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
Morlaix	Website Down	5 hours ago
Mumbai	Errors	19 hours ago
Paris	Website Down	22 hours ago
Paris	Website Down	1 day ago
Iztapalapa	Errors	1 day ago
Charlotte	Website Down	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:

• 
  Rich Peter (@peterli34923561) reported 8 minutes ago

  $AMZN --- Early July: Amazon kicked off a $25B+ bond offering to fund AI chip procurement and data center buildout. Make no mistake — this is Amazon doubling down in the AI arms race, and it's all about cementing AWS's cloud dominance. Despite the massive raise, Fitch still slapped an AA- rating on it. That tells you everything about how the market views Amazon's cash generation machine and competitive moat. Earnings watch: Q2 report drops July 30. The two numbers everyone's fixated on — AWS revenue growth (Q1 printed 28%) and how much AI capex is compressing near-term margins. 1. AWS: The AI growth engine. This is where the magic happens. Generative AI integration isn't just landing big fish like OpenAI and Anthropic — custom silicon (Trainium, Graviton) is widening the cost-performance gap. And with a backlog north of $364 BILLION, the long-term visibility here is about as solid as it gets in tech. 2. Ads: the high-margin secret weapon. Amazon's ad business has become a profit monster. Quarterly revenue keeps climbing, and thanks to the retail platform's built-in traffic, the incremental margins are insane — directly juicing the bottom line. 3.Retail is finally hitting its efficiency stride. Regionalized logistics and third-party seller integration have the retail business holding share while profitability keeps stepping up. The old "low-margin retailer" narrative is officially dead.

• 
  divyansh darbar (@theIndian0786) reported 14 minutes ago

  Dear @AmazonHelp, I purchased Amazon Prime on 4 July for 999, but due to your technical issue, it's still not activated, and no refund either. Now the price is 1499. Please honor my original purchase and activate at 999. #AmazonPrime #CustomerService #AmazonIndia

• 
  Ranjeet mete (@RanjeetKum98365) reported 17 minutes ago

  @AmazonHelp Prime membership = Faster delivery? Reality: Order 2 books, get 1 book, wait 1 week for fix. Online shopping ka fayda hi kya hua? Pay more than market to waste time. Please do better.

• 
  GUL (@gulVasikova) reported 17 minutes ago

  Why T1 Energy Fell: Investors Are Questioning the Next Phase of the AI Boom T1 Energy (NYSE: $TE ) dropped nearly 16% on Tuesday, but the selloff wasn’t caused by company-specific news. Instead, investors started questioning whether the massive AI infrastructure spending boom can continue at the same pace. The first trigger came from Samsung Electronics’ preliminary second-quarter results. Although Samsung reported strong revenue and profit driven by AI demand, investors focused on whether AI spending may start slowing after two years of explosive growth. Markets always look ahead. Even strong earnings can lead to selling if investors believe future growth may begin to cool. Another major concern came from a Reuters report that Chinese AI startup DeepSeek has quietly spent the past year developing its own AI chip. This matters because many large AI companies are beginning to design their own processors instead of relying entirely on suppliers like Nvidia, Broadcom, Samsung, and Huawei. Companies including Google, Amazon, Microsoft, Meta, and Tesla have already taken this approach. Building custom chips helps reduce costs, improve performance, and gives AI companies more control over their infrastructure. For investors, however, it also raises an important question. If more AI companies build their own chips, will demand for third-party semiconductor suppliers grow as fast as previously expected? That doesn’t mean chip demand will disappear. AI infrastructure spending is still expected to grow, but investors are beginning to think growth may become more selective instead of lifting every company across the supply chain. Why T1 Energy Was Sold Too T1 Energy doesn’t make AI chips, but it is part of the broader AI infrastructure story. AI data centers require enormous amounts of electricity, making power generation and energy infrastructure critical to the AI buildout. As companies like Microsoft, Amazon, Google, Meta, Oracle, and CoreWeave expand their AI capacity, they also need more energy. However, when investors become cautious about AI spending, they often sell the entire ecosystem—not just chipmakers. That includes data center companies, networking firms, power providers, and energy infrastructure names like T1 Energy. So Tuesday’s decline was driven more by sector-wide sentiment than anything specific to the company. Technical Selling Added More Pressure The selloff also triggered technical selling. T1 Energy fell below its 20-day and 50-day moving averages, while the MACD remains bearish, showing sellers still have control in the short term. Even so, the stock remains well above its 200-day moving average, meaning the longer-term trend hasn’t been broken. The Bigger Picture The market isn’t saying AI is over. Instead, investors are becoming more selective. Rather than assuming every AI-related company will keep soaring, they’re starting to ask which companies will capture the biggest share of future AI spending. That’s why T1 Energy fell despite no negative company news. The stock was caught in a broader reset across AI infrastructure names as investors reassessed expectations after a huge rally.

• 
  Earthen Vessel (@heathobrien51) reported 19 minutes ago

  @WesleyLHuff My Amazon app is not working. I have to find another way. When someone mentioned it could be this generation's ET I've been trying to see it.

• 
  Talikithegreat🇺🇬🇬🇧 (@Talikithegreat7) reported 22 minutes ago

  @Favwontmiss Yeah, it’s legitimately ******. The Amazon isn’t just “pretty trees” — it’s one of the planet’s biggest carbon sinks. When you cut it down

• 
  truthseeker414 (@clownworld414) reported 24 minutes ago

  @LINEMUSIC_JP @haku_circle For some reason i cant purchase this on amazon. Website is broken 😭

• 
  Chris Jaynes (@TheChrisJaynes) reported 26 minutes ago

  @WallStreetApes The error in this post is that Amazon has free shipping. It doesn't. Prime members get "free" shipping. Only, they don't. They pay their shipping in advance when they pay for their Prime membership. I don't buy enough from Amazon for the free shipping to offset the membership fee

• 
  DarkStarTales (@darkstartales) reported 28 minutes ago

  @spybrary "I know editing is expensive" that, right there is the issue. I'm not saying don't charge what you feel you need to, or, that books don't need as much editing as possible, however that 'indie' tag, usually means 'no financial backing.' It means writing your story, self editing and/or using software, and then self publishing it, because there is no other way to do it. When I wrote and published my first book, there was zero way I could afford an editor for any stage of the process, and if I had waited until I had the money, I never would have published any book, and I wouldn't be sitting here with 50+ books on Amazon. If you're wealthy, and you're writing/publishing, that's a different kettle of fish, then yes, hire an editor, but unfortunately that's not going to work for the majority of indie authors.

• 
  Bharat Mata ke Sewak (@CountryGulshan) reported 29 minutes ago

  A man ordered a ₹1,200 Wi-Fi adapter on Amazon using Cash on Delivery. When the delivery agent arrived, he made an unusual request. He said, "Cancel the order, pay me directly, and you'll still get the package." The customer was surprised. He wondered, "If I cancel the order, there will be no official record of the delivery. What if something goes wrong later?" Thinking it could be risky, he refused. He said, "I'll pay only after the order is marked as 'Delivered' in the Amazon app." The delivery agent then marked the order as delivered, and only after that did the customer make the payment. Before leaving, the delivery agent allegedly said, "No problem. You'll need us someday. Then we'll see." Now the question is: If the customer had canceled the order and paid the delivery agent directly, could he have lost his ₹1,200, or was it just a misunderstanding? What would you do? If a delivery agent asked you to cancel your online order and pay them directly, would you agree, or would you wait until the order was officially marked as Delivered in the app before paying?

• 
  Fredrick Fabian (@fredicancompose) reported 29 minutes ago

  @AmazonHelp Initiate a conversation with the delivery agent. I can't even fetch details on that, I'm tired of seeing amazon letting me down. Please fetch some details, and hurry them today, I want it delivered the same day.. today july 10th 2026

• 
  Quintessential American, God,Family,1A,2A-Texas (@Gibson5972) reported 30 minutes ago

  @EricLDaugh @VP @POTUS here, I made you list of where to start. The CORE of the VISA problem is with these, and many other companies. Meta Apple Microsoft XBOX Oracle UPS FED-X Tesla Space-X Amazon AMD Salesforce Alphabet Netflix NVIDIA PayPal ALL Universities

• 
  Ranjeet mete (@RanjeetKum98365) reported 31 minutes ago

  Dear @amazonIN @AmazonHelp Prime membership = Faster delivery? Reality: Order 2 books, get 1 book, wait 1 week for fix. Online shopping ka fayda hi kya hua? Pay more than market to waste time. Please do better. Please help this metter @PiyushGoyal sir

• 
  Smiling into the void (@EXTREMEDANCING) reported 37 minutes ago

  @JamesBryhan Am I supposed to be answering if it does have appeal or it doesn't? Because the appeal should be based on execution not concept. It's no problem for amazon to release stargate. It's a problem if it's bad.

• 
  Tushar Karmarkar (@imTkarmarkar) reported 40 minutes ago

  Ecommerce is broken beyond repair this time due to rains Bhiwandi flooded .. try to order any stuff on Amazon / flipkart and for dates they're showing beyond 25 july or even 8-10 august 💀

• 
  Saurabh Gidwani (@gidwani_saurabh) reported 41 minutes ago

  Ordered a 55” TV from Amazon (Order #405-4042041-3508367). Received a defective unit, but neither @AmazonIN nor the seller, Dawn Tech Electronics Pvt. Ltd., is replacing it despite the issue being confirmed. Very disappointing experience. A SCAM. #AmazonIndia @geekyranjit

• 
  🇺🇸Edward🇺🇸 (@cauzin_trouble) reported 42 minutes ago

  @carryingmarine @amazon My ring cam will show the proof. The darkies throw **** from halfway up the sidewalk on the porch. The very very few white dudes bring to door and set down. Ups and fed ex both white drivers. Door everytime. Raining they put under table. Call they give money back.

• 
  Mahima nayyar (@mahima_nayyar) reported 43 minutes ago

  @AmazonHelp What ****** service is this Amazon? Why do I have login to the app to talk to the customer care? Is this a really different facility for that I need to make so much of efforts Why there is no easy way lot contact customer care like other brands have?

• 
  Davis (@DavisPaipa) reported 43 minutes ago

  The easiest organic sales channel for Shopify stores in 2026 is not SEO. It's AI buyers. And most stores are invisible to them for one boring reason: empty category metafields. When ChatGPT, Claude, or Perplexity recommends products, it reads structured Shopify catalog data. Category, attributes, product metafields. No structure, no recommendation. We already lived through this once. On Amazon, listings with complete product attributes outsold identical products with empty fields. Nobody argued about it, you just filled them in. Shopify is now at that exact same moment. Category and category metafields are the new listing attributes. The problem: on a 100 product store, filling these manually takes weeks. That's why almost nobody does it, and why the stores that do are winning AI recommendations by default. We built this into Shoptank. Claude assigns the category and fills the category metafields for every product in your catalog automatically. What took a content team a quarter now runs while you sleep. If you checked your store right now, what percentage of your products have their category metafields filled?

• 
  Justin Jordan (@the_defi_dad) reported 48 minutes ago

  @blknoiz06 Trash dev and trash coin. 0 chance it recovers. It’s Amazon prime. You can’t let it run down like it did and not expect recourse. Run is over. Bummer I’m stuck on a boat with no way to sell my bag. Ansem is black devil

• 
  biswajit singh (@biswajitsingh20) reported 50 minutes ago

  The AI Reality Check Has Begun. 2026 has shown that AI adoption alone isn't enough. 📉 Microsoft cut costly AI coding licenses. 📉 Uber burned through its AI budget by April. 📉 Google shut down Mariner. 📉 Klarna rehired human support agents. 📉 Walmart capped AI usage. 📉 Duolingo dropped AI-based performance metrics. 📉 Meta ended its internal AI token leaderboard after 60T tokens were consumed in 30 days. At the same time: 📊 Meta shifted 7,000 employees into AI teams. 📊 Amazon is hiring 11,000 AI interns & graduates. 📊 Deloitte plans to hire 50,000 professionals in India while upskilling 30,000 employees. The next AI race isn't about adoption. It's about ROI, governance, and building the right human + AI workforce. #AI #FutureOfWork

• 
  Jatin Pandat (@_JatinPandat) reported 52 minutes ago

  What's one app with terrible UI... ...but amazing UX? I'll start: Amazon.

• 
  Lex | High Ticket Ecom (@dropshiplex) reported 52 minutes ago

  The mistake: Google Shopping is not research until sellers are filtered and 3 are direct. Amazon, eBay and quote pages fake demand. You need 3 independent stores with live add-to-cart for the same product type. The fix: the 3-direct filter. Count stores, not noise.

• 
  Venkatesh.BS 🌞 (@AppuVBS) reported 53 minutes ago

  @amazon Pathetic service !! I’ve been calling from three days to just return the item as it is not working for me, and your service centre guys not at all helping! What’s going on.

• 
  Toast (@CryogenicToast) reported 58 minutes ago

  @jffry8902 @Byromew @TRIGGERHAPPYV1 He is literally working under Amazon you lobotomite. He used a weapon that is against policy to carry. He’s going to get in trouble/fired. You literally are goldfish brained if you think otherwise.

• 
  Stuart Recher 🦜 (@sjrecher) reported 58 minutes ago

  One of the AI paradoxes that ******* way are just terrible Netflix, Amazon and Apple are at suggesting shows I’d like to see…

• 
  Silence is complicity. Speak up. (@zdotconnector) reported 60 minutes ago

  @DrSuzanneH7 I've been wanting to do this for a long time. I feel 100x better when in low-EMF environment. My entire nervous system lets down and I sleep like a rock. I was looking at those canopies on Amazon a while back but was not sure if they worked. Thanks very much for sharing!

• 
  Susheel Nughwal (@SusheelNughwal) reported 60 minutes ago

  @AmazonHelp What do think I am free for fill your form every day and didn't get any response. Are you making us fool on daily @JeffBezos if you are not able to run business plz shut down it don't make us fool. One of the worst company I see. Now I regart why should I ordered from Amazon.

• 
  Kathryn Smith (@Whippetartist) reported 1 hour ago

  @unreMARKLEble Because remember? Movie making studios have tried talk with Harry about his making documentary about his mother. Amazon interested to make Diana documentary with Harry. But - Harry wants others do all work, he just appears, talks, & problem of M*ghan.

• 
  Guilherme Favaron (@guifav) reported 1 hour ago

  Most production agents pay a hidden tax: they regenerate the same code on every request. New arXiv (2607.08010) from @AmazonScience turns that inference-time loop into a build-time compile step. In the CodeAct paradigm, on every repeated step the agent re-reads the instruction, re-discovers the schema, and re-writes similar code — paying in latency, tokens, and run-to-run variance. The tech behind the tech: a tool-maker that compiles recurring SOP steps into validated, versioned tools before deployment, grounded in the live system — collecting traces, reading backend schemas, generating candidates, repairing them against labeled cases. At runtime the agent calls the tool directly, falling back to codegen only when needed. On a fulfillment-center alarm-triage agent (a 44-node SOP over heterogeneous backends), swapping the codegen loop for tool calls cut production p50 latency 42%, output tokens 58%, and turns 45%. Compact structured verdicts unlock a simpler direct-call design — a further 62% p50 cut in a model-fixed ablation — and drop errors up to 53% by killing variance in repeated steps. The part worth noting: the residual errors weren't a model problem. Four of 44 SOP nodes caused nearly all of them — clarifying that spec text lifted pass@1 from 94.5% to 99.9%. The library was a one-time 800K-token build, amortized over every call. Your next latency and reliability win may not be a bigger model — it's compiling what your agent already repeats. (Malmasi et al., Amazon)