 33 minutes ago

  Tired of surprise repairs eating your savings? 🏠 Home Shield’s smart reminders keep you organized, while quick Amazon links help grab supplies fast—saving time & money before issues grow. 💪 #HomeCare #SmartHome

• 
  Govind Ram (@GovindRam69493) reported 33 minutes ago

  @AmazonHelp Your support is incapable of resolving this. I got the following message from your AI support "Customer Service Chat is still learning and will be able to solve issues like this soon."

• 
  Shailendra (Shai) Upadhyay (@supadhyaymd) reported 35 minutes ago

  @AmazonHelp Your service is extremely terrible. We have already spent more than five hours either on the phone or by text messages on Amazon do not relation. I would like an immediate call from a representative and not let this harassment continue

• 
  hamburglar (@fatfuck123456) reported 35 minutes ago

  @DahhHavin @Isaac_Dev0 @ayeejuju really? all of the drivers I ran for would tell me that working at amazon was terrible

• 
  Laura (@sloeginpsyop) reported 36 minutes ago

  @reddit_lies Well, these men are feminists and in this household they split all work down the middle, so they’re like “Who checked out the noise last time?” And he and the wife sit and argue about it until she’s like “Fine, I’m a strong independent woman” and she grabs a baseball bat and goes downstairs and finds the Amazon package of the Lego set he ordered for himself and the 186th new drink powder she ordered that was dropped off at 11PM and slammed against their door.

• 
  Dark Factor (@DarkFactorr) reported 36 minutes ago

  Netflix released more than 100 million movie ratings and offered $1 million to anyone who could predict the missing ones. To win, a team had to improve Netflix’s existing recommendation algorithm by at least 10%. Three years later, BellKor’s Pragmatic Chaos crossed the line with a 10.06% improvement. MIT professor Gilbert Strang explains the mathematics underneath that competition in a free linear algebra lecture. Netflix’s data can be arranged as a matrix with users on one side and movies on the other. Every known rating fills one cell. But no subscriber had watched all 17,770 movies, so almost the entire matrix was empty. The challenge was to infer what a person would have rated a movie they had never seen from the relatively small number of ratings they had already provided. The obvious approach is impossible. With billions of missing entries, there are infinitely many ways to complete the matrix. Strang shows why the problem becomes solvable once you assume that human taste has a simpler hidden structure. People may appear to make millions of independent choices, but many ratings are driven by a smaller number of recurring preferences: genre, actors, directors, mood, pacing or style. That means the enormous matrix can be approximated using fewer underlying patterns. Strang connects this to low-rank approximation, singular values and the nuclear norm. Instead of memorizing every rating, the algorithm searches for the compact structure capable of generating them. The winning Netflix system was ultimately an ensemble of models, not one nuclear-norm calculation. But matrix factorization was central to the competition, and Strang’s lecture explains why recommendation data can be treated as an incomplete matrix in the first place. Watch the moment he moves from one unfamiliar matrix norm to a million-dollar Netflix competition. Within minutes, a page of missing numbers becomes the machinery behind recommendations from Netflix, Amazon and nearly every platform trying to predict what you will choose next.

• 
  CF19BUZZKILL 🏴‍☠️ (@cf19buzzkillttv) reported 39 minutes ago

  I've been working on a Space Opera type of book for a very long time and I'm thinking about just publishing it on Amazon as is errors and all because I can't afford an editor. Maybe if I make enough money of the book and X I'll get it edited and re release it. IDK. Am I retarded?

• 
  🇺🇸Salty Cat Chronicles (@SaltyCatChron) reported 40 minutes ago

  ATTENTION ENTREPRENOURS, PSA BY SALTY CAT: I’m one of those people who gets into things just to understand how they work, and lately I’ve been digging into the drop‑shipping world: Printify, Amazon, all of it. What I’ve learned is this: it’s not a good value for the consumer, and it’s even worse for the small creator trying to build something online. These companies get great deals on mass‑produced, sublimation‑ready white mugs. Nothing wrong with the mugs themselves, I have some that are 25 years old and still going strong. The problem is the “production costs.” They tell you a mug will cost {x}, but when the order goes through, suddenly it costs more. No explanation, no consistency. That means the seller has to set a much higher profit margin just to avoid surprise losses. Otherwise, you can literally sell a mug and still owe money to Printify, Amazon, etc. The math simply doesn’t math. No worries... I’m keeping the shop open. Just know the prices will be much higher because POD shipping is expensive and unpredictable. If you love the art enough to want it on a mug, the link stays in the bio. If not, enjoy the images, that part is free. Personal note: I’m retired, so I’m grateful I can run these experiments without worrying about bills or making a living. My heart goes out to the small entrepreneurs trying to survive in this system. PS: There are many ways in which these companies bleed entrepreneurs until they cannot survive...subscription fees, "seller clubs" and additional fees never mentioned upfront are some of them.

• 
  Kay (@Love_This_City) reported 41 minutes ago

  @AmazonHelp Wasn't online to see your link. It's not all drivers but enough to be a problem. Deliveries to 60607 need to be educated on civility

• 
  Joecamel (@joecamel90) reported 42 minutes ago

  @GordMagill Passed an Amazon this morning with his front end doing the hula and the fender mirrors bouncing up and down a foot. Totally abusing that poor Pete, just a question of where and when the inevitable happens.

• 
  I Kaya (@kaya85kaya) reported 43 minutes ago

  @sjb987654321 @edels0n Issue is if Ai buildout of Big Ai of ~$6T in the next 4 years is to happen. Maybe half can come from core business cashflow leaving $3T to come from debt or new equity issues If $1T can be borrowed That means a wall of $2T of stock sales from Meta Google Amazon Oracle etc

• 
  Bitcoin Clown 🤡 (@bitcoin_clown) reported 50 minutes ago

  Some of the biggest companies in the world are each borrowing $25B: Oracle → $25B SpaceX → $25B Amazon → $25B Google → up to $25B The ones who don't need the money get it cheapest. The ones who do get denied — or pay 24% on a credit card. The system isn't broken. It's working exactly as designed. 🤡

• 
  HELL ON EARTH✰ (@BandmanNay) reported 55 minutes ago

  Dog ****** be in Amazon crispy asf Amiri down😭😭😭😭

• 
  Shel In Ohio (@Shel_Is_Tired) reported 58 minutes ago

  I'm seeing a lot of people talking about the fact that they're buying school supplies not just for their child, but for the classroom. Back when my daughter was in school 20 years ago I think that "buy for everyone" thing started to take hold. Whereas when I was a kid, our folks bought us a pack of pencils, a pack of notebook paper, folders for whatever subjects we had, a box of crayons etc. When my daughter was in school, it started small...just the normal supplies for HER, but also, "Please buy a box of tissues to use in the classroom" or "please buy a box of 1 gallon ziploc bags for classroom use". Now I'm seeing full-blown Amazon lists coming from teachers to the children's parents. All the while Administrators are pocketing more "salary" than ever. I'm over it. Get your kids out of public schools, not only for their actual education, but this bureaucratic nonsense of pooring sand down a rat hole.

• 
  Maltesh (@malteshk) reported 58 minutes ago

  @AmazonHelp Hello, i dont know what is happening. Can someone update me the real status? I have received tons of mails and again i have been asked to go through the return process in app which is not working.

• 
  Lee Stetson🇺🇸⚓#🟦 (@flankspeed_) reported 59 minutes ago

  @LOP_LWO @Vince39979839 Amazon has MGM and Netflix runs into antitrust issues with being the #1 streamer. You think theatrical releases are more likely under WBD/PSKY or Netflix?

• 
  Maduforo _Godwin (@EmekaGo69635875) reported 59 minutes ago

  Every serious KDP publisher uses pen names. Not to deceive anyone..... Amazon knows exactly who owns the account, It's standard publishing practice. King published as Bachman and Nora Roberts publishes as J.D. Robb..... For KDP, pen names do three specific things... They separate your niches; A reader buying one type of book from you shouldn't land on an author page that also sells something completely different. Different audiences expect different things, mixing them under one name creates a credibility problem you can't undo. They contain damage; A bad review on one title doesn't follow you to every other book in your catalog. They let you scale without looking like a content factory; Ten books across several author names looks like ten independent publishers doing their work. Ten books under one name invites scrutiny. One rule that matters: your pen name cannot closely resemble a well-known author already in your niche. That's a KDP policy violation, not just bad practice so always check on amazon if that pen name is used already.

• 
  Mark Schultz (@wordrefiner) reported 60 minutes ago

  @reeseys They already tried. They scraped my website one night, thousands of hits! It didn't do them any good because every book published has spelling, punctuation or other errors. I like your idea about returning books written by AI. Amazon would stop publishing them if they lost money.

• 
  ANANTAPUR CINEMA (@AnantapurCinema) reported 60 minutes ago

  @AmazonHelp Not working still getting same page once it went to support page @AmazonHelp