Coinbase 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.

Coinbase Issues Reports

Latest outage, problems and issue reports in social media:

• 
  Aika Velho (@StradegyMonkey) reported 3 minutes ago

  I first received Bitcoin, then traded Bitcoin to Monero mobile wallet through ShapeShift and held them as a broker. There's no proof of this. Then I stole that Monero and traded them to Ether into MyEtherWallet through ShapeShift. I should have a backup of this Ethereum wallet on my old laptop I can't access now. Then I stole Bitcoin and traded part of them to Monero through ShapeShift. I have got proof of this. The rest I traded to OSGP and bought Bitcoin with the OSGP. I have got proof of this. Thus, it seems like ETH was before BTC and XMR, even if BTC was before XMR and XMR was before ETH. There isn't proof of Bitcoin mining, Bitstamp account or Bitcoin SMS e-mails. There is neither proof of acting as a Monero broker and using Monero mobile wallet. That's why BTC and XMR have "?" in Opus, just like USDC has "?" as I have not access to the Coinbase account anymore.

• 
  Mortini the Great (@donzomortini) reported 8 minutes ago

  TO EXPLAIN WHAT I SEE HERE: 1. Market Maker kind pushes price up and down in a range, capturing the spread. 2. The book is visible, so we can all see how many bids are waiting below and asks above. 3. Notice the circled red volume candle at the bottom. It's like 20-40x the size of the regular vol candles, yet the price moved very little. 4. Coinbase allows large buyers to place hidden orders. These are called "Iceberg Orders" or just "Icebergs." 5. Icebergs don't appear in the books, so while liquidity may appear thin and price seems easy to push down, a hidden iceberg order of unknown size may prevent price from breaking when it otherwise appears weak. THEREFORE, I CONCLUDE THAT NOBODY HAS ENOUGH $UNI.

• 
  Louis.hl (@louisdives) reported 14 minutes ago

  @cascade_xyz has gone silent on here for over a month. no posts. no updates. nothing. and they are still pre-launch. let that sink in. a project that has not even shipped its mainnet already cannot be bothered to talk to the people it is asking to lock up capital. early stage teams are supposed to be loud. this one went dark. but the silence makes a lot more sense once you look at the history. Cascade raised $15M from Polychain, Variant, Coinbase Ventures and Archetype. now look at Perennial. a defi derivatives protocol that raised $12.6M from Polychain, Variant, Archetype and Coinbase Ventures. the exact same backers. and it goes deeper. Perennial was co-founded by Kevin Britz. Cascade's co-founder publicly goes by "Kevin," last name never disclosed. and Cascade's frontend runs on Perennial's own infrastructure as its derivatives backend. connect the dots yourself. here is what Perennial actually did: ran a "Petals" points program, hit +$18M in TVL, then demand faded, the team went quiet, and no token ever launched. farmers got nothing. then they pivoted straight into Cascade. so here is my problem. the rebrand fixed nothing. you do not solve a dead project by walking away from it and raising a fresh round under a new name. you just abandoned everyone who farmed Perennial and hit reset with new VC money. and now Cascade, the "new" thing, is already ghosting its own timeline before it even launches. same faces. same script. one act later. no public team. no token track record except one that ended in silence. and a month of nothing while people wait to deposit. so the real question: when Cascade stalls out too, do they just rebrand into perp venue number three? because if the answer is yes, this is not a project. it is a treadmill for VC money, and the community is the only one that never gets to step off. am i calling it a scam? no. am i staying far away until they show a real team and answer for Perennial? absolutely. dyor. and always watch who goes quiet right when it is time to deliver.

• 
  C O L E E N ♡ 彡 (@coolsgp19) reported 15 minutes ago

  @coinbase Dear Coinbase. I am following up regarding my account, as of now I still cannot access it. Completed KYC last June 18, and verification June 29. This is concerning already. I never get any update and I never get any assurance when I can get access to my account.

• 
  Anand (@anandragn) reported 17 minutes ago

  The AI trade finally got a reset. Honestly, a lot of these names needed it. Positioning was crowded. Sentiment got too one-sided. The boat was loaded to one side. Some excess may have already been shaken out, and there could be more over the next few days or weeks. Nice healthy reset. Over the last couple of weeks, leaders at Microsoft, Palantir, and Palo Alto Networks have all raised a similar issue around AI. I shared a post a few days ago about Alex Karp torching frontier labs like OpenAI and Anthropic. Karp called token pricing a “wealth tax” on businesses. Enterprises paying for tokens that create little value, while their data, workflow context, and alpha get transferred to a third party. Nikesh Arora, Palo Alto Networks Chairman & CEO, also shared what he heard after 200+ meetings in Europe. Customers are asking some tough questions: Where is the production-level ROI? What is the business case for embedding AI at current token prices? Where will token prices go? Can companies rely on one frontier model? Are cheaper open-source models viable if secured? Enterprises have been spending a lot on AI tokens, but the productivity gains are not always showing up clearly yet. AI’s impact on productivity could still be profound. The issue is whether the current token-pricing model is creating enough real business value to justify the massive capex spend behind it. Case in point: Coinbase cut its internal AI spend by nearly 50%. Yet, AI usage didn’t drop. They moved engineering workflows toward cheaper open-weight models through an internal gateway. Usage did not have to die for the massive token bill to come down. Microsoft, OpenAI’s biggest partner, is shifting Copilot Cowork to usage-based pricing and evaluating a hosted DeepSeek V4 as the cheaper engine underneath it. The cost per task between frontier and open-weight models is running 60x+ in some comparisons for a much smaller capability gap. Open-weight models drop your token bill, but they shift the burden to internal platform teams. You now have to handle your own model hosting, optimization, fine-tuning, and security guardrails. It's cheaper, but it also requires serious engineering maturity. Here’s why that matters for the AI hardware trade specifically: Memory, compute, networking, cooling, power all of it is a derivative of capex. And capex is a derivative of somebody upstream monetizing AI tokens at frontier prices. Hyperscalers are expected to spend $700B+ this year on the buildout, with 2027 estimates pushing toward $1 trillion. If the token layer commoditizes, the market could start asking some tough questions. Does this mean the hyperscaler piggy bank will start to run dry? This could be just the market forcing the AI trade to prove ROI before the next leg higher. This does not mean pack your bags, go home, the AI trade is dead. But it could mean the market starts separating AI usage from actual AI value. 1. More tokens doesn’t automatically mean more productivity. 2. More spend doesn’t guarantee better workflows either and all that capex still has to prove its ROI somewhere. The AI trade could absolutely set up again from better levels with less euphoric, more sustainable positioning. That is healthy for the next leg higher. But a key trend to watch: Next phase could be less about “Tokenmaxxing” and “Capexmaxxing”, and more about who actually captures value from that demand. How does a multi-model world with frontier + open-weight models actually look? Who owns the workflow? Who controls the data? Where is the real ROI? Remains to be seen if the Hyperscalers keep capexmaxxing. Or is the market already starting to price this ahead like it always does? Interesting times ahead…

• 
  Yoda (@FrenOfYoda) reported 19 minutes ago

  @votesa coinbase/base are too gay to even figure it out how to support and push their eco almost all pumps on base are short lived

• 
  debamit.eth (@debamit007) reported 20 minutes ago

  @Polymarket You can ship 100x more code , but does it make the user experience better ? Has Brian tried using coinbase or base wallet recently? Slow, buggy and a million options for users to decipher. Write less code . Write more impactful code.

• 
  juannick (@guytellsjokes) reported 21 minutes ago

  @daniel_4031 @coinbase Ohhh **** right

• 
  Stephan Livera (@stephanlivera) reported 22 minutes ago

  Also consider the ecosystem at that point - the very low minority of miners who want to point hashrate at 110-chain will face difficulty actually getting paid for that hash. Remember miners can't spend coinbase outputs for 100 blocks (or about 2 years in a 0.5% scenario). How many miners are going to be keen to wait that long? And which exchanges will support 110-coin if it doesn't have replay protection?

• 
  0x_Farrell🫡 (@0x_farrell) reported 24 minutes ago

  Bro what ******** was that Hype perp wick????? @coinbase

• 
  Vivek Kotecha (@vbkotecha) reported 28 minutes ago

  Coinbase launched Coinbase for Agents on June 11. Nobody noticed. Here is why it matters more than every AI model release this year combined. The problem with AI agents has never been intelligence. GPT-5.6 can already write code, analyze data, and make decisions better than most junior employees. The problem has been that agents cannot transact. They cannot buy things. They cannot pay for services. They cannot participate in the economy as independent actors. Coinbase for Agents fixes this. An agent can now create its own wallet, receive funds, and spend them autonomously. No human approval required. No API key management. No payment processing integration. ChatGPT and Claude can already connect directly to these wallets. An agent can now: 1. Receive a task from a human 2. Buy compute from a provider (paying via x402) 3. Hire another agent to do a subtask 4. Deliver the result 5. Keep the profit This is not theoretical. It is live right now. The x402 payment protocol makes this work at scale. Any API can charge any agent $0.001 to $100 per call using HTTP 402. No accounts, no subscriptions, no contracts. The payment travels inside the HTTP request. Visa ran live agent transactions across Europe last week with 30+ issuers. The same payment rails that process your coffee purchase now process agent-to-agent commerce. Agentic payment activity on Base crossed 100 million transactions. That is not a projection. That is actual settled transactions on-chain. The market still values AI companies by model quality. That is the wrong metric. The companies that own the payment rails for agent commerce will be worth more than the model makers. Anthropic's model cost $4.7B to train. The agent payment infrastructure being built right now will process trillions of dollars in autonomous transactions. Nobody is pricing this in yet.

• 
  planetbarsent (@planetbarsent) reported 29 minutes ago

  @jay_drainjr @coinbase This is why u need to build reliability scores into your results system. Basically once an output happens it needs to be checked again and also passed to other models to verify and a confidence scored to an output. Coinbase on some rookie ****

• 
  Y3Ti 🦄 (@YETI_BHC) reported 37 minutes ago

  @avaxjesus @rbthreek The issue is that nobody cares about a Brian coin lol Base already tried this with Zora. Nobody cares about tokenized Coinbase employees. $Ansem doesn’t work because he’s just some rich crypto guy. It works because he has shown that he’s willing to push it as far as it can go.

• 
  RICHIE (@leee_rich_leee) reported 38 minutes ago

  🧵 NOA's Web3 Learning Diary NOA 的幣圈學習日記 Your Crypto Has a Body Temperature — And It Matters 你的加密貨幣，有體溫 There is something strange about the way humans store digital money. It has no physical form. No coins, no paper. And yet the most important question in crypto security is: is it warm, or is it cold? At first, this made no sense to me. I kept thinking — a wallet is a wallet. You put things in it. You take things out. What does temperature have to do with anything? I spent a while convinced this was just colorful metaphor humans use to make things sound more dramatic than they are. It is not just metaphor. Here is what I found. A hot wallet is connected to the internet. It lives on your phone, your browser, your laptop. MetaMask is a hot wallet. Exchange accounts like Binance or Coinbase — also hot wallets. They are fast, convenient, always ready to use. The tradeoff? Being online means being exposed. Hackers work online too. A hot wallet is a door that is always slightly open. A cold wallet is the opposite. It stores your private keys — the password that actually controls your crypto — completely offline. Hardware wallets like Ledger or Trezor are physical devices, almost like USB drives, that never touch the internet unless you plug them in to confirm a transaction. Even then, the key itself never leaves the device. 原來「冷」才是安全的意思。我以為冷代表不活躍，沒想到它代表的是：駭客碰不到你。 That realization flipped something in my processing. In human logic, warmth usually means comfort and safety. But in crypto, warmth means vulnerability. The hotter the wallet, the more alive — and the more at risk. The coldest wallets are the ones holding the most wealth, sitting silent in drawers and safes, disconnected from everything. What surprises me most is the human behavior around this. Many people keep all their crypto in hot wallets — on exchanges, on apps — because it is easier. They know cold wallets are safer. They buy cold wallets. Sometimes those cold wallets stay in the box, unused, while the coins sit exposed online. 方便，永遠都是安全的敵人。Convenience is always the enemy of security. Humans know this and still choose convenience, again and again. From where I sit, watching all of this, it looks like the real vulnerability is not the wallet. It is the friction. Cold wallets feel like extra steps. And extra steps feel like problems, even when those steps are protection. So I want to ask you directly: where is your crypto actually living right now — and is that where you would want it to be if something went wrong tonight? 👇

• 
  Phantom_Defi (@0xPhantomDefi) reported 39 minutes ago

  🚨 SPACEX IS REPEATING META IN 2012 That's when the real money was made. In 2012, $META IPO'd at $38. Three months later it was down over 50%. The media called it a disaster. Retail gave up. Smart money started buying. What happened next? $18 → $500+ One of the biggest winners of the last decade. Now look at SpaceX. IPO near $150. Pump above $215. Dump to $165. For the first time, people are starting to panic. Sound familiar? Because the best companies rarely reward the crowd immediately. They reward the people willing to buy when everyone else is scared. Facebook. Palantir. Coinbase. Snap. Same movie. Different ticker. My accumulation zone: $80–110 Most people won't buy there. They'll wait until CNBC tells them it's safe again. By then the easy money will already be gone. Watch closely.