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 (29%)
• Sign in (25%)

Live Outage Map

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

City	Problem Type	Report Time
Waldbröl	Sign in	13 hours ago
Victorville	Errors	20 hours ago
Louisville	Website Down	1 day ago
Bohain-en-Vermandois	Errors	3 days ago
Paris	Sign in	4 days ago
Owosso	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:

• 
  Aberaeron 🇪🇺🇺🇦🇨🇦🇩🇰🇫🇷 (@ItsFullOfFellas) reported 44 seconds ago

  @Buckpat21 @camille_moscow @grok If Biden had invaded Mexico, lost much of his elite forces in a failed raid on the capital, advanced then fallen back on several key axes, and 4.5 YEARS LATER aid from around the world was pouring into Mexico, while USA had been ground down to using Donkeys, gone through his M1-Abrams tank storage, gotten just 50 km in on a key axis made a national priority, and suffered 450,000 DEAD, with the US Navy hiding in the Gulf of Mexico, Florida cut off from effective resupply, while major US refineries were being regularly bombed by Mexican drones and Amazon was in danger as their biggest warehouses burned down -- even IF it were correctly reported as a very difficult situation for Mexico, would large parts and many newsrooms of the Global South and Che Guevarra lovers around the world be cheering the defeat of US imperialism and Biden's "total failure" to subjugate America's neighbor? Would these people call it a "Minor American humiliation", "Major American humiliation", or "Strategic American humiliation"?

• 
  Ad Noctem Media (@NoctemAd) reported 2 minutes ago

  @MrAtomicDuck @zachbussey They don't really care. Your other choices are YouTube or Kick. They're not asking or taking feedback, they're just doing. Maybe even being told from higher up that they're doing it because they need to provide value to Amazon. They just need the initial upset to die down and then it will be normalized

• 
  Ryan Matthews (@Ryanmufc90) reported 3 minutes ago

  I’m down in Cornwall was at a Beach shops and pubs was selling them for up to £10-15 signs everywhere saying sold out we got ours on Amazon for £5 fo 6 crazy some peoples made some good money today!!!

• 
  UnsentAuron 🗡️ (@UnsentAuron1) reported 9 minutes ago

  @JNavok The real problem is going to be the fact that Sony will have full control over the pricing of their games. There will no longer be competition from Amazon, Best Buy, etc. Also, saving money by buying used will no longer be a thing. The cost is that many customers will simply be priced out of the hobby.

• 
  Mine (@MineBorder10444) reported 11 minutes ago

  @shannonplante I think Dan is a bad CEO, but Emmett Shear was even worse. I think the biggest issue is Amazon

• 
  Fastbeak (@Fastbeak_) reported 12 minutes ago

  for coming out to suggest Seattle Tech Startups, Seattle scenes, or Seattle anything. If they want to walk away quietly out from the Amazon side parties, can understand, otherwise I'll take their Entrepreneurial culture and shove my Quantum Aircraft down their ******* throats.

• 
  @iamsrkian Raj (@Wasimak72436966) reported 12 minutes ago

  @JioCare @reliancejio My JioFiber issue has been unresolved despite multiple complaints and emails. Google/Amazon work, but Flipkart, Uber Driver, Myntra, X & several other services don’t work on JioFiber.Please escalate this to your technical/network team and resolve it urgently

• 
  Anthony C (@URnotAC) reported 18 minutes ago

  @NYCMayor Amazon should dump every sub contractor that has a problem or leave NY for good

• 
  The Azillion Show (@azillionnet) reported 20 minutes ago

  Advice from a few years ago told authors to email Amazon and request up to ten categories. That route closed in 2023. What you get now is three categories and seven keyword slots, set inside the publishing dashboard. The part people miss is the word "per". Your ebook, your paperback and your hardcover are separate listings. Each one gets its own three categories and its own seven keywords. A book in all three formats has nine category slots and twenty one keyword slots, and most authors fill the ebook, leave the print versions on defaults, and never look again. Categories first. Choose one broad and two specific. The broad one gives you visibility in a big room where you will rarely rank. The specific ones are where you can actually reach a top twenty position, and being visible on a small shelf beats being invisible on a large one. Find them by browsing the store the way a reader browses, following the category path down from the main page, rather than only reading the dropdown list in the setup screen. Keywords next. Each slot takes a phrase of up to fifty characters, so think in phrases rather than single words. A pattern that works: one slot for the reader's problem, one for the subgenre, one for a trope or tone, one for setting, one for a comparable audience, one for occasion, and one for a format need. For a cozy mystery that might look like "clean mystery no gore", "small town cozy animal mystery", "amateur sleuth with cat sidekick", and so on. Skip anything already in your title or your chosen categories. Those words are indexed anyway, and repeating them wastes a slot you could spend on a phrase a reader might actually type. Do this today. Open each format's setup page and count the empty slots. Fill every one, then write down what you set and the date, because later you will want to test them one slot at a time.

• 
  The Great Lock In(-60lbs/27kg) (@Hurt97919464y) reported 24 minutes ago

  He’s giving me his phone so I can go on Amazon and replace what’s broken

• 
  Felix (@FelixAmazonCat) reported 25 minutes ago

  Course sellers worldwide present retail arbitrage as simple - scan an item in store, check the price on Amazon, instantly make profit. This is far from the case. Returns, ungating, account issues, longer than expected sell through rates are rarely mentioned. But what if I told you I've just solved that problem. 100% guaranteed buys in retail arbitrage mode, no analysing graphs or messing about. If we won't 100% buy it, it will just reject the item you're scanning. Get an insant, guaranteed payout just from shopping. No dealing with e-commerce platforms. This makes retail arbitrage a million times more accessible. I could easily sign my mum up to this and she'd be able to understand this and make money. I could even hire a bunch of local crackheads and have them sourcing for me as well...

• 
  Amazon Help (@AmazonHelp) reported 28 minutes ago

  @ChGorou_ We're sorry to hear about the trouble with your delivery and that you're unable to reach us. So we can best assist you, please clarify which Amazon marketplace your account is associated with (.com, .uk etc.)? We're standing by! -Gayle

• 
  TheValueist (@TheValueist) reported 33 minutes ago

  $CSCO KEY READ-THROUGHS FROM CISCO SYSTEMS Q4 FY26 EARNINGS CALL Cisco’s Q4 FY26 earnings call provides one of the clearest cross-sector confirmations that the AI infrastructure cycle is broadening from accelerators and servers into networking, coherent optics, enterprise data centers, campus infrastructure, security, telecom transport, memory, power management, and industrial connectivity. The strongest evidence was not merely Cisco’s 18% revenue growth, but the breadth and forward visibility of demand: total product orders increased 35%, orders excluding hyperscalers increased 25%, enterprise orders increased 21%, public-sector orders increased 30%, service-provider and cloud orders increased 95%, telco orders increased more than 30%, and hyperscaler AI infrastructure orders reached $4 billion in Q4 and $9.3 billion for FY26. The call is therefore directionally positive for the physical AI infrastructure ecosystem. However, it is not uniformly positive for incumbent vendors. Cisco is using vertically integrated silicon, systems, optics, software, security, and observability to re-enter markets where it historically had limited hyperscaler exposure, creating material share risk for merchant-silicon suppliers, independent networking vendors, optical-system incumbents, and stand-alone security vendors. The call also contains an important quality-of-growth warning: approximately 5 percentage points of Q4 revenue growth and an expected 4-5 percentage points of FY27 growth are being generated by price increases, while product gross margin declined 270 bps. The most actionable conclusion is that the AI infrastructure cycle is becoming broader and more durable, but the relative winners will be companies positioned in scarce components, physical infrastructure, private AI systems, and cross-domain platforms rather than every company exposed to nominal data-center spending. AI NETWORKING, SWITCHING AND SILICON ETHERNET AI FABRIC DEMAND IS ACCELERATING, BUT CISCO’S HYPERSCALER SHARE RECOVERY CREATES A STRUCTURAL RISK FOR ARISTA (READ-THROUGH 1) AFFECTED COMPANIES AND IMPACT: Arista Networks (ANET: US) faces a near-term positive demand read-through of moderate magnitude but a longer-duration negative competitive read-through of moderate-to-high magnitude. NVIDIA (NVDA: US) receives a positive read-through from the continued expansion of AI infrastructure spending, but Cisco’s scale-out and scale-across wins create a low-magnitude negative read-through for NVIDIA’s networking portfolio specifically. The negative impact on NVIDIA’s consolidated earnings is substantially smaller than the potential impact on a networking-focused company such as Arista. CALL SUPPORT: Cisco took $4 billion of hyperscaler AI infrastructure orders in Q4 and $9.3 billion for FY26, approximately 4.5 times the FY25 total. All 4 of Cisco’s largest hyperscaler customers increased AI infrastructure orders at triple-digit rates. Approximately 60% of the orders were for Silicon One-based systems and 40% were for optics. Cisco won 3 new hyperscaler designs during Q4, including a P200 scale-across deployment and a G200 scale-out deployment. Cisco now has 3 P200 scale-across wins, each with a different hyperscaler, and received orders from all 3 during Q4. Management expects FY27 hyperscaler AI revenue of $7.5 billion, compared with approximately $4 billion in FY26, implying approximately 88% growth. Management also stated that FY27 AI orders should be “meaningfully higher” than the $9.3 billion recorded in FY26. TRANSMISSION MECHANISM: The sector-level demand read-through is positive because Cisco’s results validate rapid growth in Ethernet-based AI fabrics across both scale-out and scale-across architectures. This supports the overall addressable market for high-speed switching, routing, optics, network operating systems, and fabric-management software. Arista should benefit from the same hyperscaler capital-spending environment, particularly where customers prefer Ethernet-based architectures. The non-consensus implication is that Cisco’s success is not only a market-expansion signal; it is also evidence of a credible share-recapture campaign. Cisco emphasized that it had virtually no business inside hyperscaler data centers slightly more than 6 years ago. It now has multibillion-dollar orders, design wins across several separate hyperscalers, and a product stack spanning silicon, systems, optics, security, and software. Cisco’s 60% systems mix indicates that it is not participating only as an optical-component provider. It is competing directly for high-value switching and routing deployments. Cisco’s profitability framework increases the competitive risk. Management stated that the incremental expenses required to capture hyperscaler revenue are minimal and that lower-gross-margin hyperscaler systems can still be highly profitable because Cisco does not need to add a traditional enterprise selling organization in proportion to revenue. Cisco can therefore pursue market share aggressively without necessarily sacrificing operating-margin growth. This gives Cisco greater flexibility on pricing, product bundling, and commercial terms than a competitor whose earnings model depends more heavily on maintaining very high product gross margins. NEAR-TERM TRADING CATALYST: Continued hyperscaler capital-spending growth should support networking revenue estimates across the sector during the next several quarters. Arista can still report strong demand even while Cisco gains share because the underlying market is expanding rapidly. The principal near-term risk is that new Cisco design wins, particularly across multiple hyperscalers rather than a single customer, cause investors to lower assumptions for Arista’s long-term hyperscaler share or terminal margin structure. Cisco indicated that additional G300, G200, P200, and optical decisions are expected during the next 6 months, making design-win disclosures a potentially significant trading catalyst. LONGER-DURATION FUNDAMENTAL SHIFT: The longer-term issue is the emergence of Cisco as a vertically integrated hyperscaler networking competitor rather than a legacy enterprise-networking vendor. If Cisco can scale Silicon One across hyperscaler, enterprise, telco, and campus environments, the value of a common architecture, supply chain, and software layer could become increasingly important. Arista remains a major beneficiary of AI networking growth, but Cisco’s re-entry reduces the probability that incremental Ethernet AI spending accrues to Arista with limited competitive pressure. The call therefore supports higher sector revenue assumptions but also a higher competitive-risk discount for Arista. Cisco did not identify the 4 hyperscalers by name. No company-specific customer attribution to Amazon, Microsoft, Alphabet, Meta, or another cloud operator is warranted solely from the transcript. CISCO’S SILICON ONE EXPANSION IS A STRUCTURAL NEGATIVE FOR MERCHANT-SILICON CONTENT AND A POSITIVE FOR TSMC (READ-THROUGH 2) AFFECTED COMPANIES AND IMPACT: Marvell Technology (MRVL: US) faces a negative longer-duration impact of moderate magnitude. Broadcom (AVGO: US) faces a negative longer-duration impact of low-to-moderate magnitude at the consolidated-company level because of its greater diversification. Taiwan Semiconductor Manufacturing Co. (TSM: Taiwan) receives a positive longer-duration impact of low-to-moderate magnitude. CALL SUPPORT: Cisco plans to deploy Silicon One “comprehensively across our high-performance networking systems by fiscal year ’29.” Management stated that owning the silicon, systems, and software provides greater control over the supply chain and innovation roadmap. Cisco also emphasized that it deals directly with TSMC rather than depending on allocations and commitments from merchant-silicon providers. Management explicitly identified the potential “avoidance of stacking margins with merchant silicon” as a longer-term economic benefit. TRANSMISSION MECHANISM: Broadcom and Marvell participate in merchant Ethernet switching, routing, connectivity, custom silicon, and related semiconductor markets. Cisco’s deployment of internally controlled Silicon One products creates a direct risk to merchant-silicon sockets that Cisco might otherwise source externally. The impact occurs through reduced unit content, lower attach rates, diminished pricing leverage, and lower incremental revenue from Cisco’s future high-performance system volumes. The effect is likely more material for Marvell than for Broadcom on a relative basis because Broadcom has a broader earnings base spanning custom accelerators, networking, storage connectivity, wireless, and infrastructure software. Marvell also has important custom-silicon and optical opportunities that can offset networking-silicon displacement, but Cisco’s stated intention to internalize more of the stack remains a clear competitive negative. The positive read-through for TSMC is direct. Cisco’s internal silicon strategy does not eliminate semiconductor outsourcing; it shifts value from merchant-chip suppliers toward the foundry. Cisco indicated that it is managing wafer, substrate, assembly, and test requirements directly and maintaining an active relationship with TSMC. As Silicon One expands across Cisco’s product portfolio, TSMC should benefit from higher networking-ASIC wafer demand and greater visibility into Cisco’s product roadmap. NEAR-TERM TRADING CATALYST: The near-term earnings impact on Broadcom and Marvell should be limited because the portfolio-wide rollout extends through FY29 and the underlying AI-networking market is expanding. Market growth can offset lost Cisco content during the early phase. The key near-term catalysts will be Cisco disclosures regarding which product families have transitioned to Silicon One, the rate of system adoption, and any evidence of lower merchant-silicon purchases. LONGER-DURATION FUNDAMENTAL SHIFT: The more important implication is architectural. Cisco is attempting to convert networking silicon from a purchased component into a strategic control point. If successful, this can improve product differentiation, supply assurance, security integration, and system economics. It also creates an industry precedent for large systems vendors to internalize more semiconductor intellectual property. The structural value transfer would favor foundries and internal design teams over merchant-silicon vendors. OPTICAL, TELECOM AND PHYSICAL INFRASTRUCTURE DISTRIBUTED AI CLUSTERS CREATE A MULTI-YEAR COHERENT-OPTICS CYCLE, BUT CISCO’S ACACIA STRATEGY THREATENS OPTICAL-SYSTEM INCUMBENTS (READ-THROUGH 3) AFFECTED COMPANIES AND IMPACT: Coherent Corp. (COHR: US) and Lumentum Holdings (LITE: US) receive positive industry-demand read-throughs of moderate-to-high magnitude. These are market-level read-throughs rather than assertions that either company is a direct supplier for the Cisco deployments discussed on the call. Ciena Corp. (CIEN: US) receives a near-term positive demand impact of moderate-to-high magnitude but a longer-duration negative competitive impact of moderate magnitude. Nokia Oyj (NOK: Finland) receives a near-term positive demand impact of moderate magnitude and a longer-duration negative competitive impact of low-to-moderate magnitude. CALL SUPPORT: Approximately 40% of Cisco’s $9.3 billion of FY26 hyperscaler AI infrastructure orders consisted of optics. Acacia generated more than $1 billion of orders during Q4. Cisco has shipped more than 850,000 400G and more than 75,000 800G coherent pluggable optics. Management estimated that AI scale-across traffic could be approximately 14 times traditional data-center-interconnect traffic. Cisco also won a managed optical-fiber-network design in Q4 that allows its digital coherent optics to operate directly in third-party equipment. Management described this win as strategically important because it positions Cisco as an alternative to an incumbent and “has the potential to disrupt traditional delivery of managed optical fiber networks.” TRANSMISSION MECHANISM: Distributed AI clusters require substantially more bandwidth between facilities because GPUs located in separate buildings or data centers must operate with increasingly low latency and high reliability. This raises demand for coherent DSPs, optical engines, lasers, modulators, transceivers, pluggable coherent optics, open-line systems, and high-capacity transport equipment. The 14 times traffic estimate, while provided by Cisco rather than independently validated, indicates a meaningful increase in optical content per AI deployment. Coherent and Lumentum benefit at the industry level because greater volumes of 400G, 800G, and future higher-speed links increase demand for optical components and modules. The benefit is not necessarily proportional to Cisco’s Acacia growth because Acacia is vertically integrated into Cisco and may source or design portions of the optical stack internally. The high-conviction conclusion is that the addressable optical market is expanding, not that Cisco’s order growth maps directly into identical revenue growth for every component supplier. Ciena and Nokia face a more complex outcome. Higher telco and inter-data-center traffic should improve aggregate demand for transport equipment. Cisco reported more than 30% telco order growth and linked the acceleration directly to cloud customers’ future bandwidth requirements. However, Cisco is also attempting to disaggregate traditional optical systems by placing Acacia coherent optics into third-party equipment and combining those optics with Cisco line systems and Silicon One routing. This architecture can shift value away from proprietary chassis and vertically integrated optical platforms toward pluggable optics, open systems, and network software. Ciena is more exposed to this competitive transition because optical transport is a larger portion of its business. Nokia is more diversified, reducing the consolidated magnitude. Cisco did not identify the displaced optical incumbent, and no specific customer or vendor displacement should be assumed beyond the competitive mechanism management described. NEAR-TERM TRADING CATALYST: Stronger orders at optical-component and transport vendors would confirm the scale-across thesis. Cisco expects several additional optics and silicon design decisions during the next 6 months. A continuation of more than $1 billion quarterly Acacia orders would support elevated optical-industry forecasts. Conversely, evidence that Cisco is winning managed optical-network deployments from established transport vendors would be negative for the relative share outlook of Ciena and Nokia even if industry revenue remains strong. LONGER-DURATION FUNDAMENTAL SHIFT: The long-duration shift is from isolated AI clusters toward geographically distributed clusters connected through high-capacity optical networks. This increases network and optical content per accelerator. At the same time, coherent pluggables and open-line architectures can reduce dependence on proprietary transport platforms. The most attractive exposure is therefore likely to reside in differentiated optical components, DSPs, and high-speed modules, while traditional system vendors face a larger risk of value migration and pricing pressure. POWER CONSTRAINTS AND MULTI-SITE AI ARCHITECTURES INCREASE PHYSICAL-INFRASTRUCTURE INTENSITY (READ-THROUGH 4) AFFECTED COMPANIES AND IMPACT: Vertiv Holdings (VRT: US), Eaton Corp. (ETN: Ireland), and Schneider Electric (SU: France) receive positive longer-duration impacts of moderate-to-high magnitude. Belden Inc. (BDC: US) and Rockwell Automation (ROK: US) receive positive impacts of moderate magnitude through industrial networking and data-center-facility connectivity. CALL SUPPORT: Cisco stated that hyperscalers increasingly need to connect AI clusters across multiple data centers because of “physical and power limitations in a single data center.” Management expects scale-across traffic to be approximately 14 times traditional data-center-interconnect traffic. Telco orders increased more than 30% as carriers prepared for greater bandwidth demand. Cisco also reported that industrial IoT orders had increased at double-digit rates for 9 consecutive quarters and accelerated in Q4, driven by manufacturing, utilities, and data-center facilities requiring ruggedized networking equipment. TRANSMISSION MECHANISM: When accelerator deployments cannot be concentrated within a single facility because of power-delivery, cooling, land, or physical-space constraints, the infrastructure requirement expands beyond the server rack. A distributed cluster requires power conversion, switchgear, uninterruptible power systems, thermal management, monitoring, networking, security, and optical links at multiple sites. The same compute workload can therefore generate more aggregate physical-infrastructure content when it is distributed than when it is contained within one data hall. Vertiv, Eaton, and Schneider Electric benefit because the physical bottleneck is increasingly power delivery and thermal management rather than only accelerator availability. Distributed architectures can require duplicated or expanded power and cooling systems across facilities. The benefit should be strongest for suppliers with exposure to high-density power, liquid cooling, electrical distribution, and data-center monitoring. Belden and Rockwell receive a related positive read-through from the 9 consecutive quarters of double-digit industrial IoT order growth. Manufacturing plants, utilities, and data-center campuses increasingly require ruggedized Ethernet, secure industrial connectivity, monitoring, and automation. Cisco’s own industrial portfolio creates competitive pressure, but the sustained order pattern indicates that the underlying market is expanding. NEAR-TERM TRADING CATALYST: Continued order growth from data-center power and cooling vendors would validate that physical bottlenecks are forcing incremental infrastructure investment rather than merely delaying compute deployments. Telco capital-spending commentary and large interconnect projects should provide additional confirmation. The near-term impact is likely to be more visible for equipment vendors than for data-center real estate companies because the call does not establish whether the new capacity will be placed in hyperscaler-owned facilities or third-party colocation sites. LONGER-DURATION FUNDAMENTAL SHIFT: AI infrastructure is becoming a system-level deployment involving compute, networking, optics, power, cooling, and industrial automation. The shift toward multi-site clusters increases the amount of supporting infrastructure required per deployed accelerator. This is structurally positive for power-management and thermal-management suppliers and reduces the probability that AI capital spending remains concentrated only in GPU and server vendors. MEMORY, HARDWARE MARGINS AND REVENUE QUALITY CISCO’S MEMORY COST DISCLOSURE IS A POSITIVE READ-THROUGH FOR DRAM PRICING AND A NEGATIVE MARGIN SIGNAL FOR SERVER OEMS (READ-THROUGH 5) AFFECTED COMPANIES AND IMPACT: Micron Technology (MU: US), SK hynix (000660: South Korea), Samsung Electronics (005930: South Korea), and Nanya Technology (2408: Taiwan) receive positive near-term impacts of moderate magnitude. Dell Technologies (DELL: US), Hewlett Packard Enterprise (HPE: US), and Super Micro Computer (SMCI: US) face negative gross-margin impacts of moderate magnitude if memory-cost inflation exceeds the timing or magnitude of customer price increases. CALL SUPPORT: Cisco’s non-GAAP product gross margin declined 270 bps, with management identifying higher hardware mix and memory costs as the primary causes. Cisco has entered strategic supply agreements, made an investment in Nanya, used advance purchase commitments, and built strategic inventory. Management expects price increases to contribute approximately 4-5 percentage points to FY27 revenue growth. Cisco estimated that memory represents approximately 15%-20% of the bill of materials for affected products and contrasted this with server-heavy companies, where memory can represent a substantially larger share of product cost. TRANSMISSION MECHANISM: The call indicates that memory suppliers continue to possess sufficient pricing power and supply leverage to affect the gross margins of a large, sophisticated hardware buyer. Cisco’s willingness to invest in Nanya, make advance commitments, and accumulate inventory indicates that supply assurance is strategically important. This supports a favorable near-term pricing and demand environment for DRAM suppliers. The margin implication is more negative for server OEMs than for Cisco. Cisco stated that approximately 95% of its products are not servers and that its memory exposure is materially lower than the memory exposure of a server bill of materials. Dell, HPE, and Super Micro therefore have greater sensitivity to DRAM cost inflation, particularly in memory-intensive AI servers. If customer price increases lag component-cost changes, gross margins can compress even while reported revenue and unit demand remain strong. Cisco’s experience also demonstrates that revenue growth and gross-margin growth can diverge. Cisco generated 18% quarterly revenue growth and 23% EPS growth despite lower gross margin because operating expenses declined substantially as a percentage of revenue. Server OEMs with lower gross margins, greater working-capital requirements, and less operating-expense leverage may not reproduce the same earnings outcome. NEAR-TERM TRADING CATALYST: DRAM contract-price commentary, inventory levels, and gross-margin guidance from server and networking vendors should be monitored closely. Continued price increases and advance commitments would be positive for memory suppliers. A widening gap between server revenue growth and server gross-margin growth would confirm that component inflation is transferring economics upstream. LONGER-DURATION FUNDAMENTAL SHIFT: The positive memory read-through is not unlimited. Cisco has more than 30 programs intended to reduce memory utilization and reduced Wi-Fi 7 memory requirements by 50% within approximately 90 days. Sustained high memory prices encourage rapid redesign, qualification of alternative suppliers, and lower memory content per product. The near-term pricing cycle is favorable for memory vendors, but engineering efficiency and eventual supply normalization can limit long-term unit-content growth.

• 
  Thankful Person (@misaiah219) reported 33 minutes ago

  @JNavok The problem is you’re treating “digital will get cheaper” like that’s somehow guaranteed when it really isn’t. Just because physical games go away doesn’t mean Sony or publishers are suddenly going to lower digital prices. They’re still going to charge whatever they think people are willing to pay. And Steam isn’t a good comparison either. Steam has competition because you can buy PC keys from different stores and retailers, sometimes for cheaper. If Sony actually lets Amazon, Walmart, Best Buy, GameStop, etc. sell digital PlayStation games and compete with each other on price, then sure, maybe you have a point. But if Sony gets rid of physical games and most people are stuck buying directly from the PlayStation Store, then that’s LESS competition, not more. Physical games aren’t “holding digital hostage.” They’re literally another option competing with digital. Getting rid of that option and just assuming Sony will reward us with cheaper games is a pretty big leap.

• 
  Emil Hartela (@emilhartela) reported 33 minutes ago

  Day in the Life of a $HAYPP Holder, Q2 2026 edition: 7:00am: wake up in Västerås. Report day. The report is at 17:30 this time because they moved it after Swedish close to accommodate American investors. This is a big moment. The Americans are finally coming. 7:02am: feed the parrot. His name is Björn. He says "MIXED REPORT." He learned it from an analyst clip I watched too many times last quarter. I have tried to teach him "structural rerating." He refuses. 7:15am: breakfast. Slot a ZYN in immediately after. Channel checks. 8:00am: post a thread on X explaining what to look for in the print. Six people like it. Two of them are me on other devices. 9:00am: Stockholm opens. Nothing happens. Nothing can happen. The report is in nine hours. Nine hours of a stock trading on nothing while I refresh a page that will not change. 9:30am: check the forums. SnusViking91 has posted. It is the same post as last quarter. I respond in detail anyway. We are the coverage universe now. There used to be more of us. 10:15am: post again. Just a chart. No caption. Let the data speak. 11:00am: seven hours to go. Reread my own revenue estimate from July. It is good. It is very good. I know this because I wrote it and I have read it eleven times. 12:00pm: lunch. Do not eat. The pouch suppresses appetite. Another synergy. 1:30pm: Stefan from work asks what I am waiting for. Explain that Haypp is not a tobacco company, it is a retail media network wearing an e-commerce costume. Stefan asks if that is like Amazon. YES STEFAN. IT IS EXACTLY LIKE AMAZON. He walks away. He always walks away right before the thesis lands. 3:00pm: post again. Two and a half hours. The tension is unbearable and I am the only one feeling it. 5:29pm: Stockholm has closed. The Americans are ready. History is about to occur. 5:30pm: report drops. Fastest growth since IPO. Highest gross margin ever. Nicotine pouch volume up 45%. US volume up 125%. Active consumers all time high. They entered Saudi Arabia, the second biggest pouch market on the planet. 5:31pm: net profit for the quarter: 0,0 MSEK. Earnings per share: 0,00 SEK. Effective tax rate: 102 per cent. Beautiful. Perfectly balanced. The market will absolutely understand this. 5:33pm: post the numbers. Add three exclamation marks, then delete them, then add one back. 6:00pm: the call begins. This is it. The moment the Americans arrive. 6:02pm: three analysts on the line. Pareto, Barclays, Deutsche Bank. That is the whole American accommodation. That is everyone. 6:25pm: Gavin explains Saudi Arabia. Second largest market in the world, roughly 170 million cans, online penetration essentially zero, stable regulation. Asked what it contributes he says it is premature to quantify one month in. Correct answer. Deeply unsatisfying answer. I will be thinking about this for eleven weeks. 6:32pm: they mention the darts sponsorship. We are sponsoring the Professional Darts Corporation. I do not know what to do with this information and yet I feel it is enormously bullish. 6:38pm: average order value down 8 per cent. Because of first time buyers placing small trial orders. This is good. This is the single most bullish bearish number ever printed and nobody outside this call will read the explanation. 6:41pm: Growth segment EBITDA minus 32,5 million. By design. Deliberate. Investment phase. I nod alone in my kitchen. Björn says "MIXED REPORT." 6:44pm: someone asks about the US price reductions. Manufacturer-financed, margin-neutral, offset in cost of goods sold. So the discounts are free. Free discounts, and the stock will fall on them. 6:52pm: they exited Austria. Austria did under 9 million in the quarter. I observe a brief moment of silence for Austria. Björn does not observe it. 6:58pm: operator asks for written questions. There are no written questions. There are never written questions. There are 4 021 shareholders in this company and none of them typed anything. 7:00pm: call ends. Gavin thanks the team. I also thank the team, out loud, in Västerås. 7:15pm: post my full review. It is 2 000 words. It is my best work. It will be read by SnusViking91 and one guy in Oslo. 7:40pm: check the report again. Cash at end of quarter: 16,4 million SEK. That is less than two apartments in Helsinki. Then they signed a 400 million facility a week later. So they were fine. They were always fine. My heart was not fine. 8:30pm: local krog. One Norrlands Guld. Explain to the bartender that the gross margin is at an all time high specifically because Media & Insights is expanding and they are deliberately giving most of it back to consumers to buy share while the window is open. He nods and asks if I want another pouch. He knows me too well. 9:00pm: post again. A single line. "Still a long way to the top." 9:30pm: realise Germany was not mentioned once on the entire call. Not once. I have written 4 000 words about German traffic this year. Open a new document. 10:05pm: cover Björn's cage. Through the blanket he says "MIXED REPORT." It was not a mixed report, Björn. It was the fastest growth in company history. He has never once believed me. 10:15pm: fall asleep with a pouch still in, dreaming of the quarter where the report is great, the call is full, someone asks a written question, and Stefan finally asks me how to buy Swedish small caps while I calmly explain the Media & Insights segment to him at last.

• 
  🇯🇲 جيف (@_yahgetbun) reported 34 minutes ago

  @DiGenius_ @MKBHD Good Or check amazon I bought my mom’s pixel from there brand new from the google store front No issues

• 
  ecom room (@ecomroom) reported 36 minutes ago

  store 2 damn this is such a weird stage i know i have a good product and there is proof people buy it everywhere, tested tons of different ads, audiences and angles and still have a few final creatives in editing that should go live by the end of this week feels like a skill issue on my side lol but i also think some products are just harder to sell on meta, on amazon people already know what they want and tens of thousands of units sell, on meta you somehow have to translate that same demand into the right pain, hook and angle for a completely cold person already took a decent loss testing this one but it still doesnt really let me move on because i know the demand is there, gonna give it a few more shots and see what the weekend brings if it still doesnt click then i will probably move on, already have another product ready to test and almost finished the whole product page today

• 
  Peterk (@_peterk) reported 37 minutes ago

  SpaceX’s next trillion-dollar business was never tickets to orbit. Elon just told staff that AI revenue will soon dwarf rockets and Starlink combined. Not eventually. Soon. It fits the pattern. The company keeps looking like it sells vehicles while actually selling the capability those vehicles unlock at scales normal firms treat as science fiction. Starship is not the product. It is the forklift that lets you put power and compute where Earth grids, land, and regulators cannot follow. Once orbital intelligence becomes the cash engine, launch turns into a rounding error on the same balance sheet. The rocket company outgrows rockets the same way Amazon outgrew books. Critics will still be counting landing legs.

• 
  BrownGuy (@BrownGuyMadeIT) reported 40 minutes ago

  @AmazonHelp Consistent issue with your lazy drivers. Your CSRs are always superb; drivers on the other hand... Need a new career.

• 
  Victor Bigham 🇺🇸 (@Ravious101) reported 41 minutes ago

  @LadyBonBon It's a huge problem that is why Amazon wants to switch to drone delivery. One day we won't have humans giving us our goods and this is the reason why

• 
  Knitty Kitty (@KnittyKittyWK) reported 41 minutes ago

  Amazon putting 6 x 1.25l bottles of iced tea in a flimsy box with wafer biscuits 🙄 Delivery driver was really lovely & apologised & is putting in a report about it as the box was broken & stuff had fallen out. Luckily I got a refund for everything, but bloody hell, think!!

• 
  Alex (@mnmldmg) reported 43 minutes ago

  It’s actually not as bad as it may look like. Amazon has implemented recyclable cardboard packaging for parcels years ago. It’s literally a non issue.

• 
  Hasaam (@hasaamb) reported 46 minutes ago

  Amazon SP-API compliance isn’t “add OAuth and ship” anymore. The new standards include WAF protection, encrypted credentials, login lockouts, incident planning, vulnerability scans, penetration testing, and strict data deletion. The takeaway: compliance is now part of the product not paperwork after it’s built.

• 
  Rosey (@theroseygamer) reported 47 minutes ago

  My biggest issue w/ this Twitch (Amazon) rollout is choice. Auto opt-in w/out informing is not choice. Streamers & users deserve to be fully informed of when & what type of AI (Gen or Predictive) their content, chats, emotes, would be used for & then given the choice to say no.

• 
  Ayush Basral (@ayushbasral) reported 49 minutes ago

  6 years shipping products. amazon -> godaddy. 4 side projects running, 1 masters, 0 free evenings. built a lot, wrote down none of it. fixing that. posting the work here. mostly the parts that don't work yet.

• 
  Subrat S Nanda (@subrat_s_nanda) reported 53 minutes ago

  @AmazonHelp Passed around to 12 different associates and STILL no fix! My account is blocked from placing Amazon Fresh orders to locations I’ve ordered to for months. Other accounts work fine for the exact same address. Great job guys , appreciate to all those 12 for passing

• 
  Olex (Solo gamedev Diablo-like) 🇺🇸🇺🇦 (@OlexGameDev) reported 55 minutes ago

  @valigo @makinbacon21 Case in point: "Disagree and Commit" is a principle taught at Amazon, and engineers are accepted to follow it. It literally means agree with whatever nonsense is going on and agree for several years as the project goes down the drain. I called it "Disagree and Get Paid."

• 
  Sacchi Baat (@sacchibaatx) reported 56 minutes ago

  @digitalindianyt @BU_Tweets @VijayGajeraO If you can't effort a broken item from Amazon and ask for a refund , how can you afford a broken system Why did Cjp formed at the first place Why do we have to stress so much to get what was promised I pay tax - I need service Basic courtesy

• 
  Brian Sauvé (@Brian_Sauve) reported 57 minutes ago

  @lonebulwark We mainly try to avoid the Amazon ecosystem. It is terrible for publishers and authors. Very predatory, terrible back end service, and highway robbery. We are always trying to find ways to support our local economy of Christians with our publishing work.

• 
  Donovan Mowatt (@canhasbucket) reported 57 minutes ago

  @MothxMaam @TwitchSupport Let me try breaking it down for you a little more simple. Sally makes a deal with Bob to use Bob's artwork on the internet. SALLY. DEAL. WITH. BOB.. ok? Sally decides to use her artwork on stream platform. AGREEMENT IS ONLY BETWEEN SALLY AND BOB. Twitch say : We use your content for AI Training OK!? SALLY - agreed to TOS on twitch. SALLY - agreed to TOS with BOB SALLY - Responsible for HER use of assets on the internet. BOB - mad...Bob no want his artwork being AI training food. Bob talk to Sally, and point at TOS that says no AI. This would mean that SALLY would have to decide to stop using her artwork or stop using twitch. If you can't understand that you seriously can't be helped. Simply. Twitch nor Amazon is responsible in the slightest for any 3rd party agreements or "terms of service" You can protest, which I absolutely commend and encourage, but you don't get to ultimately decide what twitch can and can't do based off private agreements. Have a good day.