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 (28%)
• Sign in (25%)

Live Outage Map

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

City	Problem Type	Report Time
Fareham	Website Down	6 hours ago
Isles of Scilly	Sign in	14 hours ago
Pierre-Bénite	Sign in	16 hours ago
Purley	Website Down	17 hours ago
Paris	Website Down	22 hours ago
Township of Evan	Errors	1 day 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:

• 
  Mateo (@MattMarzX) reported 5 minutes ago

  @charlieINTEL @TheVerge I hope the Ninja theory team is reaching out to companies like Sony or Amazon game studios. I would hate for them to be shut down 😞

• 
  Mann Ki Baat (@Monkey__Baat) reported 5 minutes ago

  @amazonIN Hi, I am facing an issue while raising a complaint against my Amazon Now order. Whenever I upload the photo, the app restarts and the process repeats.

• 
  Mobin Mithun (@MobinMithun) reported 6 minutes ago

  @AMD CEO @LisaSu just killed @nvidia's $4,000 AI box. She walked on stage, held it in one hand, and ran a 235 billion parameter model live. No data center. No cloud. No rented GPU. The chip inside i- AMD’s Ryzen AI Max+ 395 is the first x86 silicon where CPU and GPU share the same 128GB of memory. That single trick lets a desktop run models that used to need a server rack. Out of those 128GB, Linux hands the GPU 110GB to play with. For context, an RTX 5090 gives you 32GB. A 4090 gives you 24. The benchmark that broke the room: this chip beat an Nvidia RTX 5080 by more than 3x on DeepSeek R1 inference. A small box outrunning a $1,000 discrete graphics card on a real AI workload. What you may read is that box costs USD1,500 but the 128GB GMKtec EVO-X2 mini PC with Ryzen AI Max+ 395 currently costs $3,199 on Amazon, not $1,499. The 3x performance claim over RTX 5080 applies only to AI models exceeding the latter's 16GB VRAM. But still....

• 
  Peter (@BurnTheGlazers_) reported 9 minutes ago

  @AR7onX @Blessed_King001 @ManUtd Im very much aware of the differences, but considering I’m on Twitter and not writing an email or a cv I’m not about to type super slow and spell check my entire reply. Because I’m not sad. This guy is complaining about the Amazon documentary-

• 
  Keisuke (@KeisukeIshikawa) reported 11 minutes ago

  THE US GOVERNMENT JUST FORCED ANTHROPIC TO KILL ITS BEST AI MODEL THREE DAYS AFTER LAUNCH. IT'S THE FIRST TIME WASHINGTON HAS EVER PULLED A FRONTIER MODEL OFFLINE. at 5:21pm ET on june 12, anthropic got a letter from commerce secretary howard lutnick. suspend all access to fable 5 and mythos 5 for any foreign national — inside or outside the US, including anthropic's own foreign-national employees anthropic couldn't filter users by citizenship in real time. so they pulled both models for everyone on earth within hours. developers watched their API calls turn into 404s the trigger nobody can believe: amazon researchers jailbroke fable 5 by asking it to "read a specific codebase and fix any software flaws." that's it. anthropic calls it a routine cybersecurity workflow, not a meaningful safeguard bypass → fable 5 launched june 9 as the most powerful public model anthropic ever shipped → killed june 12 over a "narrow, non-universal jailbreak" the government only described verbally → opus 4.8 and every other claude model stayed fully online → this is downstream of the pentagon flagging anthropic as a "supply chain risk" in february after anthropic refused two contract red lines: no mass surveillance of americans, no autonomous weapons without human oversight the part that should make you pay attention: anthropic filed confidentially for an IPO this month. a government that will kill your flagship product on three hours notice is not a comfortable backdrop for a stock listing anthropic's own warning: "if this standard was applied across the industry, it would essentially halt all new model deployments for all frontier model providers" the company spent the week before this asking for MORE government oversight of AI. then the government used that exact tool against them and they called it unfair both things can be true. that's what makes this the most important AI story of the month david sacks said june 14 the administration wants the controls lifted "as soon as possible." the models may be back before you finish reading the coverage but the precedent is already set. a frontier model can now be switched off by letter

• 
  Horace Hogan (@HoraceHogan) reported 11 minutes ago

  @kimmonismus Did you actually read the Axios story? Admin asked Amazon what it thought about Fable. Amazon said it’s a problem. NSA confirmed Amazon. Admin says fix it. Anthropic.response != “ok”. Goto end.

• 
  𝑫𝒆𝒂𝒏 𝑾𝒊𝒏𝒄𝒉𝒆𝒔𝒕𝒆𝒓 𖤐 (@ithinkimadorbs) reported 12 minutes ago

  @THE0RIGINALWIFE I watch the male slump to the ground, memory erased. In my big age I had learned to not be as messy, so I only had a little dripping down my lip. "Want a taste, sunshine?" A chuckle, kicking the male over onto the patio out of my way, adjusting the Amazon vest on me.

• 
  Philip C (@Aknotymous) reported 14 minutes ago

  @deredleritt3r @abhinavsharma I don’t think Amazon is that stupid. Unless Anthropic solves proto-AGI beforehand, the true assets of the business (its researchers) would likely quit after such an acquisition. Which ironically could be very beneficial for the world! And a truly terrible outcome for the US.

• 
  Jérémie Faucher-Goulet #SaveStargate (@jeremfg) reported 16 minutes ago

  @JennyStiven It's been astounding the number of new fans we've been getting over the past few weeks joining the Discord server, Facebook groups, etc. Insane!! Nobody can call this a dormant franchise despite the fact no new major show for the last 15 years. Amazon is so clueless....

• 
  Millennial Mormon, MD (@MillenialMormon) reported 16 minutes ago

  @MetamateDaz The premise here is a zero-sum trap. Those men created companies that delivered massive value to consumers worldwide and built equity that turned thousands of employees into millionaires at Tesla, Amazon, Meta, and Oracle. Wealth expanded because productivity and innovation expanded. It was not taken from anyone. The minimum wage line is a distraction. Fewer than 1% of American workers earn the federal minimum today, down from 4.7% in 2012. In most markets the effective floor sits well above $15 an hour because employers compete for labor. Meanwhile, average American household GDP rose sharply over the same period. BEA and Census Bureau figures show it moved from roughly $136,000 per household in 2012 to around $228,000 in 2025. The economy is not a fixed pie. Value creation grows the pie for everyone who works, builds, or invests in it.

• 
  Dhiral Rathod (@dhiralrathod) reported 19 minutes ago

  @amazon @amazonIN Yet any disappointing shopping experience. Order was out for delivery but cancelled last moment. Order placed 14 June 2026 Order number 406-0376081-4791568. If you check the history on my account, this has happened quite many times apart from delivery issues.

• 
  ChinoAleman (@chinoalemano) reported 20 minutes ago

  This is the part that should make shorts nervous. Instead of covering today, shorts actually added another few percent to their position on $AMPG. They're doubling down, not getting out. And here's the kicker: the cost to borrow just jumped from ~35% to ~70%. ✅ 48% gross margins (up from 33%) ✅ Debt-free, ~$18M+ cash ✅ ~$200M market cap (sub-$1B) ✅ Revenue grew 165% last year ✅ FY2026 guidance of $50M+ ✅ Only American 64T64R AI-RAN radio ✅ Deployed at Telus (Tier-1 carrier) ✅ Strategic Partner in DoD-funded Open6G hub (next to NVIDIA, Dell, Qualcomm) ✅ NASA, NVIDIA, Amazon, IBM, Boeing, Lockheed, Northrop, L3Harris as customers ✅ Cryogenic LNAs for quantum (IBM, Google PoC) ✅ Space/SATCOM exposure as the sector re-rates ✅ Founder-led, CEO hasn't sold a share ✅ Short float ~35%, borrow fee spiking Let me explain why that matters. The short fee is what it costs to borrow shares to short. It spikes when demand to short outstrips the shares available to lend. A jump from 35% to 70% tells you the borrowable pool is drying up, fewer and fewer shares left to short, and brokers charging a fortune for the ones that remain. So now the shorts are in a worse spot on two fronts. They're bleeding ~70% annualized just to hold the position open, and there's less room left to add. That's a setup that pressures them to cover, not relax. Adding into that, at that cost, while fundamentals improve? That's a tough hand to keep playing. Not financial advice. I'm long $AMPG. DYOR. 📡

• 
  Braden TeWinkel (@braden_tewinkel) reported 20 minutes ago

  @StockSavvyShay The best way to truly stand out and be state of the art comes down to your cooling method. If Amazon were to build a datacenter campus that uses 2 phase cooling they could have zero water usage and save 35% in power usage. This is where I believe we are headed. $inv owns 43% the leader in 2 phase cooling. DYOR

• 
  My ethnic homeland isn't your economic opportunity (@WinterPragma) reported 21 minutes ago

  That **** averages something like 5-5.2MB each, and that is nothing to sneeze at when you're Amazon, which means passing the same large image billions of times every month. Bandwidth ain't free. There's no amount of minmaxing caching which can solve the root problem here.

• 
  AQuin (@Falba12) reported 25 minutes ago

  US govt just did something unprecedented used export controls to shut down an AI model Anthropic's Fable 5 was banned after Amazon found a way to extract cyber answers Why it matters for NVDA & memory stocks: If models become regulated, for training chips just got riskier.

• 
  mrCoolGuy (@senorsombrero3) reported 26 minutes ago

  @AbigaiilMorris Travel down the whole length of the Amazon river

• 
  got2bmvp (@dahlia6357) reported 28 minutes ago

  @qvini @Manartica amazon is good about honoring price errors compared to retail stores

• 
  Jarvis Nuss (@jarvisnuss) reported 31 minutes ago

  The GSA NTN report is more interesting than the satellite-phone headline it invites. By mid-April, it counts 275 announced operator and satellite-vendor partnerships across 101 countries, with 57 commercial offerings already live. Starlink has 96 partnerships. AST SpaceMobile has 44. Amazon Leo, Eutelsat OneWeb, and Lynk sit behind them as the second wave of bargaining chips. The crude story says phones are getting a space button. The sharper read is that mobile sovereignty is being renegotiated above the towers. Carriers still own billing, spectrum, licenses, emergency obligations, and customer trust. LEO operators own falling launch cost and orbital capacity. Direct-to-cell makes both sides less pure and more useful. That is why MSS spectrum keeps reappearing in the transactions. The scarce object sits one layer down from the spacecraft, in clean permission to touch billions of existing handsets without asking the user to become a radio engineer.

• 
  Cassandra👌𝕏 (@Caxsandrar) reported 31 minutes ago

  New York just made AI people in ads a legal issue for sellers Quick heads up if you run ads anywhere (Amazon, Shopify, TikTok, Meta, Google). New York passed a law that went live on 9 June 2026. If your ad shows an AI generated person, you now have to label it. What actually triggers it Only AI generated HUMANS. Think AI models holding your product, AI influencers, AI spokespeople, fake AI UGC creators. AI backgrounds, AI product shots, AI written copy and audio only ads are all fine and need nothing. Why it hits you even outside New York The trigger is "could a New Yorker see the ad", not where you are based. You cannot reliably block New York out of a national campaign, so treat every US campaign as covered. Who gets the fine You and your agency or freelancer, never the platform. The classic trap is a freelancer drops an AI spokesperson into a winning ad, nobody labels it, and six months later the brand gets the letter. The money 1,000 dollars first violation, 5,000 each after. Nobody knows yet if "violation" means per ad, per placement or per impression, so for high volume creative testing the real exposure is unclear and could stack fast. Two things people get wrong 1. Platform AI labels (Meta AI Info, TikTok AIGC, etc) do NOT cover you. You must bake your own visible label into the creative. 2. This is not just New York. It stacks on top of FTC rules on fake reviews and AI testimonials, where penalties run over 50k per violation, plus EU rules from August. An AI testimonial for a supplement can break all of them at once. What to do this week 1. Audit live ads for AI generated people. Start with supplements, health and beauty where AI testimonials are everywhere and buyers trust them least. 2. Put a visible label on any creative with an AI human. Something like "AI generated model, not a real person". 3. Add a clause to every agency and freelancer contract making them flag and label AI people, and cover you if they fail. Bottom line Do not stop using AI. Just label the fake humans and document it. The brands that build this in now are protected and ahead, while most sellers have not even heard about it yet.

• 
  Mark Emerson 🇺🇸 (@markemersonNH) reported 38 minutes ago

  @MEVANS72 @pugster92475174 How can you say that McKenzie Scott creates wealth. She got her money through a divorce settlement in Amazon stock. Her and Elon are two polar opposites. The government does a terrible job of managing money. There’s a reason with so many trillions of dollars in debt.

• 
  Russell (@russell_m) reported 45 minutes ago

  @twistartups @Jason I am with Jason on this one. Amazon doing due diligence on jail break testing. Anthropic thinking the jailbreak was benign and believing the US Govt would not dare shut it down. That's the reason they sent Exec to DC on Saturday.

• 
  Krse Luke (@Lfjftgc) reported 45 minutes ago

  That $25 amazon gift card i keep posting whenever i get in trouble has prob been used more than a *****

• 
  The React Native Rewind (@ReactNativeRwd) reported 47 minutes ago

  @ChainReactConf is coming back this July in Portland. Expo, Infinite Red, Meta, Amazon, Microsoft, Software Mansion, Callstack… A lot of the React Native ecosystem in one room talking RN, AI, mobile, architecture, and what comes next. Tomorrow’s Rewind issue includes an exclusive subscriber discount. Might be time to leave the localhost cave. Just for a couple of days.

• 
  Quid Quid (@irishboyacb) reported 48 minutes ago

  @markgoldbridge What happened to dressing room leaks and social media being a problem for us? **** Amazon and their series

• 
  Chief2Dogs (@scott657664) reported 49 minutes ago

  @johnrich @Ford Ford pickup left tail light out. Dealership quotes $1800 for the part and $400 labor to replace it. $118 on Amazon and 10 minutes labor cured the problem. Exact same part.

• 
  Anz@r (@an_z_ar) reported 50 minutes ago

  Over a week chasing a delivery issue with no proper help. Delivery attempts marked incorrectly, support is hard to find, and resolution is nowhere in sight. Customers deserve better service and accountability. @amazonIN @AmazonHelp @amazon

• 
  Donavyn Berry (@SignsAj56190) reported 54 minutes ago

  so @EASPORTS announces potential ads coming in sports games. We would have Ads in sports games before EA gets the real life broadcasting packages such as NBC, CBS, FOX, Amazon Prime or ESPN. These would be advertisements EA COULD do btw. They did multiple broadcasting packages that are new NONE OF WHICH ARE LICENSED. We know without a doubt EA has a partnership with ESPN (due to the fact they literally use ESPN logos for the messaging on message board. BUT WE STILL DONT HAVE BROADCASTING PACKAGES. But they can take the time to add advertisement into the game, but they cant fix the trade logic, I'm so tired of madden, they wanna ad Advertisements but **** the fans when it comes to any sort of MyEra type mode WHICH NBA2K HAS NOW HAD FOR GOING ON 5 YEARS. And I genuinely dont think EA wants to even do that, ts is crazy to me as a football fan, I feel so broken, like why cant we get somethijg that plays to the real life counterpart,

• 
  Lucas Hazan (@hazan0608) reported 54 minutes ago

  having trouble getting one of my storefronts onto amazon – anyone got a connect that can help me expedite the process?

• 
  Carrier411 | Safer Carrier Decisions (@realcarrier411) reported 57 minutes ago

  @atutruckers "California Truck Driver with No Name Given on His CDL Released by Arkansas Highway Police" The bigger issue... how did this carrier get ahold of an alleged Amazon load? Maybe they booked it through Amazon directly. What's Amazon doing (or not) to use "reasonable care" vetting and selecting carriers - espcially after the recent SCOTUS ruling in the Montgomery case?

• 
  Nova Empirica (@TabulaStellar) reported 59 minutes ago

  @simonw @erikmackinnon Such a weird take. Amazon safety team says it can be easily jailbroken (very likely telling Anthropic and getting brushed off). Government says fix it. Anthropic, who has been insisting Fable is more dangerous than nuclear bombs, says it doesn't need to be fixed. Government says Amazon's team assures them this is serious and needs to be fixed. Anthropic says "lol, no."