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
Bohain-en-Vermandois	Errors	2 days ago
Paris	Sign in	3 days ago
Owosso	Website Down	4 days ago
Washington	Website Down	4 days ago
Paris	Website Down	4 days ago
Reynosa	Website Down	5 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:

• 
  Schulz Duggan (@Schulz_Research) reported 3 minutes ago

  CoreWeave's adjusted EBITDA margin last quarter was 56%. Its adjusted operating margin was 1%. Both numbers are real. Tonight tells you which one the market decides to believe. CoreWeave reports Q2 after the close. This is not a cloud earnings report. It is the first stress test of the exact financing model Nvidia just announced at $500 billion scale two days ago. Here is what actually matters. 1/ The 55 points that disappear between two "adjusted" numbers. Q1 2026: revenue $2.1B, up 112% YoY. Adjusted EBITDA $1.2B — a 56% margin. Adjusted operating income: $21 million. A 1% margin. The gap is depreciation, roughly $1.15B in the quarter. For most businesses, D&A is a non-cash accounting artifact you can reasonably add back. For a GPU fleet, it is the single most economically real cost on the page. The asset has a hard service life, the replacement cycle is set by Nvidia's own 1–2 year architecture cadence, and every generation makes the last one relatively less rentable. Adding back depreciation on a GPU business is adding back the business. Consensus for Q2 is roughly $2.5B of revenue, +109% YoY, on a loss of about $1.21–1.24 per share. Watch which margin management leads with. 2/ Interest is now a quarter of revenue. Q1 interest expense: $536 million, against $2.1B of revenue. Long-term debt: $22.7B, closing the quarter near $25B all-in. Then this week, two days after Nvidia's $500B announcement, CoreWeave closed a $2.6 billion GPU-backed loan. That is the thesis made literal. Nvidia said its chips are an "investable asset." CoreWeave just borrowed $2.6B against them. This is no longer a framework — it is executed paper. Q1 operating cash flow was actually positive at $3.0B. Capex was $6.8–7.7B. 2026 capex guidance is $31–35B against a revenue run-rate near $10B. Capex is running at roughly three times revenue. The gap is debt. Minsky's taxonomy (Stabilizing an Unstable Economy, ~pp. 220–228) sorts borrowers three ways: hedge units cover principal and interest from operations; speculative units cover interest but must roll principal; Ponzi units cover neither. On these numbers CoreWeave is squarely a speculative unit. That is not an accusation — it is the standard financing structure for building infrastructure. Minsky's actual point is that the category is not fixed. It moves with the price and willingness of outside capital, and nothing inside the business has to change for a company to slide from one box to the next. Which is why the refinancing calendar matters more than the backlog: tens of billions in predominantly GPU-collateralized loans across CoreWeave, Nebius, Lambda, Crusoe and Applied Digital come due between 2026 and 2028. 3/ Nvidia is the supplier, the shareholder, and the customer. Nvidia sells CoreWeave the GPUs. Nvidia invested $2B in CoreWeave in Q1. Nvidia signed a $6.3B take-or-pay capacity backstop running through April 2032. Read that third one carefully. Nvidia has contractually agreed to buy unsold capacity from a customer it also supplies and part-owns. In game-theory terms that is a commitment device (Dixit & Nalebuff, The Art of Strategy, ~pp. 174, 198). It makes CoreWeave financeable — a lender underwriting GPUs is really underwriting Nvidia's backstop. It works. It also means the demand signal and the supply signal in this chain are no longer independent variables. If you want to know what Nvidia's $500B platform looks like in practice, it looks like this deal, replicated. CoreWeave is the show home. Tonight you get to walk through it. 4/ And here is the part almost nobody frames correctly. Why do neoclouds exist at all, when Microsoft and Amazon have infinitely more capital? Because a hyperscaler building new capacity has to permit, construct and interconnect — a 3–5 year cycle. A neocloud that already holds powered sites only has to install GPUs — 6–18 months. The neocloud's moat was never the GPUs. Anyone with capital can buy GPUs; Nvidia is now actively helping them do it. The moat is holding energized capacity while everyone else waits in an interconnection queue. Which means the correct metric for this entire sector is not backlog, not GPU count, not even revenue. It is active megawatts. And here is the number that should frame tonight: CoreWeave and Nebius have each contracted roughly 3.5 GW of power capacity — and the vast majority of it is not yet energized. Contracted power is a promise. Energized power is a business. The distance between those two is the entire investment case, and it is measured in years, not dollars. $99.4 billion of backlog cannot be delivered without electrons. Nvidia understood this precisely — its investment in Texas power developer Lancium is structured so that an additional $1B tranche is contingent on milestones that explicitly include grid hookups. The most sophisticated buyer of this thesis wrote the bottleneck directly into its payment terms. 5/ So what to actually watch tonight. In order of what moves the stock: — Energized megawatts, not contracted. If management gives contracted capacity without an energization schedule, that is an answer. — Depreciation and adjusted operating margin. Does the 1% go up or down. — Capex guidance for 2026. $31–35B stands. Any raise is a bigger funding gap, not a bigger business. — Terms on the $2.6B GPU-backed loan — rate, tenor, advance rate against hardware. That tells you what the market actually thinks a used GPU is worth. — Customer concentration. Microsoft was 62% of 2024 revenue. Ten clients now committed at $1B+. Progress here is genuinely de-risking. — Founder selling. Jefferies flagged it as the next catalyst, not earnings. Three scenarios: Bull — revenue beats, energization schedule given with dates, adjusted operating margin expands, capex held. The financing model looks self-funding and every neocloud reprices up. Base — revenue in line, backlog up again, margins flat, power discussed in contracted terms only. The stock trades on capex guidance alone. Bear — capex raised without matching energization, operating margin compresses further, or any softness in the 2026–2028 refinancing commentary. Then the market stops paying for the backlog and starts pricing the interest expense. 6/ Where I could be wrong. The circular-financing critique is now well-worn — io-fund, Jefferies and others have all written it. Being right about a risk everyone has already named does not make money. Positive $3.0B operating cash flow is a real defense, and I do not want to wave it away. This business generates cash; the question is only whether it generates it faster than the fleet depreciates and the debt comes due. And the $99.4B backlog is contracted revenue from creditworthy counterparties, not vapor. If energization lands on schedule, almost everything above resolves benignly and the bears look silly. The bet tonight is not on whether AI demand is real. It is on whether the electrons show up on time. Not investment advice. Q1 figures from CoreWeave's Q1 2026 release and 8-K; consensus and capex guidance as published; power capacity figures per industry reporting.

• 
  Dan Mullaney (@agenticgrowth) reported 9 minutes ago

  @alex_alexiuc The Amazon story is the more interesting data point in this list, because it's not a headcount argument, it's a monitoring argument. $1.8M and 860% over budget on a menial coding task doesn't happen because AI is expensive, it happens because nobody had a real-time view of spend against expectation until the invoice showed up. That's a solvable problem with the same rigor companies already apply to cloud spend, but most AI usage still gets treated like a corporate card with no limit instead of metered infrastructure. The headcount numbers say AI isn't replacing people yet. The Amazon number says the bigger near-term risk might just be nobody watching the meter.

• 
  Ayush Garg (@ayush_garg1001) reported 10 minutes ago

  Amazon sde 1 intern batch 2027 assessment given right now: 1)a 2 pointer min cost based array question to make all elements the same 2)a nodejs based codebase where creation and dleetion of a product were bugged.Had to fix the api routes and schema to fix it 3) continued below...

• 
  Perc Cousins (@PercCousins2) reported 15 minutes ago

  @SamAntar Mamdami is an idiot, but Amazon, etc using subs or independent contractors to avoid liability is a major issue

• 
  Donald Ferguson (@HMRC4EVR) reported 15 minutes ago

  @dilophosaurid The Cardbot figures I've got off Amazon haven't had issues, but slower delivery. Re-check ebay, filtering 'US only' to see options. Fgures like Heavy Iron W (Blastrain to lesser degree) were higher on Amazon. Heavy W was $90 more than eBay listing that had import fees included

• 
  Benny | The Bench Trades (@TheBenchTrades) reported 15 minutes ago

  THE READ: an index down 0.19% is hiding a 2%-plus sell in Alphabet and Amazon. Rotation, not de-risking — IWM green, HYG flat, VIX under 16. Kills it: SPY back over 774.54, or GOOGL green by the close. Proof, not hype. @TheBenchTrades. Not financial advice. Educational only.

• 
  Gradient (@GradientX0) reported 17 minutes ago

  2. Unitree G1 — the one you can actually order Unitree took the opposite bet: don’t wait for the robot to be perfect, ship something affordable and let volume and iteration do the rest. What’s actually happening: as of February 2026, the G1 is listed for sale on Amazon US at $17,990, with a lower-spec base configuration starting around $13,500–$16,000 depending on source. Unitree shipped 5,500+ units in 2025 alone — more than every other humanoid robot maker combined, according to multiple industry trackers (though AgiBot disputes this with its own 5,168-unit figure and Omdia’s #1 ranking). Unitree’s 2026 target is 10,000–20,000 units shipped. Hardware: roughly 19–23 degrees of freedom depending on configuration, and a notably smaller frame than the other two — about 130 cm (4’3”) tall, versus ~168–173 cm for Optimus and Figure. Battery life sits around 2 hours in standard operation, well below Figure’s 20+ hours or Optimus’s 8-hour shift. The honest problem: the G1 is genuinely capable for the price, but it’s not competing on raw industrial capability — it’s competing on accessibility. Serious secondary development requires the separate, quote-only EDU version rather than the consumer-facing retail unit. Short battery life also limits it to research, education, and light commercial roles rather than continuous industrial shifts. Where it’s used today: education, research labs, and light commercial pilots — it’s the platform startups and universities buy because it’s the only one they actually can.

• 
  Joshi (@JoshidaTrap) reported 21 minutes ago

  @AmazonHelp the link sent does not fix the issue, and the only email i got was telling me iv been locked out and the helpline number 1800 064 122 does not work.

• 
  MarxMan (@MarxistsWon) reported 21 minutes ago

  I doubt we are quite there yet with current technology, but not too far away. Modern women, especially American women, are in DEEP trouble when men can order something like this off Amazon.

• 
  snipy.in (@snipy_in) reported 23 minutes ago

  JUST IN: US market ends down as Iran peace deal optimism fades Wall Street closed lower as investors grew pessimistic about Middle East stability. Amazon and Alphabet dipped, weighing on the S&P 500 and Nasdaq. Brent crude futures held near one-week highs amid rising oil prices and uncertainty. Investors await crucial inflation data that will shape Federal Reserve policy expectations.

• 
  wanneracademy (@wanneracademy) reported 25 minutes ago

  $32,000 last month. From YouTube and Spotify combined. But here's the thing, relying on one platform is a mistake. I broke down exactly how I turned my faceless channel into multiple income streams: memberships, Spotify, Amazon, Dailymotion, Tubi TV, digital guides, and more. This is how you stop relying on a single platform and actually scale. Comment "SNOOZE" and I'll send you the full 20-minute breakdown.

• 
  Ryzm (@Goeun_6121) reported 26 minutes ago

  US Close | August 11, 2026 The S&P 500 fell 0.33% to 7,727.79 and the Nasdaq Composite lost 0.60% to 26,445.00. The Nasdaq 100 declined 0.33% to 29,525.48. SOX gained 0.87% to 12,098.47, recovering part of Monday’s selloff. The 10-year Treasury yield eased to 4.68%. WTI settled 1.3% higher at $83.20. Reuters, Nasdaq The index fell. Semiconductor beta and small caps moved the other way. What changed? Brent traded above $90 before settling at $88.91 after Iran said the Strait of Hormuz would remain closed unless the US accepted its conditions. WTI added another $1.07 after Monday’s 5% jump. Reuters The rates response was more restrained. The 10-year yield finished near 4.68%, down from 4.72% late Monday, and the two-year was near 4.23%. The curve stayed roughly 45 basis points positive. Oil kept the inflation problem alive, while Treasuries faded Monday’s selloff before CPI. That distinction matters. The S&P’s decline was concentrated in megacaps and lacked confirmation across risk assets. Amazon fell more than 2%, Alphabet lost almost 4%, and SpaceX dropped 5.2%. The Dow fell 0.35% to 53,787.57. The Russell 2000 gained about 0.3% to 3,027.12, DXY was nearly unchanged around 99.82 and VIX stayed near 15.3. Why did chips hold? SOX recovered 0.87% after Monday’s 2.94% drop. Nvidia finished near flat, and the group held despite another rise in crude. The 10-year yield’s retreat removed the second leg of Monday’s pressure. The funding story also shifted. Apollo, Blackstone and KKR rose after Nvidia announced partnerships with six financial firms to mobilize more than $500 billion for AI compute infrastructure over time. Nvidia describes independent pools of third-party capital. The market rewarded the asset managers that would package the financing, while the largest cloud buyers lagged. Nvidia AI spending remains in place. More of its risk is moving toward credit structures, collateral values and completion schedules. The financing terms and deployment pace remain undisclosed, so today’s price action shows enthusiasm for the fee pool before the capital has been committed. What reaches Korea? SOX back above 12,000 gives Samsung Electronics, SK hynix and Korean equipment names a better opening signal than Monday’s selloff did. The cleaner test is still foreign cash flow into the two memory megacaps. US chips recovered while Amazon, Alphabet, Dell and Oracle weakened, so a broad AI-beta reading would be too generous. Oil is the second channel. WTI at $83.20 and Brent near $89 keep pressure on Korea’s import bill and the won. If USD/KRW rises while the 10-year remains near 4.68%, the oil channel will matter more than the US duration signal. Tomorrow July CPI is due at 8:30 a.m. ET. The consensus looks for headline inflation to slow to 3.4% year over year from 3.5%. The release will test whether Treasuries were right to fade Tuesday’s oil rise. US Bureau of Labor Statistics Watch the 10-year around 4.70%, SOX at 12,000 and WTI above $83. A hot print that pushes yields through 4.70% would reconnect oil with the discount rate. A softer print with SOX holding 12,000 would leave Tuesday looking like megacap concentration risk inside a resilient AI trade.

• 
  drugranker (@rankdrugs) reported 29 minutes ago

  @PlumbNick The problem for Amazon started with Bezos. His culture of frugality encouraged regular PIP's and high turnover between 2007-17. Many NW locals who got burnt in their first stint never went back despite lucrative 2'nd stint offers. The company is now forced to hire foreigners

• 
  K.T. Carlisle (@KT_Carlisle) reported 35 minutes ago

  Alright, folks. It's about time for a rant. I'm pissed as hell, and if you've ever considered running a Featured Deal through @BookBub, you might want to stick around, read about my experience, and think twice before throwing your money away like I did. What they have done to me is negligent at best, predatory at worst, and I'm so disgusted by their behavior, complete lack of support, and contentment to steal from the authors they claim to assist. Back in June, I decided to apply for a BookBub Featured Deal. I'd heard mixed things about the platform; some authors raved about how their Featured Deals helped them sell hundreds of copies, while others lamented the lack of movement in that department. As someone who hasn't done that much advertising for my recently published books, I thought it was worth checking out. Without getting into too much behind-the-scenes detail, I keep running into the same obstacle when it comes to acquisitions interest: I haven't hit that magic 3,000 lifetime sales for a single title yet to be considered worth the risk, despite the fact that I've managed to sell about half that without any advertising whatsoever. But I digress. I was excited when I got approved for my Featured Deal. I thought it was a sign of good things to come, and even though I was nervous about the price tag and the scary "No Refunds" messaging attached to my purchase, I figured, "Maybe this'll help me reach that magic number." I got approval, purchased the spot, and waited for August to arrive so I could set my promotion price through Amazon. BookBub makes it pretty clear that failure to set the promo price in time will result in your Featured Deal not being run, so I made sure to log into my KDP dashboard every single day leading up to the deal to set that promo price. My deal was set to run on August 10, but by the time August 7 arrived, it became pretty clear to me that my worst fears were coming true. The KDP Enrollment period hadn't reset when it indicated it would at the time I made my purchase back in June, which meant I couldn't set the promo price on my book by the launch date for my BookBub Featured Deal. For those of you who may be unfamiliar with how KDP Enrollment periods work, essentially, it's a 90-day period that Amazon sets, which dictates how often you can set promo prices. So, for instance, if you set your book on sale for $0.99 from August 1 – 8, you won't be able to do so again until ~October 31 or thereabouts. The infuriating thing about this is that these enrollment period dates fluctuate across titles, and there's not a set start/end date, so there's a little bit of guessing involved. That's where I got into trouble with my Featured Deal. I thought I was well within the reset margin so I could set the promo, but I was off by literally one day, and that one day cost me $202 that—as BookBub loves to remind me—I am unable to get back. Once I realized the mistake on August 7, I immediately reached out to BookBub support. They did not respond. I reached out to them again three more times between August 8 – 9. No answer. On August 10, I reached out again after receiving the email that my Featured Deal was not run. No response. I finally checked their "No Refund" policy, and within it found that I could reapply for a Featured Deal and have the date moved back, so I thought, "Okay, I can deal with that. It's not what I planned, but at least I'm not completely losing the money and I can just push back the spot." Only after submitting for a new Featured Deal and explaining the situation did I receive a response. At first, it seemed like they were willing to work with me, asking me about the new dates I'd like to select and if that worked for me. I approved everything and was relieved (albeit slightly annoyed about the new timeline) that everything was going to work out after all. Then, BookBub hit me with, "Okay, that'll be another $202." Excuse me? You want me to pay you an additional $202, so I'm now paying double the price for ONE ad? That makes absolutely no sense to me. When I pushed back on this, I was told that this was BookBub's policy because I "took a spot from one of their other partners." That doesn't compute to me. I paid for a service, the service was not provided due to circumstances entirely outside of my control (as is often the nature of the publishing industry and most people who work within it are flexible enough to understand that this is the case; **** happens, deadlines change, you work with the affected parties in a reasonable manner, and come to an agreeable solution—basic customer service/human decency regardless of any industry you work in as far as I'm concerned). BookBub hasn't gotten this memo, it appears. Instead of simply changing the date of my deal and applying the money I already paid, they're demanding more. So, my only source of recourse at this stage is to report the charge to my credit card and attempt to recover the funds that way (which, again, were not used to run any ads whatsoever). I fully intend to do this, but since the charge itself isn't fraudulent, I doubt this will go through, and I'll be out $202 that, honestly, I could barely afford to pay when I purchased the spot back in June but I was so desperate for a win that I did it anyway. Never. Again. Moral of the story? This industry will convince you that you aren't good enough. It will tempt you into pouring hundreds of dollars you don't have into the hope that you'll finally be desirable. And when that's not enough either, it will punish you for trying to get ahead. Don't let that happen to you. There are other ways to get your books out there without greedy entities like this blatantly stealing from you. I wish I never let comparison rob me of my joy and my money, but here we are. Let my experience be a lesson to you so you don't fall into this same trap.

• 
  Kris Hardy (@KHsportsfan) reported 36 minutes ago

  Idc if people live there lol. Take their land one more time for old times sake lol. Idc who’s impacted I want this and I don’t feel bad lol. Migrate to the Amazon it’s just down the road I’m sure lol. South America isn’t that big lol.

• 
  DARANATH KUMAR (@KumarDaranath) reported 38 minutes ago

  @AmazonHelp ECR has ALSO been delaying and making empty promises for the last 5 DAYS without scheduling the pickup! Stop passing the buck between teams. If Amazon cannot send a courier agent and issue a FREE PICKUP DATE immediately, save your explanations for the Consumer Court tomorrow!

• 
  David Caraway (@CarawayDJ) reported 40 minutes ago

  @RoccoRidesTesla Google and HERE have extensive fleets mapping the roads daily. So does TomTom, but to a lesser extent. Tesla's navigation problems are a $$$ problem. They pulled a Delta/Amazon LEO.

• 
  Lagbhag Raaz 😷 (@i_Raaaz) reported 43 minutes ago

  Hi @amazonIN , my return was already picked up by your courier partner, but the order status is still showing as "Return not picked up." Because of this, my refund is being delayed. Please look into the issue and update the return status at the earliest. #Amazon #CustomerSupport

• 
  ⊹.˚✧⋆❀ Chun-Li ♡⋆ ✫⊹｡☪︎ (@CosmicKarma__) reported 44 minutes ago

  you ARE NOT employed or insured by Amazon. it is a small-town wanna-be rich person who buys the trucks and hires the drivers as a subcontractor. and don't you DARE point ANY of this out if you wanna keep your job. you better drive that broken truck, starve for lunch, & be quiet.

• 
  Rebecca (@Rebeccaou8v) reported 45 minutes ago

  @AmazonHelp @amazon Shocking experience. Been lied to be 3 different people. Had a chat disconnected before he even let me explain. He shut my feelings down, I have 2emails stating refund processing. Another additional email stating refund processed successfully.

• 
  Scott Bramley (@ScottBramley1) reported 48 minutes ago

  @andyburnham Betting shops aren’t killing our high streets — they’re there because people use them. Amazon, online shopping and retail parks changed our towns. Stop wasting taxpayers’ money trying to engineer high streets and focus on real issues like illegal immigration.

• 
  Paul Harri (@PaulHar10765609) reported 51 minutes ago

  This man has bought his trainers from my shop but has the nerve to call small business dodgy. Won’t deal with the actual problem of Amazon who pay very little tax whilst destroying small business. @andyburnham

• 
  Michaelsssss 🇺🇸 (@Michaelsssss41) reported 53 minutes ago

  @NYCMayor No problem Amazon will just use UPS, FedEx and USPS and pass savings zone to consumer All the subcontractors would be fired

• 
  Kothandaraman S 🇮🇳 🕉️ (@saijayam) reported 54 minutes ago

  @amazonIN @amazon India Team - For a significant price difference in a product (-18% within 10 days), why can't you simply issue a credit for the price difference. Now a return is initiated and a new order is placed!

• 
  Arihant (@_Goyal_Arihant) reported 55 minutes ago

  My orders are repeatedly being cancelled automatically as “suspicious.” Order #406-9113547-0285136 (₹294) was also cancelled after successful payment. Please review my account and resolve this issue so I can place orders normally. Screenshots attached. @AmazonHelp

• 
  MargaB (@MargaBTweet) reported 60 minutes ago

  @brij_chaggar @PeterStefanovi2 And you can't have Amazon and healthy high streets. Seems people have chosen. Trouble is, they want it all.

• 
  Saundra Edgell (@saundra_edgell) reported 1 hour ago

  @TMZ @HarveyLevinTMZ I have 409 more videos, and I need help. They were recorded at Amazon with Meta glasses over the course of six months. They vary in length from 20 seconds to FIVE MINUTES!!!!! The aspect ratios differ and are not optimized for YouTube. All I know how to do is voice-overs in Adobe. Still, they would be more entertaining to edit with vid recordings of me explaining what's happening, as I'm one of the funniest people on the planet with seven years in that warehouse dealing with rogue AI. The ratios are causing me problems in OBS studios. Help.

• 
  fbs (@fleetblackship) reported 1 hour ago

  @AlexNoonan6 I feel that I also deserve some restitution for handling last meter delivery. I've carried thousands of boxes up the stairs and then broken them down. They should give me an amazon vest.

• 
  mushy_vocado (@KingJoffrey2020) reported 1 hour ago

  @maggiemoda Theres 2 types of Amazon drivers. There Flex, which is what you are talking about. Individuals who use their own cars. But drivers with Amazon branded vans, work for a subcontracted third party. This is the big issue.

• 
  Clever Alias (@CT0347635474175) reported 1 hour ago

  @Jason Sure, Jason... I'm sure the "Studies" you read proved your silly point. Why would anyone imagine businesses are affected adversely when the cost of one of their inputs rises artificially thanks for myopic do-gooders like you? I appreciate the confidence with which you present "the data." Let me offer some of the data you apparently didn't review. "Prices only rose 1-2%." Congratulations. You just described a pay raise funded by the customers, many of whom earn minimum wage themselves and now pay more for the food they can barely afford. You've taxed the poor to pay the poor and called it progress. "No businesses shuttered." You checked some of the surviving businesses. Did you count the ones that never opened or the ones that never grew? The restaurant that was almost viable at $12/hour labor but not at $20? That business doesn't appear in your data, Jason, because it never existed. You can't measure the job that was never created, the location that was never leased, the supplier that was never contracted. Economists call this the "unseen." Bastiat wrote about it in 1850. It apparently hasn't made it into your research yet. "It didn't drive massive automation." You're posting this in 2026. Self-order kiosks are now standard in every major fast food chain. Automated drink dispensers. Robot fry cooks being piloted by multiple chains. Amazon eliminated cashiers entirely with "Just Walk Out" technology. But sure, no automation. Were you not paying attention, or do you think those kiosks were installed because the companies just love touchscreens? "Individuals earning under $40,000 spend 100% of their income." Yes. That's called "having no savings." You're describing financial fragility and calling it "monetary velocity" as though rebranding poverty as an economic engine is clever. These people aren't spending because they're stimulating the economy. They're spending because they have no choice, and the ill-informed like you continue to drive up the cost of living with economically destructive but oh-so-virtuous interventions in the marketplace. And my favorite part: your last paragraph. You admit, in your own words, that minimum wage increases "depress" jobs for teenagers, who "aren't worth $15-20 an hour." So your solution is a government-mandated two-tier labor system where bureaucrats decide which workers deserve the "real" wage and which ones don't. You just described the exact central planning problem that free market advocates have been warning about, and you did it while arguing against free and voluntary market pricing. You dis-proved your own point and called it "good business." Maybe read those studies again...