"Compute market isn't real" is half right. It's real. It just won't look like oil. It'll look like power. History lesson first. NYMEX launched electricity futures in March 1996 with great fanfare: Palo Verde, California-Oregon Border, later Cinergy, Entergy, PJM. By 1999, power futures traded fewer than 200,000 contracts for the year. Natural gas did 19 million. Gasoline did 8.7 million. PJM futures averaged 13 contracts a day. The market was so thin that an order of 25-30 contracts could move the price. In February 2002, NYMEX delisted all of it. But power contract didn't die. It went local. In 2025, EEX traded 9,330 TWh of European power derivatives. ICE did a record 7.8bn MWh of US power. Nord Pool cleared 1,231 TWh of spot. PJM alone billed $80.5B. So power still trades in some size. Just nothing like oil. Rough math (volume x average price): about $1.5T a year of US + European power futures notional, vs about $50T for Brent + WTI. Power is roughly 3% of oil, even though the physical "spot" power market is enormous. Why? Three things oil has that power doesn't. 1/ Storage. A barrel sits in a tank for months. Electricity barely stores at all (US utility-scale batteries built through 2020 average only about 3 hours of output.) 2/ Transport. A barrel gets on a tanker and goes anywhere. A megawatt-hour is stuck on its grid. 3/ Supply inelasticity. In a longer term horizon, Oil supply source is limited and is often ineslatic to the rising Oil demand. This results in the price sensitivity for Oil globally. Meanwhile, electricity is supplied locally via multiple means (solar, nuclear, hydro, geothermal...). Even if there is electricity shortage, given enough time and capital, power supply is often quicker to catch up to the demand. Elastic supply + No storage + No shipping = no global price. You get a thousand local, hourly products. A MWh in Germany at 6pm is a different commodity from a MWh in PJM at 3am. Now swap "MWh" for "GPU-hour." 1/ No storage. An idle H100-hour is gone forever. You can't put it in a tank. 2/ No shipping. It's tied to a site: its power, its network, its latency, its contract terms. An H100 in Virginia isn't an H100 in Singapore. 3/ Elastic supply. Fabs, HBM and power are tight today, but give it time and capital and supply catches up, just like power. On top of that, every chip generation partially obsoletes the last one. Imagine if a new grade of crude made your barrels worth less every 18 months. So when we asks whether compute is a standardizable commodity: Yes, the way POWER is. Not the way OIL is. I expect regional indices, hourly and monthly contracts, bilateral OTC deals, and basis everywhere. I highly doubt there will be a standardized Brent of compute. But, there could be a PJM of compute. Very different

Finally got our Proof of Reserves out. Due to the unique design of Tenbin assets, there were quite a bit of juggling we had to do with the PoR setup. Thanks @chainlink for accomodating and helping us with the setup! Check the doc for more details:

ibovespa -17.6%, BRL -11% (worst day since '99), brazil 10Y 12.9% → 16% in one session I said it monday: BRL sell is overshoot. Fiscal fear is real, but selic 13.75% + a 16% 10Y is a lot of carry to get paid for. Panic sellers have set a new entry

Someone tried to sell ~$630 of tokenized reddit (RDDTc) on base and got quoted $0.26 lol. 99.95% price impact Exactly what i meant last week. no atomic mint/redeem = you're stuck with whatever AMM liquidity exists 1:1 backed ≠ liquid

Have spent few days looking for on-chain stock that offers the BEST & consistent pricing across sizes. But disappointed that only 1 provider ticked all the boxes for me... Mainly two observations: 1. There is a large amount of arb flow that make the historical executed stock price on-chain looks better than it actually is (this applies uniformly to stocks that heavily rely on AMM, such as @bstocksfinance, @RobinhoodCrypto, @base stock). Most retail traders will likely suffer 15-25bps in spread from market price on average, while arber compete to capture negative spread. PropAMM helps here but limited coverage with size still. 2. All AMM based stocks does not have the atomic mint and redeem feature enabled (unlike @Ondo/@xStocksFi ). This significantly constrains the access to the primary liquidity sources (i.e. Alpaca's brokerage giving access to underlying market). This forces all the swaps to rely on AMM liquidity. Conclusion: AMM still suck for stock trading. Stock providers with atomic mint/redeem suck less. But the latter is still more expensive than to trade directly via local brokers in most cases. Still better than AMM on average. Given enough time, perhaps propAMM will improve the price execution for retail flow. But unfortunately, we are still far from "cheap" stock trading on-chain for most stock providers.

This wraps up @blockchain_rio. Thanks to everyone who participated in our product survey. We received amazing feedback from the community and truly appreciate everyone who shared their thoughts. Please keep your eyes out for the survey airdrop & more product updates!

Met a lot of amazing people in Rio and got lots of great feedback on Tenbin! Thanks to everyone who completed our surveys 👍 Now heading to Sao Paulo. Down to catch anyone in town! DM open ;)

This is 2026, DeFi today is no longer the same DeFi back in 2023. The same old DeFi playbook won't work anymore (nor is it sustainable as proven many times in the market). Gotta carve our own path. Not ez, but has to be done.

Since 2021: MXN TIIE return (MXN T-bill equivalent): +78.8% U.S. short-term rates: +19.0% Even after currency moves, holding pesos + earning Mexico's overnight rate beat US T-bill rates by 59.9%. In plain English: the MXN peso carry trade paid more yield + the currency upside.