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 (49%)
• Errors (27%)
• Sign in (24%)

Live Outage Map

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

City	Problem Type	Report Time
Annecy	Website Down	3 hours ago
Frankfurt am Main	Website Down	23 hours ago
Bridgeport	Sign in	1 day ago
Rochester	Errors	2 days ago
Saint-Apollinaire	Errors	2 days ago
Noisy-le-Sec	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:

• 
  Cindy (@CindyBalesUSA) reported 29 seconds ago

  @rigo48520173 I’m still trying to get my packages delivered to my 14’x36’ covered porch that also happens to have zero steps between it and the driveway. Instead, most of my Amazon packages are delivered by the USPS, who would rather leave them in the rain, day after day after day. Honest to God, 1 stride would fix the problem, but that’s too much to ask. Rain or shine they also stick my mail underneath the box, sticking out about a quarter way. We call 2 pm at my house WTF time. Everyday. Can’t wait to see what they do for an encore.

• 
  Amazon Help (@AmazonHelp) reported 1 minute ago

  @ramOG250925 @amazonIN @ramOG250925 Kindly copy the link > paste it on a 'web browser' > login to your Amazon account and share your details. Our team will investigate get back to you within 6-12 hours via email. -Sravan

• 
  Smokey Bear (Charles M.) (@Smokey_Bear2024) reported 4 minutes ago

  How nice. Did you all know that SpaceX sold $25 billion in bonds after selling a bunch of you suckers $8[ billion I debt through the IPO? SpaceX is flat broke and busted hemorrhaging cash like blood spilling from stuck pig. That's not the only issue, apparently Amazon is

• 
  Anna writes (@anna_writes0) reported 8 minutes ago

  Looking for the prompt he used??? There you go Replace [YOUR NICHE] with any niche you're interested in. Prompt 👇 Act as an experienced Amazon KDP market researcher and publishing strategist. Help me find profitable, evergreen non-fiction book topics in the [YOUR NICHE] niche with strong search demand on Amazon. Focus on topics that solve real problems and can be expanded into a series of related books. For each topic: 1. Suggest a high-converting, SEO-friendly title. 2. Suggest logical spin-off books that naturally connect with the main topic. 3. Prioritize beginner-friendly subjects with long-term demand. 4. Avoid highly competitive niches that are difficult for new publishers to enter. 5. Focus on practical guides that people actively search for and are willing to buy. 6. Recommend topics that can build a recognizable author brand and allow for cross-selling between books. Think like a successful KDP publisher and prioritize books that provide clear value, solve everyday problems, and have the potential to generate consistent long-term sales.

• 
  🌟 LIN 🌺💫 (@wispiicoatl) reported 10 minutes ago

  imo spindle is in the same boat as glitch, they aren’t rlly indie anymore given how connected they are to Amazon I’m not here to Spread Leaks for fun, and based on former employees’ statements I think 2 spoiler-free screencaps is not the biggest threat to artists over there

• 
  not the Brad you hoped for (@the_Brad_youget) reported 15 minutes ago

  @DrDeath1776 There’s already a solution to the hardtack problem though. Pilot bread is a durable cracker, slightly denser than saltines, about 1/4 thick, that can be eaten straight. They’re actually delicious. If you could find a way to make these without the seed oils, that’s the thing to stock up on. The 2lb box used to be $5 delivered by Walmart in like 2019 but it’s now $20+ if you can find it. They sell it in 9oz cans on Amazon for $25/lb, more than a buck a cracker, labelled 30 year shelf life. What kind of equipment and ingredients do you have access to? I’ve thought of making these to stack in widemouth jars. Cans are the better option but I don’t yet have a sealer.

• 
  Black.Silq (@BlackSilk529) reported 15 minutes ago

  I'm trying to shed atleast 30 pounds... I still like being a fluffy bunny but I want to bring it down a bit...still want to keep the curves tho...maybe add with the weights and protein... I'm aiming to be a true Amazon 💀😭

• 
  Sgjs (@Sgjs32408) reported 17 minutes ago

  Does anyone have a problem with the customer service with Amazon? They speak a version of English that is hard to understand. I give them a 1 out of 10 for customer service speaking an understandable English. The English they speak is sure not American English.

• 
  69kov (@levikov) reported 19 minutes ago

  Instagram will ban u for showing a ****** then promote a woman who has never existed to 400,000 followers then let Samsung pay her $14,000 for a single post then recommend her content to ur mother 6 times a day then watch ur mother buy everything she recommends then take 0% of the commission same platform. same algorithm. same terms of service that say "authentic content only." ur ****** was a violation. the AI character that makes $50,000 a month selling vitamins to ur mom was a "recommended creator" n the reason this works is the single most uncomfortable truth in the entire creator economy: the algorithm does not measure authenticity. the algorithm measures watch time, saves, shares, n click thru rate. a fake grandmother who gives genuinely useful health advice scores higher on every single one of those metrics than a real influencer who films 3 hours a day in a ring light the algorithm cannot tell. the audience cannot tell. the brand safety tools cannot tell. the only entity on earth that knows the grandmother isn't real is the operator running her from a laptop in his apartment n nobody is coming to fix this. there is no regulation. the FTC wrote its endorsement guidelines for REAL PEOPLE. the legal framework hasn't caught up to fictional characters with parasocial relationships n brand deals. by the time it does, the operators who started now will have already made millions here's the build. every step. nothing held back face: google flow. $99/mo. warm woman, early 60s, kitchen background. lock the face. lock the room. lock the outfit. NEVER change it. the sameness is how ur mom's brain files her as someone she knows voice: real older woman cloned from fiverr. $300 one time. 60 min of varied audio. elevenlabs $99/mo. she sounds real bc the voice IS a real grandmother. ur mom can't tell bc there's nothing off. the warmth is genuine. it's just wearing a different face content: 6 videos/day. 4 pure education + 2 product. 18-second scripts. "if u're over 50 n ur knees crack every time u stand up, this is for u." the algorithm pushes the educational ones for free. when a product video crosses 2-3% click thru, tiktok spends the BRAND'S ad budget to push it further. u pay $0. tiktok pays DM funnel: "comment SLEEP." auto-DM. amazon idea list w 12 products. 31-41% click thru. one video = 400,000 views = 2,000 comments = $3,976 in commissions 8:47pm exploit: post when she empties her cart. 24-hour amazon cookie covers everything. $14 product becomes a $14.20 commission bc of the $340 already in her cart. same video at 2pm = $0.56 6 characters across 6 niches: $180-300K/mo. $1,278/mo in tools. 99% margin Instagram banned a real woman for a ****** n promoted a fake woman to 400,000 followers. the platform's own algorithm chose the fake one. the platform's own brand safety tools approved her. the platform's own recommendation engine showed her to ur mom 6 times a day n ur mom thanked her in the comments every morning at 6am ap3x. 1:1 coaching directly into the operation. not a course. we built the system $70M+ runs on. fill out the form. applying costs u 2 minutes n nothing else. we hand-pick n reject most. link in bio the real woman got banned for a ******. the fake woman got promoted to 400,000 followers. Instagram is begging u to build the fake one n u're still out here filming urself in a ring light for 51 followers lmfaooo

• 
  Ram (@Ensbsjjsikzbw) reported 21 minutes ago

  @AmazonHelp @amazonIN you said the issue has been escalated even after informing you guys the order got cancelled the second time ?

• 
  All that glitters is not gold 😎 (@MissSG__) reported 22 minutes ago

  This isn't fair. I always go out my way to return packages back to the original Amazon delivery station if I have issues delivering them. However, I learned that I should stop marking packages as "unable to deliver" because I usualy end up being able to deliver it anyway.

• 
  V_Game_Veteran (@D_G_Add) reported 22 minutes ago

  An Amazon Web Services failure caused the PSN outage not a hack by anonymous.

• 
  Nyan Nyan (@CatNyanpital) reported 28 minutes ago

  @firstadopter Amazon is so bogged down by bureaucracy they needed a ERD and 6 meetings about the letter before voting on whether to even read the letter

• 
  Chris Weston (@ChrisWeston_PS) reported 32 minutes ago

  📊A Traders' Week Ahead Overview... The top-down macro forces will continue to dominate cross-asset moves this coming week... Incoming US-Iran headlines and the price action in crude, the broader energy markets and the ag's will impact rates, USTs and by extension the USD and gold..at this point, a 4-5% gap lower in crude futures tomorrow seems fair.. Secondly, there remains big focus on the moves in the US/Korean/Chinese AI plays, and whether we see further drawdown. Our NAS100 cash index closed Friday below the mid-June swing low, with the sellers in control and few seemingly prepared to buy this dip with any conviction - given the weaker open tomorrow eyed for crude, unless we see new news, the NAS100 should reopen +0.5% to +1%.. The Fed meeting is the main event risk for markets this week, with USD IR swaps implying a 38% chance of a 25bp hike - the Fed wont hike at this meeting, which given the pricing in IR swaps, suggests some risk of initial USD weakness.. Trading over the FOMC meeting is for the brave, or for those with an ultra fast news reactive algo... US core PCE inflation could move the dial, although the estimates from economists are usually on the money... Tomorrow's CXMT debut on the Star exchange, is also pretty funky - pre-IPO perp pricing implies a near 400% gain for the stock on open.. we look to see if the stock can pull further investor capital from other equity holdings, and keep kicking ahead of SK Hynix and Samsung's earnings.. On the corp earnings side, 32% of the S&P500 market cap are set to report numbers, with the key focus on Meta, Amazon, and Microsoft and whether they join Alphabet in lifting their 2026 capex plans.. For the EU equity/index ballers, 21% of the EU Stoxx market cap hit us with numbers this coming week... Aussie June monthly/Q2 CPI will be closely watched by the RBA and also market players - with AUD IR swaps currently pricing a 32% chance of a 25bp hike at the August RBA meeting.. for the AUD momentum traders, AUDCHF is the play, having broken to new trend highs, with AUDJPY and AUDCAD not far off.. While a continuation of the recent move is somewhat conditional of Aus Q2 CPI and EU CPI, and I am not one for trading over key news, short posi's in EURAUD look there for the taking too..

• 
  Metro (@LewisL442006) reported 32 minutes ago

  @luceishere_ I use the Turtle beach stealth 600 Gen 3 wireless headset, works for Ps5 Xbox and Pc I’ve not had any issues with them so far I’ve had them a few weeks now and I like them 🔥 they are being sold at Argos and Amazon

• 
  pavan kumar (@rajapavanprata) reported 33 minutes ago

  @AmazonHelp Give me a callback and sort out the issue ..I will share contact details through dm

• 
  Yugandhar Tripathi (@Yugandhar_19) reported 34 minutes ago

  @Nishant2908 @TheLongInvest I bought MSTR all the way from 95 to 350 and from 350 to here just doubled down extra hard below 120 xD I have Amazon RSUs from job side so don't buy that but very bullish on AMZN (maybe bias xD NFA DYOR). I buy monthly individually picked US stocks that are undervalued that I share anyway on posts time to time like recently CRCL COIN HOOD etc being mindful of my crypto space exposure so META and GOOG and NVDA too. Time preference for any pick is minimum 5 years here so I don't worry much about buying companies that are oversold on RSI on day or week level which I am confident are not going out of business.

• 
  Alex Napier Holland 🦍 (@NapierHolland) reported 35 minutes ago

  @BladeoftheS No, the UK has high tax because of a bloated welfare state, low productivity and insufficient entrepreneurial activity. Your sole achievement in life is to write one of the most badly-reviewed books on Amazon. Sit down, shut up and listen to people who actually understand economics. Talentless narcissist.

• 
  OldWomble 🏴󠁧󠁢󠁥󠁮󠁧󠁿🇮🇱🇬🇧 (@OldWomble) reported 35 minutes ago

  @AmazonHelp Here’s what I’m getting, Switch countries to resolve issue To keep shopping, you need to fix an issue in one of your other Amazon accounts. I DONT HAVE ANY OTHER ACCOUNTS

• 
  Ryan Mitchell (@TheRealMitch) reported 35 minutes ago

  @mitulkanji I'm cool with automation, since most humans prove themselves to be utter morons. From Mcdonalds workers not being able to figure out how to put a cheeseburger and chicken nuggets in the bag when the bag says, "cheeseburger and chicken nuggets" on the side, to the Amazon workers who will somehow **** up scanning an item and placing it into a box, as if it's rocket science...I'll gladly welcome automation if it fixes these problems.

• 
  David Welker (@welkerlaw) reported 40 minutes ago

  @BobbyBorkIII This sounds like a problem for Costco, not Amazon, to solve.

• 
  Suryansh Tiwari (@Suryanshti777) reported 43 minutes ago

  Jensen Huang runs the most valuable chip company on Earth. He also just admitted, on record, that he let one of the biggest AI bets of the decade slip through his fingers. It wasn't a lack of belief in Anthropic. It was an outdated assumption. Huang's old playbook said: if a startup needs capital, it goes to VCs — that's just how the game works. He didn't grasp that Anthropic's capital needs were an entirely different order of magnitude, the kind no venture fund could realistically underwrite. By the time that clicked for him, Google and Amazon had already moved in and written the checks. Here's the part that actually matters: Huang didn't dodge it. He said it plainly — his old model of "how funding works" broke, and he was slow to notice. That's the real skill in this story. Not predicting the future perfectly. Noticing the moment your assumptions stop matching reality — and admitting it before someone points it out for you.

• 
  Alonzo Harris (@HashemMelech048) reported 48 minutes ago

  @JasonMBrodsky Let me ask you something since you believe, against any reasonable evidence to do so, an anonymous source NY Times article- Do you think the stockpile in this scenario was unknown 2 weeks ago? Like some clerical error? lol Do you think they began this round of fighting without a full accounting of resources? Do you think someone who understand manufacturing was like but urs fine well Amazon prime sone more? All in all, I’m at the point of questioning your competency in analyzing the situation.

• 
  Asymmetric Investor (@gu_yunhong43226) reported 53 minutes ago

  $MU $SPCX $SKHY $META $NVDA $AMZN $MSFT $CRWV Distributor or Gas Station? — The Structure of Today's AI Market This is a time when you have to invest knowing that no decision comes with a right answer. The answer key can change at any moment. Still, there is something an investor can do: separate what people are underestimating from what they are overestimating, build an investment idea on that distinction, and decide how to hold the position. That is the hard part of investing. Knowledge helps with judgment, but what the market rewards is not knowledge itself. It is reading which direction people's beliefs are migrating. The decision at today's price level — does it break down from here, or does it bounce — is the same kind of question. Both sides have logically sound arguments. So let me start by laying out the bear case. 1. Why did it fall? — Leverage I believe the biggest driver of the decline was leveraged money. Once prices slipped, liquidations and margin calls followed, and the selling fed on itself. But the trigger was something else: anxiety over memory prices, and the fact that this cycle's investment has shifted from self-funded to leveraged. A data center built with debt is a fundamentally different object from one built with your own cash. With your own cash, a demand miss ends as a depreciation problem. With debt, it becomes a credit problem. That is exactly where the market has grown sensitive. 2. Drawing the Map — The Stakeholders In the business of building AI data centers today, there are broadly four groups of stakeholders: Big Tech (Google, Amazon, Microsoft) Compute lessors (SpaceX, Meta) Neoclouds (CoreWeave, IREN, Nebius, Applied Digital) Stargate (Oracle–SoftBank–OpenAI) On the model side, three companies are competing, with the open-source camp standing outside them: Anthropic (Claude) OpenAI (ChatGPT) Alphabet (Gemini) Overlay this map onto the oil industry and it becomes easier to understand. Memory is crude oil. Nvidia is the company that designs refineries and sells the equipment. Big Tech clouds are the major refiners; neoclouds are the small refiners. And the model companies are patent holders collecting royalties on their blending formulas. 3. Distributor or Gas Station? This picture reveals the essence of the battle underway. The companies that control compute face one of two fates. If open source grows and many models compete, compute providers become distributors who control the sales channel for intelligence. Like Walmart, they decide which model sells at what price, and they keep the margin. But if a handful of models capture the market, compute providers become gas stations — following whatever price the refiner sets, collecting a thin margin. That is why compute providers cannot help but cheer for open source. Renting out compute to run open-source models leaves them more margin, and the more models there are, the more pricing power shifts toward the distribution layer. 4. Everyone's Ledger Nvidia's calculation goes like this. It needs refiners of every size, small to large, so that its customer count stays high and their competition preserves Nvidia's pricing power. So Nvidia wants compute providers to earn fat margins. But right now, a large share of that margin is being captured by Anthropic and OpenAI — the companies that opened up the market in the first place. With that market-opening role largely complete, the picture that favors Nvidia is one where model-company margins shrink — in other words, where open source grows. If model margins compress, AI prices fall, demand rises, and compute providers capture that margin, thickening Nvidia's customer base. Big Tech has its own ledger. Burning enormous capex, they do not want pricing power ceded to the model companies. So they, too, back open source. But they also dislike Nvidia continually breeding new rival refiners. So they have weaponized their scale — partnering with Broadcom to design their own ASICs and entering a war of attrition, using lower inference costs to squeeze out competitors. The model companies' calculation is the exact opposite. Flush with capital, they do not want open source to grow. The bigger open source gets, the weaker their pricing power becomes. To sum up: on the open-source front, Nvidia and Big Tech are on the same side, with the model companies across the line. But on the chip-and-memory front, Nvidia and Big Tech collide head-on. Big Tech must secure HBM and TSMC capacity to make its ASICs efficient; Nvidia must secure even more HBM to defend its moat. This is why no one can stop buying, no matter how expensive memory gets. The red-ocean fight to thin out the number of refiners still lies ahead, and the fewer suppliers remain, the thicker the pricing power and the moat — so the game is to lock up memory first. 5. And All of It Sits on Debt The problem is that this entire fight is being waged on top of loans collateralized by future revenue. Every participant faces a supply shortage. As price and performance rise, compute demand explodes — build all you want, it is still not enough. So everyone is pulling forward debt against nothing more than future revenue contracts. OpenAI's and Anthropic's compute contracts flow through as RPO (remaining performance obligations) at SpaceX, Amazon, Google, and Microsoft, while Oracle is concentrated on OpenAI alone. The neoclouds, in turn, borrowed against these same contracts on a priority basis. Private credit and insurance money went into those loans, and Nvidia's own capital went in as well — stacking up an enormous pile of credit. Because loans only get made once chips are secured first (via LTAs), no one can tell where real demand ends and phantom demand begins in today's chip scarcity. Yet on the back of these massive contracts and orders, the memory makers have kicked off large-scale capacity expansion. If this demand turns out to be phantom — if data centers tip into oversupply, or if actual AI demand cannot support the capex — the conditions are in place for that mountain of credit to collapse into a financial crisis. I also suspect Big Tech's aggressive bond-market raises have a calculation beyond securing compute: by preempting and pressuring the credit markets, they cap their competitors' ability to raise money. Economies of scale mean that as rivals disappear, the payback on their hardware investment grows stronger, backed by pricing power. 6. So Is It a Bubble? This is a structure with a very high probability of producing a bubble — enormous borrowing stacked on top of demand that is hard to forecast. But declaring an "inevitable collapse" is a different matter. If the demand is real — if AI scaling keeps generating demand — the same structure could lead to even stronger expansion and cash generation, and the market could go much bigger. A bubble's preconditions being in place is not the same as a bubble being confirmed. What we can read at this point is what investors are afraid of. And personally, I find the opposite possibility hard to dismiss: that when Rubin launches and inference prices fall, a genuine cost-down phase opens up — and AI crosses into a structure that actually makes money. 7. My Conclusion Back to the opening question: what remains is separating the underestimated from the overestimated. What I think the market underestimates is this: until Rubin arrives, compute can only be scarce. In a supply-shortage regime, bubbles are hard to pop. So at this point, I lean toward betting on the train headed into the bubble. Conversely, the things the market fears — the limits of financing, credit anxiety, the macro environment — are not groundless. The market's current pickiness is understandable. But I see it as a question of when those fears materialize, not which direction is right. We can compare this cycle to history. But this event, too, is independent. The rhyme may be similar; the ending can differ. And if you are invested, there is one thing to keep in mind to the end: none of us knows anything. So build positions not on conviction but on distinction — on what is underestimated and what is overestimated — and stay ready to change when the answer key changes. That, I think, is what investing means in a time like this.

• 
  netgate123 (@netgate123) reported 53 minutes ago

  @samhogan Another issue worth raising is that Amazon is Anthropic’s largest investor, so its decisions also reflect Amazon’s interests. Don’t just hide in the shadows and try to rake in profits, Amazon!

• 
  Douglas Porter (@ahhyeah) reported 54 minutes ago

  @Swiftor I used it to automate my receipt reporting and now that’s broken. I’ll have to manually go to Amazon and screenshot them and upload them adding hours to my workflow.

• 
  netgate123 (@netgate123) reported 55 minutes ago

  @Mononofu @JensenHuang Another issue worth raising is that Amazon is Anthropic’s largest investor, so its decisions also reflect Amazon’s interests. Don’t just hide in the shadows and try to rake in profits, Amazon!

• 
  Pbzzzz2323 (@pbzzzz23233) reported 56 minutes ago

  @AGRobBonta @jaketapper @TheLeadCNN Free Market takes care of PRICES BiddyBrain. Not GOVT. People evaluate whether the price is too high, and if it is, they don’t buy it. Enough people shy away, the price comes down. ITS CALLED SUPPLY AND DEMAND, and it’s Econ 101. AMAZON and NETFLIX must have donated a bundle.

• 
  Victor John Ekereobong (@d_ekereobong) reported 56 minutes ago

  @lokoso418 Thanks for this, but I have this question, can I publish my Paperback on Amazon + Expanded Distribution ON, AND same book as eBook on Google Play, and other platform, without having issues?

• 
  Avinash Verma (@VermaSgit) reported 58 minutes ago

  @AmazonHelp I have made an order for vacum cleaner and when received parts are missing. Complaint raised but no response since more than 15 days. Customer Care is making fool since then. Order summary Order placed 7 July 2026 Order number 171-8779647-6007550 Resolve issue.