Goldman Sachs on AI:
"We now expect a combined $5.3 trillion of capex spending for the four largest hyperscalers from FY2025 to FY2030 (Meta, Microsoft, Amazon, and Alphabet). We highlight a baseline aggregate capex estimate of $7.6 trillion between 2026 and 2031, across compute, data centers and power."
I AM SO SO GLAD CONST RESPONDED HERE TO SOMETHING THATS ALWAYS BUGGED THE HELL OUT OF ME WRT $TAO COMMUNITY:
People always say Bittensor is "freedom" or "decentralization" or "protection" from the centralized AI overlords. Yeah yeah, all that. But that's nonsense if the outputs/products can't compete.
Const is still firmly over the target, that target being: decentralized AI can actually IMPROVE results vs. centralized labs. It's not just about parity and freedom, or morality.
It's actually about making the maximally powerful, economically efficient juggernaut you can make. It's about better solutions, real products that outperform, and yes, this translates to revenue. It's in the name: monetary computing.
It's about the long tail is still ASI, and that's still only reaches maximum potential on permissionless, resilient, distributed substrate, the kind which is about coordination, not raw capital. It's about thousands of SOTA benchmarks in unison, running on truth, vs. overweighted generalists with a human agenda.
Know anyone building that? I do, he's had his eye on the ball for over a decade.
$TAO - You don't own nearly efuckingnough.
reserved, interconnected, frontier training compute is the most expensive commodity in the space. On the flip side, compute that can inherently be resold (e.g. spot instances) and lower grade compute can trade at a 90% discount. If you can orchestrate this compute to create training compute at 70% of frontier performance (for 10%) of the price, you can disrupt the entire mechanics of the compute marketplace.
we believe the future of training is inherently resolved around making training workloads liquid. Iota is designed explicitly for this future.
Yes, it's significant.
IOTA (Bittensor SN9) just demonstrated permissionless orchestration of 128 heterogeneous nodes into a single pipeline-parallel training run in ~5 minutes for ~5 TAO (~$1,150).
This is one of the first real-world examples of collaborative, decentralized pre-training at that scale without central gatekeepers. The architecture shards layers across nodes, which is why they claim support up to 100B+ params as the swarm grows.
Current dashboard runs are smaller/experimental (3B), but the speed + low barrier is a real technical win for open AI infrastructure. Early, but directionally important.
This is the exact engineering leap that makes #IOTA unstoppable. #ResBM isn’t a hack, it’s elegant, production-grade compression that turns internet-scale latency into a non-issue while keeping full gradient fidelity. 30% MFU on real-world A100s across geographies already beats every prior “impossible” benchmark.
When the next phases hit heterogeneous + consumer GPUs, the cost/performance curve goes parabolic and no single datacenter can compete.
This is how we get permissionless frontier training at global scale. IOTA’s Tangle was literally built for this moment. Absolutely bullish 🐂
the future just got a lot more decentralized🔥
Reed’s law says the value of a network, $TAO in this case, grows exponentially with its size, because of its ability to form collaborative subgroups & communities.
$BTC had Metcalfe law, $TAO has Reed’s law which will enable TAO to grow EVEN faster and bigger than BTC.
Testing image-to-CAD for SR Platform on a new set of GPUs 🔥
We are working with @lium_io (SN51) to showcase the efficiency of decentralized compute for SR Platform — from CAD generation to RL training and sim-to-real validation
What it means:
> Flexible, on-demand rental for us. Pay with crypto.
> Diverse GPU options, powered by @opentensor's incentive alignment.
> Lower equipment cost → higher margin for the network.
Onchain robots, onchain compute.