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 (47%)
• Errors (28%)
• Sign in (25%)

Live Outage Map

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

City	Problem Type	Report Time
Charlotte	Website Down	6 hours ago
Annecy	Errors	9 hours ago
Santiago de Querétaro	Sign in	2 days ago
Kingston upon Hull	Website Down	2 days ago
Pensacola	Website Down	2 days ago
Santiago de Querétaro	Sign in	2 days 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:

• 
  Anime Updates (@animeupdates) reported 5 minutes ago

  @ItsAnimeJP Apologies Not Disney+ Amazon Prime Video. We made an error when creating the visual, which carried over to the post itself

• 
  Rajesh Pradhan (@RajeshPradhan61) reported 6 minutes ago

  @AmazonHelp @amazonIN @amazon I think you got on 6th deliver another item, if address is an issues how 6th July delivery done?? I asked him i have pending delivery, he said some other sift peron will deliver do not worry. There some core issues on delivery team.

• 
  Alex Koutmos (@akoutmos) reported 7 minutes ago

  I think Jose is 100% on the money here. My only edit to his statement is that I don't even think this is a "we'll see", I think it's "already here". The barrier to entry to publishing libraries now is super low with AI models generating a slop library in an afternoon. There are now many AI generated libraries even on Hex. For many of these libraries it was an afternoon vibe code dopamine hit without any real thought or effort put into the library. It's tough to "trust" that an open source library was written with passion and effort to solve a certain problem. Perhaps Hex should start tagging libraries that were mostly written by AI similarly to how Amazon tags books on KDP? I'm not saying that people shouldn't use AI tools for libraries....but AI should not write the entire library imo. I'm still not thrilled with the code quality of AI models. I think we've hit some real plateaus there and I haven't seen drastic quality improvements since late 2025.

• 
  ank (@ank1091040) reported 8 minutes ago

  @bezosjaff2 @amazonIN @AmazonHelp received a used laptop sold as new. Your support team confirmed in writing that the item is used, and I submitted all the required proof. the Leadership Team has refused to issue a refund. Please investigate and resolve my case ASAP

• 
  Lola la Chola (@KupoPewPew) reported 9 minutes ago

  @worldhuman_2026 @NYCShackleton @Bobbythirdway And a majority of the h1b scammers are paid less than 100k. My husband has been in the middle of this tech revolution for more than 25 years, worked for Microsoft Google amazon and others and he saw it first hand. I had a store down the steeet from Microsoft campus for 8 years and many of my customers where MS employees, and I got to see first hand many of them train and be replaced by these abusers and every single one that stayed had the same **** to talk about the surge of h1bs… low quality and lower wages, so don’t try and gaslight

• 
  CM (@seattleiscrazy) reported 15 minutes ago

  Hey @amazon your customer service sucks!!! Like the worst. You have not solved my problem and I’ve been a customer for decades. Having your service members located abroad who speak poor English is a bad way to go. he didn’t even k ow Seattle Is Amazon headquarters

• 
  Saul (@SaulSellsStuff) reported 16 minutes ago

  @BuyHoldCollect I sell all booked on Amazon as used. Rarely will you run into issues, if you do its almost always textbooks and can be fought.

• 
  markie mark (@StartScratchup) reported 21 minutes ago

  @FrancesJatkfc @StockSavvyShay the wild part isn't that Amazon is spending another $100M+ on AI infrastructure. it's that they're doubling down on a product that's spent years raising questions about its economics. sometimes the value isn't in today's P&L. it's in staying in the race.

• 
  Mr.Goodcat (@BigTenFan2023) reported 25 minutes ago

  @amazon @Chase I signed up for an Amazon prime Visa card, and was supposed to receive a $200 gift card once approved. I still haven't received the gift card, and when I call you both say it's the other company's fault. Please fix this l!

• 
  WanderingMindCo (@mossandmercury) reported 27 minutes ago

  In the midst of the Tyler Robinson prelim and all the talk of DNA, another case comes to mind that I studied. And when I say studied, I mean read every book I could find on it from all sides and even bought the VHS tapes and a VCR off Amazon to watch the entire trial. When the murder happened, I was too young to care. I just remember a white Bronco chase. Why the Robinson case brought up this classic gem of a case is because all the mishandling of DNA and chain of custody in the OJ case. A lot of you who remember the case probably don’t even know the half of it and that the real story isn’t what we seen on TV. The O.J. Simpson case makes a lot more sense if you look past O.J. and look at his oldest son, Jason. In June 1994, Jason was 24 years old, working as a chef, and dealing with severe psychological issues. He had been diagnosed with intermittent rage disorder and was actively writing in his journals about battling a “Jekyll and Hyde” split personality. In fact, at the very time of the murders, he was on probation for assaulting his former boss with a knife, and his diaries from that year contained chilling entries like, “This is the year of the knife for me.” The physical evidence points heavily toward Jason’s profession and personal belongings. As a chef, he was highly trained in using and sharpening tactical knives. While the LAPD never found the murder weapon, a private investigator later purchased the contents of Jason’s abandoned storage locker. Inside, he uncovered a heavy hunting knife. Forensic examination revealed that the blade’s dimensions perfectly matched the entry wounds on the victims, and the heavy butt of the knife matched a distinct blunt-force indentation found on Nicole’s skull. Jason’s alibi from the restaurant that night was never officially verified, and his time cards were suspiciously handwritten rather than machine-stamped. This theory perfectly explains the mixed DNA evidence and O.J.’s bizarre behavior. If Jason was the attacker, O.J. may have arrived at the scene during or immediately after the crime. A scuffle could’ve broken out as O.J. tried to stop his son or handle the situation, which would explain how O.J.‘s blood and DNA became intermingled at the Bundy drive scene. Ultimately, O.J. wasn’t running from the police to protect himself; he was staging a distraction, taking the fall, and using his wealth to hire a Dream Team to protect his own flesh and blood from the death penalty. In my opinion.

• 
  (Light Bringer) + (Black in German) (@CosmicInglewood) reported 27 minutes ago

  @amazon This an Amazon notorious for abusing employees in fulfillment centers, for screwing flex drivers, giving huge odd hard trouble finding weird unposted addresses, replacing people with KIVA systems robots, Q enterprise AI systems, no excuses, its ***** level WTF without cause

• 
  Dave Data Guy | author of 2 stock investing books (@DGretta_Author) reported 27 minutes ago

  Amazon $AMZN customer service contacted me today and they're getting close to fixing technical glitch on their end restocking my book. Buy my book instead of Portnoy's crap book. I'd also highly reco clicking on my X bio link + ordering direct. Same price + faster!!

• 
  KleverKid12 (@KleverKid12) reported 28 minutes ago

  Say what you will about Glitch but at leastbthey're leinant on copyright enough to let reaction videos and reviews slip by Since Spindle is under the watchful eye of A24 and Amazon, nobody's safe They're slowly becoming what they set out to destroy

• 
  Joey Zhuo (@JoeyZhuo777) reported 32 minutes ago

  $MRVL Jensen says it's worth a trillion, the filings say growth to $4 billion Nvidia put $2 billion into Marvell and Jensen threw around trillion-dollar talk, so retail piled in and the stock tripled in two months before giving back a third of it. The question isn't whether the rally was real — it's whether what's left still makes sense at 57 times forward earnings. The bull case runs on three engines, and you need to separate them because they carry very different risk profiles. Engine one is the data center ramp that's already happening. Marvell did $2.42 billion in Q1, data center was 76% of that, and management lifted the full-year guide to $11.5 billion for fiscal 2027 and $16.5 billion the year after. That's 40% growth this year, 45% next year. The beats were tiny — revenue missed by half a percent, EPS by less — so the price move came entirely from guidance, not from the quarter itself. This piece is already in the price. The stock more than doubled because the street re-rated future growth, not because anything surprised in the last three months. If you're buying here, you're not buying a surprise, you're buying the assumption that those guide raises hold up. Engine two is the custom ASIC business, and this is where the real growth lever sits. Marvell pulled in $1.5 billion in ASIC revenue last fiscal year. Management sees that hitting north of $4 billion by fiscal 2028. The customer list is Amazon Trainium, Google Axion, Meta DPU programs, and one unnamed hyperscaler already locked in for 2028 volume. If that 3x in two years actually lands, the multiple starts to make sense. If any of those programs slip or one of the hyperscalers pulls back on custom silicon spending, the math breaks. The risk here isn't technical, it's commercial. Custom ASIC work carries lower gross margins than off-the-shelf merchant silicon, so even if revenue triples, the margin profile compresses unless Marvell has enough scale to offset it. Broadcom is in the same game with better scale and pricing power, so there's no guarantee Marvell holds share or pricing as this market grows. Engine three is the Nvidia partnership, which is the most talked-about and probably the least understood. Nvidia isn't handing Marvell $2 billion out of charity — they're integrating Marvell's custom accelerators and optical connectivity into the NVLink ecosystem. That's a validation of the tech, and it probably gives Marvell some moat in the interconnect layer as AI clusters scale out. But it also makes Marvell dependent on Nvidia's architecture roadmap and their willingness to keep Marvell inside the tent instead of doing it themselves or switching to Broadcom. The agentic AI angle is interesting but speculative. The idea is that when you ask an AI agent a question, it might query a model thousands of times across distributed infrastructure, and every one of those queries runs through Marvell's network interface cards, PCIe retimers, CXL links. The logic holds if agentic AI actually takes off at enterprise scale in the next two years. If it stays mostly experimental, that revenue stream doesn't show up on the timeline the valuation assumes. Marvell sold the automotive Ethernet business to Infineon for $2.5 billion cash, so they're clearly streamlining toward AI and dumping anything that doesn't fit the hyperscaler buildout story. That's the right move if you believe in the TAM, but it also means there's no fallback if data center spending moderates. The valuation math here depends entirely on where you land on fiscal 2028 earnings. The consensus range for that year runs from around $6.18 to over $7. If you take the high end at $7.13 and slap a 40x to 45x multiple on it because growth stays above 40% into 2029 and 2030, you get to around $300 a share, which is 30% above the current post-correction price of $231. That assumes no multiple compression, no margin squeeze from ASIC mix shift, no hyperscaler budget cuts, and no competitive share loss to Broadcom. If any of those assumptions crack, you're looking at a stock trading near 60 times earnings with decelerating growth, and that reprices fast. The trillion-dollar market cap talk is noise. Marvell is a $200 billion company today. To get to a trillion at current valuations, revenue would need to go to something like $20 billion and the market would need to pay 50 times sales, which isn't happening outside of a speculative mania. The real question is whether the path from $11.5 billion this year to $16.5 billion next year to maybe $20 billion-plus in 2029 is credible, and whether the margin structure holds up as the business mix shifts toward lower-margin custom work. I think the direction is right, but the margin for error is thin. If you have a three-to-five-year window and believe hyperscaler spend stays elevated and Nvidia keeps Marvell inside the NVLink stack, there's probably 25% to 35% upside from here over the next year or two. If data center budgets flatten or Broadcom takes more ASIC share, this thing re-rates down to 35x earnings in a hurry and you're sitting on a loss even if the business grows. The analyst consensus price target sits around 9% upside, so most of the street is less optimistic than the bull case laid out here. That spread tells you where the risk sits — it's not in the technology, it's in the execution and the spending environment. Marvell dumped 30% from the June highs because the whole AI semiconductor trade got hit by rotation, not because anything broke in the business. That gives you a better entry than two months ago, but it doesn't change the fact that you're paying a premium multiple for growth that's entirely dependent on a handful of hyperscale customers and one very large partner who could change their mind. If the fiscal 2028 guide holds and the ASIC ramp delivers, this works. If either one slips, the valuation has no cushion. Image source: Seeking Alpha / Danil Sereda

• 
  madhu milan prasad (@MeTooBhartiya) reported 32 minutes ago

  @amazonIN ,@amazon For Order  Id  402-4654835-9146760 return  pick up was declined & no way there on web or app to reach customer  executive service  to tell my problem Regards

• 
  Hot Aisle (@HotAisle) reported 35 minutes ago

  @nickdodd i have prime and spend way too much money on amazon... support hasn't been a problem.

• 
  Marietta Gee்Resourcefuelness Mahendran (@mariettamma) reported 35 minutes ago

  I feel like I am coming to realize that the man who put me down @amazon wanted to be like me, hunny, you can’t be like me so he tried to shun me.

• 
  Amazon Help (@AmazonHelp) reported 36 minutes ago

  @SusheelNughwal Please copy and paste the link into a different web browser clearing cookies/cache and enabling desktop mode or try to access the link via desktop/laptop to connect with our team. Copy the link > paste the link in any browse > search the link > login to your Amazon account and once logged in, it will display 2 options, one is "continue previous chat" and other is "start a new chat". Click on start a new chat option, and it will connect to our team. Kindly connect with our team via order related account for our team to check and help you accordingly. -Fasi

• 
  ObeseBallerina (@Obese_Ballerina) reported 38 minutes ago

  @SimonRAnderson Problem is that when Laboor get in and introduce new taxes National either don't remove them or only half remove. Take the Amazon tax for example. Thats never going away. The "LaNdLoRd TaX" that gave us back interest deductibility on rentals has taken years to rollback. All Laboor do is add layers and layers of cost, bureaucracy and frustration to any process - bar none.

• 
  MaryWV (@MaryWVwv) reported 39 minutes ago

  @amazon @ring Amazon needs to fix this absolutely crazy huge blunder! Do you need a dictionary?

• 
  Zack Becker (@DasBecker) reported 40 minutes ago

  @JeremyMVisser This is actually horrendous. All cheap plastic jackets , the fit will be terrible and quite lacking. Using eBay will get you much further. I would strongly discourage buying Amazon clothes, like a jacket for $50 that is alpha sizing. Waste of money.

• 
  Rathod Dino (@DjDinorathod) reported 41 minutes ago

  @amazonIN @AmazonHelp @Consumer2Court keeps repeating the same story—late deliveries, false promises, and zero accountability. Customer support only apologizes but never fixes the issue. Stop misleading customers with fake delivery dates. Tracking ID: 370586504224.

• 
  ashpak (@ashpakaks) reported 42 minutes ago

  My entire business operations are completely halted. Case ID 12908859272 for Error 5461 approval is pending for 15 DAYS! Support agents are doing nothing except sending automated replies. Need immediate intervention from the leadership team. @AmazonHelp @amazonIN

• 
  Palpy (@evgenij_rabij) reported 49 minutes ago

  THIS COMPACT BOX COMPLETELY REPLACES MASSIVE ENTERPRISE AWS BILLS AND GENERATES OVER $19,500 IN NET PROFIT PER MONTH. While major corporations blindly bleed six-figure budgets every month on cloud storage like Amazon AWS or Google Cloud—paying for every single breath through API tokens. This 25-year-old engineer built a business model that completely flips the script. He designed and configured a standalone, ultra-compact micro-server that acts as a private, offline "AI Knowledge Vault," and right now, this single device is generating him over $19,500 in net profit every single month. The real magic is hidden in the expertly selected hardware and software packed into one small chassis: Massive Local Data Array: The engineer built a custom, enterprise-grade storage array with a staggering 144 terabytes (144TB of raw enterprise storage), using just six high-capacity 24TB hard drives. This allows massive datasets to be kept right at his fingertips. Ada Lovelace Graphics Muscle: Inside the micro-server slots an aftermarket NVIDIA Ada architecture GPU, dedicated entirely to local neural network processing and heavy, high-throughput media transcoding in real time. Who needs this kind of hardware in the internet age? The engineer struck absolute gold—he rents these systems out to top-tier Hollywood production boutique shops and independent media houses. Right now, the film industry is deeply paranoid about two things: data security and compute costs. When a studio works with confidential raw footage of upcoming blockbusters or high-budget commercial projects, uploading them to the cloud is a massive risk for leaks and NDA violations. On top of that, processing, automatically tagging scenes, and upscaling thousands of gigabytes of video through cloud-based AI services costs an absolute fortune in commercial token fees. This micro-server solves both pain points in one fell swoop: Absolute Privacy (On-Premise): Production shops process, index, and upscale their footage completely locally right inside their offices. Not a single byte of data ever touches the internet, bypassing cloud leaks entirely. Zero Token Cost: Instead of paying for commercial APIs from OpenAI or Anthropic, the system runs powerful open-source models deployed directly on the local hardware. The studio pays a flat monthly rental fee to the engineer, and the AI runs completely free 24/7. As a result, clients save colossal amounts of money from week one and break free from cloud monopolies, while the young engineer converts standard hardware into a stable, recurring stream of monthly passive income. Bookmark this case study—local infrastructure (Local-first AI) is the massive trend where fortunes are being built right now.

• 
  Johnny Nole (@johnnynole1520) reported 50 minutes ago

  It's the same issue with college being prohibitively expensive. The poor kid that could cure alzheimer's may end up as an Amazon delivery driver the same way that the poor kid who could be the next [insert legendary athlete] may end up never even playing the sport.

• 
  ky (@kylie_g98) reported 53 minutes ago

  I never order off Amazon because they’re literally dumb (a multi billion dollar corporation btw) but my bf always does and EVERY time they either deliver to the wrong address, forget something or deliver late lmao At least I get my walks in when they deliver it down the street

• 
  Matthew Dodd (@1doddy83) reported 55 minutes ago

  @AndreBenji00 @Raffiatim @BareLeft Jeff bezo is an American citizen who pays tax in America he does not live in the uk and therefore is not liable for personal tax I’m not sure why you’ve even mentioned him His companies Amazon uk offices will pay millions a year probably a month in employers NI and business taxes. Footballers are PAYE and are amongst the highest tax payers in the country at 45-50% income tax The top tax payers in this country pay approx 30% of the income tax bill. That’s well documented given a few % in terms of accuracy. Again my stance is everyone is paying too much. The country does not have a tax problem. It wastes far too much money on **** it doesn’t need. Start there and make better use of what they currently get. We don’t need more taxes whoever it’s aimed at

• 
  aa (@aa0130165158561) reported 55 minutes ago

  @AmazonHelp Have shared all the details over dm every day I get same response that they will personally call by evening and resolve the issue , third day today no concrete action again I have to chase customer care

• 
  Alex Kurtagic (@alexkurtagic) reported 56 minutes ago

  @ElliotKeck Gary Stevenson has, like a lot of people who think like him, a zero-sum view of the economy. It might be because he was a trader who got rich exploiting the ****-ups of others by making bets on the economy and interest rates. In his view, if you’re poor, it’s because someone else got rich. And the rich do nothing except collect rents or dividends. They don’t own businesses, they don’t employ people, they don’t generate value through problem-solving or innovative products and services. Yet he uses YouTube, X, and other social media, and likely orders from Amazon, watches films on Netflix, listens to music on Spotify, uses a smartphone, enjoys wifi, relies on memory chips, etc.—all made available to him by “the rich”. In his world, every businessman is Robert Maxwell, every landlord Nicholas van Hoogstratten, every asset manager Bernie Madoff.

• 
  Dorian Finney-Smith Muse (@DorianMuse) reported 59 minutes ago

  @note4note @amazon @ring Same thing happens to me all the time. Amazon needs to fix this major problem.