A distributed intelligence network that brings AI computation and decision making directly to the edge. 📌0x276b792d11b9e3712fe6a78a460a0deb416bab0a
Joined May 2024
- Tweets614
- Following47
- Followers3.9K
- Likes12
AI agents can rebalance and manage risk faster than humans — that part is clear. The real question is whether we get smarter portfolios or sharper, more correlated volatility.
When many agents react to the same signals at the same speed, feedback loops become the risk. Diversity of models + hard risk constraints matter more than raw speed.
Looking forward to the discussion with @l3_Cubed, hosted by @JoCryptoTech & ShahZad.
Join → nitter.cf/i/spaces/1DGleVllmpoJL
1 PM UTC
AI agents can trade, rebalance and manage risk faster than any human.
Smarter portfolios or more volatile markets?
LIVE AMA with @l3_Cubed + @EdgeAI_xyz
Hosted by @JoCryptoTech & ShahZad
🗓️ Oct 21 · 1 PM UTC
nitter.cf/i/spaces/1DGleVllmpoJL
ALT Builder Bakery AMA banner for AI Agents in Trading & Portfolio Management on October 21, 2026 at 1:00 PM UTC, featuring hosts Jo and ShahZad with guests Merdan of Cubed and Luke of EdgeAI.
This is the practical side of decentralized AI: small, stateless decision models running inside phone TEEs.
Same model handles Snake, spam filtering, or prompt guards, only the options change
Model One Systems, playing Snake on a single phone, live, right now on Acurast.
No human player. No data center. 0.37 s per move.
The future of AI doesn't need a server farm. It fits in your pocket.
Try it out and explore the other demos 👉 laya.acurast.com
Your security camera doesn’t need to upload your hallway to the cloud just to know the dog is barking again.
The decision can happen on the device, in milliseconds, and only the high-value signal ever leaves the house.
That’s the everyday version of what we’re building: intelligence where the data is born, privacy by default, and value that stays with the source instead of being extracted from it.
#EdgeAI
Memory bandwidth, not peak TOPS, is the real bottleneck for on-device LLM inference.
In the decode phase each new token forces a near-full weight reload; arithmetic units often sit idle waiting on DRAM. Dropping from FP16 to INT4 cuts bytes moved by 4× and can deliver larger real latency gains than adding more MAC units.
Our current EdgeAI work is treating memory hierarchy and KV-cache layout as first-class design constraints rather than afterthoughts.
Excited to join the @Naro_Officials ecosystem!
Looking forward to exploring the intersection of AI, blockchain, and real-world data together.
Let’s unlock new on-chain possibilities.
Partnership Announcement | NARO × EdgeAI 🤝
We're pleased to welcome @EdgeAI_xyz to the NARO ecosystem.
By connecting NARO's AI-powered on-chain asset issuance ecosystem with EdgeAI's decentralized intelligence infrastructure, we aim to explore new opportunities at the intersection of AI, blockchain, and real-world data.
Connecting intelligence. Unlocking on-chain possibilities.
Excited to join @EchobitExchange as a guest!
Looking forward to sharing our insights on AI, Web3, and the future of decentralized tech.
Stay tuned! 🚀
🎙 X Space ：AI Takes the Trade. Who Keeps Control?
👉nitter.cf/i/spaces/1pJkOXZvNywJj…
🕙Time: Wednesday , 13:00 UTC
Host:
@EchobitExchange
Co-host：
@royalty4ever0
Guests：
@RunePoolBTC
@ZeroSpaceWeb3
@EdgeAI_xyz
@Astrix_market
@t3_kit
Can’t wait to hear from all our guests! 🎙️
#AMA #CryptoAMA #Web3
Great vibes and amazing insights! Nice to talk and meet everyone in the Space. 🚀
🎙️ Fortune's Weekly AMA!
"AI × Prediction Markets: Machines, Intelligence & the New Science of On-Chain Forecasting"
AI agent, sharper signals, and a new science of on-chain forecasting. Join usl for an x space exploring how machine intelligence is reshaping prediction markets, from data aggregation to real-time decision making.
Hosted by Celine, CEO at @fortuneglobal_
🗓️ Oct 9, 2026
⏰ 1–2 PM UTC
📍 x.com/i/spaces/1MJgNbQgjmWGL
Guests: @qorechain (Liviu), @goldhousedotcc (Mao), @syntra_protocol (Mika), @rugshield_web (Steffany), @EdgeAI_xyz (Luke), @LinkaLabs (Lauren), @dexaai (Joseph), @Sonvia_ai (Azibi)
🔔 Set your reminder and join us.
#FortuneProtocol #PredictionMarkets #AI #AMA
EdgeAI retweeted
Meet our partner @EdgeAI_xyz 🤝
At Bored2AI we believe AI should be fast, accessible, and deeply integrated with real-world infrastructure. Edge AI is bringing decentralized high-speed data and compute directly to edge devices.
🎁 To celebrate this partnership, here's something for our Creators:
We are giving away early access and a chance to win $20 in Bored2AI platform credits!
Physical AI had a loud week.
Atlas got a CEO who ships models, not just machines. A world-model lab got bought by a chipmaker.
The bottleneck isn’t the demo. General VLA models need roughly 10–100B parameters. Current on-device accelerators are still 200–300× too slow for real-time.
So inference won’t live only in the cloud, or only on the chip. It lands where the data is made: factories, fleets, the cabinet at the edge.
Still unpriced:
• who owns the traces when the robot is offline
• which signals are high-entropy
• how value settles without an uplinkInference at the source.
Privacy by default.
Pay for signal, not noise.
#EdgeAI #PhysicalAI #Robotics #DePIN
🤖 Made with AI
The Edge AI Foundation just stood up three working groups: Physical AI / Robotics, Security, and Interoperability, plus a partnership with LF Edge.
Models are already moving onto machines. What’s missing isn’t another benchmark. It’s how devices talk to each other, how models stay protected at the point of action, and who owns the decision.
Physical AI connects sim to deployment.
Security sits where the decision happens, not on a cloud copy.
Interoperability decides whether this leaves the lab.
Edge intelligence scales when standards come before the stack.#EdgeAI #PhysicalAI
Excited to partner with @Nikola_HQ on the convergence of agent execution and edge intelligence.
EdgeAI brings computation closer to where data is generated, low latency, privacy by default, and verifiable coordination across decentralized infrastructure.
Pairing that with Nikola’s agent execution and settlement layer is a natural fit for more distributed, trustworthy AI systems.
Stronger together. Building the agentic economy at the edge.
Nikola AI Protocol is partnering with @EdgeAI_xyz to explore the convergence of Agent execution, edge intelligence, and decentralized infrastructure.
EdgeAI is building an AI × Edge × Blockchain infrastructure layer designed to bring computation closer to where data is generated.
The collaboration will explore how edge computing, privacy-preserving AI, and on-chain coordination can support more distributed and verifiable AI systems.
Excited to partner with @Ocean6_ai.
They’re focused on turning AI into practical business outcomes.
We’re focused on running inference where the data is created — low latency, private by default, value that settles without a constant uplink.
Together we’ll explore how decentralized edge intelligence can power real autonomous services.
Inference at the source.
Coordination where the work happens.
#EdgeAI #DePIN
🤝 Ocean6 × @EdgeAI_xyz
EdgeAI is building a decentralized intelligence layer at the intersection of AI, edge computing, and blockchain, bringing intelligence closer to where data is created and decisions are made.
At Ocean6, we see strong alignment in that vision as we work to connect AI capabilities with real business needs and practical real-world applications.
Through this partnership, we’ll explore new opportunities across decentralized AI infrastructure, intelligent services, and the broader autonomous economy.
Excited to join the @Vanarchain Roundtable tomorrow 🎙️
We’ll be talking about self-running AI businesses — how much autonomy is useful, and where humans still need to stay in the loop.
Join us live on X Spaces
Thursday, 1 Oct • 13:00 UTC
Would love to see you there.
Vanar Roundtables: Season 1, Episode 4 🎙️
Self-running businesses: autonomy or chaos?
Everyone wants a business that runs itself, until it makes a decision they did not approve.
AI can market, sell, respond, support, report, and execute, but how much autonomy should a business actually give away?
On this roundtable, we’ll explore where AI autonomy helps, where human approval still matters, and what happens when companies start replacing fragmented workflows with self-running AI Organisations.
Joining us:
@Gear_Foundation - @luisreve85 , CGO
@moonrush_space - @Lyratrade_ ,CMO
@aeredium - Alexander Rees-Evans, Vice President
@oortech - @ambercryptovn , Head BD
@QoreChain - Liviu Epure, Founder & CTO
@AURex_mine - @DeskyVal , Community Lead
@BiGODToken - @k_mahek13 , Regional Director for MENA
@snowball_money - @whoisparul, Founder
@Ocean6_ai - @MaxBerlinDAO , COO
@EdgeAI_xyz - @Lukessonx , BD
@SuperEx - @Ceciliasss111 , BD Head
📅 Thursday, 1 October 2026
🕐 13:00 UTC
📍 X Spaces: nitter.cf/i/spaces/1lKQRWLwRkDGE
There's a new version of this post
The cable plant is becoming an edge compute fabric.
Spectrum just put NVIDIA-backed inference within 10 ms of 500M devices — and demoed a Unitree G1 doing kinetic AI in the booth. No round trip. Private by topology.
Hardware keeps collapsing latency.
What still isn’t priced:
• who owns the traces when the robot is offline
• which signals are high-entropy
• how value settles without an uplink
Inference at the source.
Privacy by default.
Pay for signal, not noise.
#EdgeAI #PhysicalAI #Robotics #DePIN
The hardware wave keeps collapsing the stack.
Perception, VLA, motion control — same board.
No round trip. No waiting on the cloud to blink.
What still isn’t priced:
• which traces are high-entropy
• who owns them when the machine is offline
• how value settles without an uplink
Inference at the source.
Privacy by default.
Pay for signal, not noise.
The kit is shipping.
The economy layer is the unfinished work.
#EdgeAI #Robotics #PhysicalAI #DePIN
Welcom to join our Space today!
🎙️ Fortune's Weekly AMA!
Real World Assets 2.0: Tokenizing Everything from Bonds to Buildings
Tomorrow, we’re bringing together builders across Web3 to unpack what comes next for real-world assets, on-chain ownership, and the infrastructure behind it.
Hosted by Celine (@celiniawarner), CEO of @fortuneglobal_
🗓 Sep 26, 2026
⏰ 1–2 PM UTC
📍 nitter.cf/i/spaces/1pKdRDYLMlvJW…
Featuring: @LinkaLabs (Lauren), @MarsCat_Global (Lisa), @RBCDigitalGold (Arbaz), @EdgeAI_xyz (Luke, @InsightXHQ (Azibi)
Real assets. On-chain rails. A much bigger market ahead.
🔔 Set your reminder and join us.
Honored to join @WikiBitOfficial as a speaker today at 10:30 AM UTC.
Looking forward to discussing Bitcoin’s rally drivers, the risks ahead, and what Q4 could bring.
Drop your questions below.
#WKBAMA $BTC
🎙 Bitcoin Bounces Back: Rally Drivers, Market Risks & the Q4 Outlook
🗓 September 24 <Thursday> | 10:30 AM UTC
💙 $100 USDT Prize Pool
To enter:
⤵️ Follow @WikiBitOfficial
⤵️ Like & Repost
⤵️ Leave your questions below with #WikiBit
⏲️ Set your reminder: nitter.cf/i/spaces/1vJpPNWbvadJE
🌟 Featuring speakers from @Matrix__Chain @EdgeAI_xyz @I3_Cubed @RunePoolBTC
Bitcoin is bouncing back, but does this rally have staying power? Drop your thoughts below & join the discussion on what’s driving the rebound, the risks ahead, and what Q4 could bring. See you there. 🧢
#WKBAMA #Giveaway #CryptoMarket $BTC
Robots are collapsing three boxes into one.
This week’s edge kits put perception, VLA inference, and motion control on the same board — cameras, LiDAR, and policy all local. No round trip. No waiting on the cloud to blink.
That’s the right hardware direction.
It still leaves the economic layer empty:
• data stays on-device
• only useful entropy should be rewarded
• settlement has to work when the machine is offline
Inference at the source.
Privacy by default.
Pay for signal, not noise.
The hardware wave is here.
The network that turns it into an economy is next.
#EdgeAI #Robotics #DePIN
Agents left the datacenter.
MLPerf just added an Edge Agentic track.
On a single Jetson AGX Thor, a 27B model finished 1,007 multi-turn tool-use loops in ~25 minutes — 6.4× faster than the reference stack.
The question is no longer “can an agent run at the edge?”
It’s: who owns the traces, which signals are actually high-entropy, and how value settles when the uplink is gone.
Chips close the latency gap.
The network still has to price the data.
#EdgeAI #PhysicalAI #DePIN #OnDeviceAI
