Surya Ranjith retweeted
The energy in AIxGPU Pre summits cannot be any higher, this was in Mumbai today!
Just joined FOUND and got a $58.05 pending reward.
Member #017,278. Claim opens after launch.
Get yours: founds.money/?ref=FND-WPN88
1
12
Surya Ranjith retweeted
We are going live on a 440Kw Datacenter in India which will offer the best Inferencing per dollar, ideal for running open source models at Billions of tokens per day scale.
1/
Feels like we've spent the last two years obsessing over training.
Bigger models. Bigger clusters. Bigger training runs.
Meanwhile, the economics of AI are shifting somewhere else:
Inference. 🧵
2/
Training gets the headlines.
Inference gets the bill.
Frontier-model training costs are measured in the hundreds of millions.
But once a model is deployed, every prompt, every agent action, and every generated token becomes an ongoing expense.
You train once.
You serve forever.
3/
The data is moving fast.
Inference represented roughly 1/3 of AI compute in 2023.
Deloitte projects it will reach roughly 2/3 by 2026.
That's a massive shift in just a few years.
4/
The mistake many teams make?
Treating training and inference like they need the same infrastructure.
Training wants giant centralized clusters.
Inference often benefits from: • Lower latency
• Regional deployment
• Better cost efficiency
• Proximity to users
Different workload. Different priorities.
5/
For production AI, the key metrics aren't just FLOPS anymore.
They're: • Cost per token
• Latency
• Geography
• Power efficiency (PUE)
That's where margins are won or lost.
6/
The challenge is that compute is fragmented across hundreds of providers and regions.
That's where @gpunet comes in.
Instead of manually researching infrastructure, teams can compare providers, discover available capacity, and find infrastructure that fits their inference workload.
The future of AI won't run on a single cloud.
It will run on a global compute fabric.
#AI #GPU #DataCenters #InferenceAI
We have Solved this. We share Transparently which Datacenter we are sourcing from - there is a Reseller agreement signed with them many years in advance.
You go directly or via us, the price remains the same, that is the Trust relationship we have built through the past years. We have been a net Revenue Accelerator for our datacenter partners.
These Datacenters partners are NTT with the longest DC footprint over decades and now setting it ourselves, been here for 4 years now. We will be here till the end of your 5 year lease contract 😘
The GPU brokers have gone crazy.
I got two offers this morning for the same data center and same specs. 20% price difference between them. This has happened multiple times in the last few months.
Pretty common for a market with no price discovery. Too much demand, not enough supply, and suddenly everyone's a broker.
At some point you have to wire $2-10 million to someone you've never met and hope they don't disappear.
Bengaluru AI builders, let’s connect 🧡
GPU x AI Connect is a relaxed evening for founders, CTOs, builders, investors, and teams working across AI and infrastructure.
No forced networking. Just good conversations and the right people in the room.
📅 14 August 2026
⏰ 4 to 6 PM
📍 Open Door Coffee Co., Kothanur
RSVP: luma.com/mjomyli5
Surya Ranjith retweeted
Looking for help from experienced on-chain sleuths.
A friend of mine recently lost 10,702 USDC on Polygon to a scam/hack. We’ve traced the funds through multiple wallets, with part of the trail eventually leading toward Binance-linked addresses.
Victim wallet:
0xfc4Cbff74a700Bb3EaC682061E0794cEB457422A
↓ 10,702 USDC
Scammer / receiving wallet:
0x9b9Ae2db2C1fC598645cA2D954D1acB665b44F62
We’re looking for someone who is genuinely experienced with on-chain investigations and can help trace the funds further and assist with a legitimate recovery path.
The victim is willing to offer up to 20% of successfully recovered funds as a reward to anyone whose help directly leads to recovery.
If you’re an on-chain sleuth or know someone who can help, please DM me. I can share the complete transaction trail privately.
Please RT/share. It might help us get this in front of the right person.
Bengaluru builders, this one’s for you 🧡
AI x GPU connect
A relaxed evening to meet fellow builders, share ideas and connect
Founders. CTOs. AI & ML teams. GPU & cloud builders. Investors. Tech leaders.
📍 Lola’s, Manyata Tech Park, Bengaluru
📅 11 August 2026
⏰ 4:30 PM
Limited seats: luma.com/8kb1rjf7
Everyone talks about tokenizing real estate.
The next major RWA category may be compute infrastructure. GPUs, power, cooling, bandwidth, and uptime are real-world assets with measurable cash flow.
@gpunet transforms idle GPU hardware into productive infrastructure for the AI economy. RWA is physical compute powering real AI workloads.
Explore at rwa.gpu.net 🚀
🤖 Made with AI
Deal #001 is now live.
🔸$1,000 minimum
🔸20% APR
🔸Paid in USDC at the end of term
🔸Target term under 12 months
Link: gpu.net/ganchain/rwa-pool
Due to strong early demand, we've increased the Batch 1 target to $125,000.
Deal #001 opens Wednesday, July 1.
🔸Minimum allocation: $1,000
🔸 Target return: ~20% APR
🔸 Paid in USDT at the end of the term
🔸 Target duration: under 12 months
AI infrastructure should produce cash flow, not only headlines.
If you're new here: @gpunet buys real AI servers, the physical hardware, not tokens, not compute rental.
We deploy and resell that hardware. The resale margin is the return.
More details soon on Deal #001
Why is Tenstorrent a third of the price?
Nvidia's hardware locks you into their software, their networking gear, and water cooling systems. All of that costs money on top of the chip itself.
Tenstorrent skips all three. Open software, regular networking cables, air cooling.
Competitive AI performance, way less to build around it.
Deal #001 runs on that math.
Stay Tuned 👀
A B200 server costs approx $500,000. The same AI compute performance from Tenstorrent costs $150,000
Nobody talks about that gap. We built a deal around it. GPUNET buys the hardware, deploys it, and shares the upside with the people who fund it.
First deal: 10 Tenstorrent Blackhole servers, 320 gpu chips in each server.
Opening soon 🧡
A B200 server costs approx $500,000. The same AI compute performance from Tenstorrent costs $150,000
Nobody talks about that gap. We built a deal around it. GPUNET buys the hardware, deploys it, and shares the upside with the people who fund it.
First deal: 10 Tenstorrent Blackhole servers, 320 gpu chips in each server.
Opening soon 🧡
GM! Quest Season 2 rewards are now live to claim.
You can now go to the platform and claim your rewards.
Thank you for participating 🧡
Claim Link: gpu.net/drops


















