SWIFT currently handles about $5T in cross-border transactions per day, connecting around 11,500 institutions across nearly 200 countries and territories.
But traditional systems still have weaknesses: • Transaction costs: about 1–4% • Settlement time: 1–5 business days
Meanwhile, stablecoins, tokenized deposits, and blockchains are directly competing on speed + cost.
SWIFT also has to change. The network has just rolled out a blockchain-based ledger, and HSBC and Standard Chartered have carried out the first realtime transactions, cutting settlement time down to just a few seconds.
And the market scale could be even bigger:
BNY: stablecoin + tokenized cash could reach $3.7T by 2030. Citigroup: tokenized securities could reach $5.5T in the same period.
This is no longer a crypto-versus-banking story about replacing banks.
The banks are already building the rails.
I think the real competition won’t be SWIFT vs crypto, but rather which payments system can settle the fastest with the lowest cost.
Do you think stablecoins could become the main cross-border payments infrastructure in the next 5–10 years?
AI is minting billionaires faster than almost anything else.
According to a report by Altrata cited by Fortune, the number of global billionaires in 2025 reached a record high of 3,795 people, with total wealth of $15.1T, up 12.8% year over year.
Notably, AI is one of the biggest drivers behind this increase in wealth. As technology companies expand their investments and operations in AI, the assets of many founders and major shareholders are also rising sharply.
There are currently 29 billionaires with wealth exceeding $50B, including Larry Page, Elon Musk, and Jeff Bezos.
Another figure is pretty wild too:
The $15.1T in billionaire wealth is equivalent to nearly 1/4 of the total market capitalization of stocks in the S&P 500.
AI isn’t just changing the tech industry.
It’s changing who owns the wealth.
I think the bigger question is: if AI continues to generate productivity and profits at this scale, how far will the wealth gap widen?
Do you think AI will create more billionaires, or will it eventually distribute value more widely?
🚨 JUST IN: Cronos just hit the emergency stop button.
Cronos has temporarily halted the blockchain after the attacker withdrew funds from Tectonic, the largest lending protocol on the network.
Damage is estimated at about $75M: • ~$6M was transferred to Ethereum • ~$60M is still stuck on Cronos • Cronos has stopped block production to address the incident
Tectonic currently holds about $121.6M, equivalent to ~46% of all DeFi on Cronos.
Cronos: “Pause the chain.” Attacker: “Wait… WHAT?” 💀
Now the biggest question is: how will the $60M that’s stuck be handled—rollback, blacklist the attacker’s address, or restart the chain?
Zcash just made private transactions 14x faster on mobile.
Zakura, the development team for the Zcash ecosystem client, has just released Zakura Common — a new cryptographic library and protocol designed to significantly boost Zcash performance.
Most notably is the wallet experience: in some cases, transaction creation time can drop from over 3 seconds to under 200ms.
Zakura Common has been open-sourced under the MIT/Apache 2.0 license and does not require a network upgrade. Zakura 1.3.0 will move to this new technology platform.
I think this is a fairly important upgrade for Zcash, because privacy doesn’t just need to be strong in cryptography—it also needs to be fast enough for users to actually want to use it.
14x faster proofs on mobile. Privacy UX just got a serious upgrade.
Do you think improving speed like this is enough to help ZEC compete more strongly in the privacy space?
The sandwich bot made $295M. Then someone figured out how to sandwich the sandwicher.
JaredfromSubway.eth, a bot specializing in executing sandwich attacks on Ethereum, has earned a total of about 117,007 ETH, equivalent to nearly $295M at current prices since March 2023.
But by June 2026, this very bot became the victim. An attacker deployed 66 fake smart contracts, exploiting the bot’s automated transaction logic to steal at least $7.5M in ETH and stablecoins.
The funds were then sent into Tornado Cash and have not yet been recovered.
In essence, a sandwich bot makes money by spotting large transactions in the mempool, buying first → waiting for the transaction to push the price → selling immediately afterward.
But this time, someone found a way to turn that exact logic into a trap.
Even more interesting is that sandwich attack activity is also dropping sharply: monthly revenue was around $10M in the late 2024 period, falling to about $2.5M in October 2025.
Real MEV is a game where the hunter can become the prey.
Do you think this is one of Ethereum’s most notable “reverse exploits”?
🚨 JUST IN: Trump Coins says it never authorized $GOLD.
realtrumpcoins has denied launching, promoting, or authorizing any digital token, calling the claims “completely false.” The team says it is working with law enforcement to investigate.
The $GOLD story is even more noteworthy: after an account allegedly related to it posted the contract, the market cap once jumped to $66M, then the tweet was deleted and the token crashed from $55M to about ~$1M in roughly 30 seconds. 💀
Finally, those behind it are said to have sold all of the tokens and made around 9,784.6 SOL (~$1.01M).
A bull run isn’t necessarily the scariest part. A fake official account is. 😭
Do you think this kind of scam will show up even more as the market heats up?
A trader on Polymarket has just placed a bet of nearly $938K that the Fed will raise interest rates by 25 bps at the September meeting.
This account previously earned more than $270K in less than a week; it now holds 2.02 million shares at an average price of 39.3¢.
More notably, this wallet is also holding about $3.06M betting that the Fed will not cut 25 bps, with an average entry price of 98.2¢.
So instead of just betting on “cut or hold,” this trader is strongly leaning toward the scenario:
No rate cut. Maybe even a hike.
The next FOMC meeting will be held on 15–16/9.
What I find interesting is this: this isn’t some anonymous trader who just jumped into the market. Someone who has just made over $270K is putting nearly $1M on a scenario that much of the market may not yet be pricing in.
Conviction or a degen bet?
Do you think the Fed will cut, hold, or hike in September?
🚨 JUST IN: Solana is upgrading everything except the price. 💀
Transaction V1 launches 9/9. Lower rental fees begin next week. Slot time continues to be shortened. Alpenglow is expected to go live on mainnet in October. And in November, there’s also Scale or Die.
The Solana team is like:
“The network is already fast… but what if it’s even faster?” 😭
And for $SOL holders:
“Sure, the tech upgrade is great… but when will the price be upgraded?” 💀
Tech is scaling. Now let’s see if the ecosystem can scale too.
🚨 The U.S. could be taking control of Venezuela’s oil pipeline.
President Donald Trump said that a new oil deal could bring the U.S. 55% of output from production projects in Venezuela.
According to reports, the project will focus on 17 strategic fields in the Orinoco Belt and Lake Maracaibo, with proposed production rights of up to 100 years.
Notably, the plan could also attract nearly $100B in private investment capital, while Venezuela expects to collect up to $209B in tax revenue.
But there’s one major issue: the entire deal has not yet been announced, the name of the private operator has not been disclosed, and the legal basis still raises many questions.
If this massive flow of capital is actually deployed, the impact won’t just be limited to Venezuela.
Oil → inflation → liquidity → risk assets.
I’ll keep an eye on whether this agreement truly becomes a boost to global oil supply, or whether it’s just a plan on paper.
Do you think the U.S. is trying to control Venezuela’s oil supply, or is this simply a gigantic investment deal?
Bitcoin just got its first quantum-resistant test on mainnet.
StarkWare has just completed a Quantum-Resistant Bitcoin (QSB) transaction on the Bitcoin mainnet without needing a soft fork or hard fork.
The transaction only cost 10,000 sats, the test lasted a few hours, and the GPU cost was about $150–200.
Notably: QSB helps reduce the risk from quantum attacks, but users still need to proactively move their BTC. It cannot protect assets whose public key has already been exposed.
AI is making investment scams harder to spot — and far more expensive.
In 2025, losses from investment scams in the U.S. exceeded $8B, up 38% from 2024, with 144,041 consumers reporting being victimized. Notably, AI is helping scammers upgrade nearly the entire process:
• Clone voices and generate deepfake videos
• Impersonate financial experts or celebrities
• Run social media ads for fake projects
• Create trading websites with fake balances and profits A fairly sophisticated tactic is to have victims successfully withdraw a small amount to build trust. Then the platform asks them to deposit more money or pay a “fee” to withdraw a larger amount.
Meanwhile, FBI data shows that reported damages from cybercrime in the U.S. in 2025 reached up to $20.877B, including complaints related to crypto totaling about $11.37B in losses.
AI didn’t make scams smarter. It made scammers scalable.
In my view, the biggest risk is no longer a crypto project looking “too fake,” but rather a scam packaged by AI with enough professionalism to appear like a real investment opportunity.
Do you think AI will help detect scams faster, or will it ultimately help scammers stay one step ahead?
OpenAI just cut Cursor off after the SpaceX acquisition.
OpenAI has just announced that it will end its model supply contract for Cursor, with the expected access stop date being 12/11/2026.
The direct reason is that SpaceX has acquired Cursor, completing the deal on 14/8. OpenAI said it cannot be certain that SpaceX will use its technology in accordance with the service terms, and it also cited a history of contract disputes with companies owned by Elon Musk.
Notably, Cursor is not directly accused by OpenAI of violating the contract. The termination was triggered by the change-of-control clause following the acquisition. OpenAI also said that future models such as Astra will not be provided to Cursor.
Cursor is currently not entirely dependent on OpenAI. The platform has integrated many other models, and SpaceX wants to push Grok + AI coding + compute after the deal.
One acquisition just turned an AI partnership into a model-access war.
In my view, this isn’t only about Cursor losing access to OpenAI. It shows that AI infrastructure is increasingly being pulled into competition between ecosystems, where owning an application can determine which models are allowed to run inside it.
Do you think Cursor will fully tilt toward Grok after losing OpenAI, or will it still find a way to keep OpenAI through an API/gateway?