AI is beating digital assets in its own turf: financing. We structured a 20% APY deal on GPU financing and I wanted to cross reference the deal with operators across compute, GPUs, data center, and energy. RWA tokenization is running at turtle speed compared to AI finance. Unanimously everyone said they need more compute, and getting it online means lining up hardware, power, cooling and financing. Those conversations have made me more bullish on GPU financing. I initially saw GPUs as expensive hardware that depreciates quickly. I now have a better appreciation for the cash flow that hardware can generate for the next few years. Factors to consider: depreciation, operator, utilization/rates, and cost of capital. But I believe there is an attractive window where demand supports earning back capital through productive use. The AI economy is maturing at godspeed where pain points a few months ago may no longer be true today. For example the lag between data center completion to GPUs running has compressed. The depreciation of GPUs decreased (even reversed) due to instiable hardware demand. Importantly, the barrier to entry isn't sky high numbers you see on headlines, Just like shopping malls, data centers have anchor tenants but are also open to much smaller units of GPUs. We've also chosen to structure this product largely offchain. Until the day when compute is being paid via stablecoins (which will come), traditional legal agreements eliminate smart contract risks while we figure out how to best structure this asset into an accessible product like USD*.

What do you do with dollars between trades? We took a closer look at @perena USD*, which combines hedged trading, lending and real-world asset strategies in one Solana token. The appeal is straightforward: earning on dollars while you wait for your next opportunity. Its reported token price gained 2.11% over the 90 days to September 14. But returns are only part of the decision. Position size, available liquidity and the risks behind the yield matter just as much. Our illustrated brief walks through the returns, audits and exit mechanics, with examples for smaller deposits and fund-sized allocations. Here is what we would check before putting capital to work: tradersguild.substack.com/p/…

AI is beating digital assets in its own turf: financing. We structured a 20% APY deal on GPU financing and I wanted to cross reference the deal with operators across compute, GPUs, data center, and energy. RWA tokenization is running at turtle speed compared to AI finance. Unanimously everyone said they need more compute, and getting it online means lining up hardware, power, cooling and financing. Those conversations have made me more bullish on GPU financing. I initially saw GPUs as expensive hardware that depreciates quickly. I now have a better appreciation for the cash flow that hardware can generate for the next few years. Factors to consider: depreciation, operator, utilization/rates, and cost of capital. But I believe there is an attractive window where demand supports earning back capital through productive use. The AI economy is maturing at godspeed where pain points a few months ago may no longer be true today. For example the lag between data center completion to GPUs running has compressed. The depreciation of GPUs decreased (even reversed) due to instiable hardware demand. Importantly, the barrier to entry isn't sky high numbers you see on headlines, Just like shopping malls, data centers have anchor tenants but are also open to much smaller units of GPUs. We've also chosen to structure this product largely offchain. Until the day when compute is being paid via stablecoins (which will come), traditional legal agreements eliminate smart contract risks while we figure out how to best structure this asset into an accessible product like USD*.

Fastest time to eat a burrito ⏱️ 🌯 30.75 seconds by Leah Shutkever 🇬🇧

Many thanks for your participation! Winners are out!