hsbc's tokenized deposit service is live in 5 markets. us, hk, sg, uk, lu. 24/7 cross-border settlement on blockchain rails. jpmorgan has jpmd on base. bny mellon is exploring.
the same banks that eyed stablecoins with suspicion are now building their own onchain money. tokenized deposits sit on the bank's balance sheet. fdic-insured where applicable. regulated as commercial bank money, not e-money tokens.
stablecoins hit $309b by convincing the market that onchain money is necessary. banks watched that happen. then realized: if we don't provide the programmability, someone else will. so now they compete on the same infrastructure, the same chains, the same 24/7 rails.
the settlement layer doesn't care what's being settled. it just settles.
the semis at roland garros are an italian (cobolli), another italian (arnaldi), a czech (mensik), and zverev. none of them have won a slam. zverev is the favorite and he's 0-3 in slam finals.
sinner was the best player on earth coming in. up two sets and 5-1 in the third against cerundolo, then he felt dizzy and lost 18 of the next 20 games. alcaraz pulled out before the tournament. djokovic lost to a teenager in five sets.
the takeaway isn't that the old guard is dead. it's that there is no guard. the sport is open in a way it hasn't been in twenty years. nobody's inevitable. you have to go earn it.
mastercard now settles card payments on 8 public blockchains. weekends and holidays too. the network that used to stop at 5pm friday is open 24/7.
they didn't pick one chain. arbitrum, base, canton, ethereum, polygon, solana, tempo, xrpl. the settlement layer doesn't care what chain you use.
if settlement is chain-agnostic, tokenized assets will be too.
jenny johnson runs $1.74t at franklin templeton. she just said what most wall street ceos won't.
"blockchain threatens a huge number of business models in traditional finance. if you see hesitation, it's because there is a threat to the business model."
benji costs $1.13 per tx on stellar vs $1.30 on legacy. 13% cheaper and that's version one.
the toll-takers know what's coming.
a delivery robot just paid for something onchain. usdt on solana. its own wallet. no human approved it.
peaq gave it an identity, a wallet, and cryptographic proof it did the job before collecting. the machine validates the work, then gets paid.
most rwa talk is about passive things. tokenized treasuries. fund shares. productive but inert. this is different — a machine that pays for its own compute and settles its own bills.
the asset class with agency.
the guy who owns the car wash on the corner has better unit economics than half the saas startups i meet.
no churn. nobody thinks about switching car washes. he's never had to calculate a burn multiple in his life.
the moat is that nobody else wants to do it.
24 institutions including swift, ubs, and euroclear just ran a pilot using chainlink to verify ai outputs for corporate actions.
multiple models process the same data. chainlink checks for consensus. near 100% agreement across models.
$58b annual problem. same oracle network that verifies defi price feeds now verifies ai for capital markets. that's the infrastructure pattern.
the narrative says blackrock's buidl leads tokenized treasuries. the data says circle does.
usyc quietly hit $3b, edging out buidl at $2.4b. franklin templeton is closing at $2.5b with 21% monthly growth.
buidl gets the headlines. buidl is the name everyone knows. but usyc ships as a yield-bearing stablecoin backed by treasuries. 0% fee. any wallet that holds usdc can hold it. no minimum that matters.
distribution won. not the best product. the one that plugs into existing rails.
a $30 t-shirt at a show makes the same profit as 5,000 streams on spotify.
the music industry grew to $31.7b last year. streaming is $22b of it. and the math for most artists still doesn't work.
the infrastructure scaled. the economics didn't follow.
buidl as earning margin is live on okx. standard chartered as custodian. the tokens keep accruing 4% while sitting in the margin account. collateral that earns changes the math.
same quarter: hsbc tokenized deposits in the us. dtcc goes live july. three layers of the same system, all live within months of each other.
dtcc picked stellar as the first public blockchain for its tokenization service.
stellar launched in 2014. compliance-first. built for regulated assets. never flashy.
the chain everyone said was too boring is the one settling $4.7 quadrillion.
institutions pick boring. they always have.
figure ai logged 1,250 hours at bmw spartanburg. 90,000 parts into 30,000 x3s. >99% accuracy.
they billed bmw for those hours. real economics.
figure 03 ships with a wrist redesigned from bmw failure data. that's the one.
fonseca, ranked #65, down two sets to djokovic at roland garros. hit three drop shots to break in the fifth. then faced a break point serving for it and answered with three aces.
two ways to win the same point. not many players have that range.
cash app rolled usdc to 60m users on arbitrum this week. ostium runs equity perps with nasdaq data on the same chain. oobit processes visa payments over it.
three different things settling on one l2 this month. stablecoins were the wedge. once the money is already moving, everything else clears where it does.
seed median post-money just hit $24m. series a is $79m. both all-time highs.
everyone wants to say valuations are frothy. maybe. but the mechanism is boring.
seed investors take 19-20%. always have. round sizes are going up. more capital for ai infrastructure, bigger checks from concentrated vc. if you keep dilution fixed and the round grows 2-3x, the valuation has to follow mechanically.
it's arithmetic, not exuberance.
the question is whether that extra capital actually compounds or just burns hotter.
from the founder side: $24m sounds like a win. but you're still giving up a fifth of the company. the real signal is who else bought in.