Track market-moving stories before they get noisy

Shiba Inu (SHIB): Trading near $0.00000521, below the 100-day average (~$0.00000530), which now acts as resistance. Buyers defended the $0.0000050 round number, but the bounce is on low volume. Bulls need a daily close above $0.00000530 (ideally $0.00000542) to target the 200-day line (~$0.00000563).Zcash (ZEC): Trading near $1,196, down roughly 30% from its late-September high of about $1,700, and now testing the 50-day average. Monero (XMR): The larger trend remains up, and falling volume suggests sellers aren't pressing. XMR is range-bound between $515 and $545. A close above $545 targets $560 and $590, while a close below $515 points to $483 and then the 100-day average near $472.Dogecoin (DOGE): Trading near $0.0845 after bouncing about 5% from $0.0800, which buyers defended. Price sits below all key averages, RSI is below 50, and bounce volume is modest. Shiba Inu Sellers Keep Control  Shiba Inu is trading near $0.00000521 after a sharp drop on Monday. The price fell under the orange 100-day moving average, which now sits around $0.00000530 and acts as resistance.

Today's candle is small and green, with a long lower wick down to $0.00000502. That shows buyers stepped in near the $0.0000050 round number. Support sits at $0.00000500 first, then at $0.00000475 to $0.00000490, where SHIB bounced in September. Resistance is at $0.00000530 to $0.00000542, then $0.00000563 at the 200-day line.

SHIB/USDT Chart by TradingViewThe bigger picture is mixed. SHIB did not reach the 200-day average (black line, near $0.00000563) in late September and has been moving lower since. The price is also below the blue and light blue averages around $0.00000540, which keeps short-term control with sellers. Volume is low on today's bounce. A rise on weak volume often does not last, so this bounce needs more proof.

HOT Stories

You Might Also Like

For buyers to take back control, SHIB needs a daily close above $0.00000530, and preferably above $0.00000542. That would put the 200-day line back in play, with $0.00000600 as the next big target.

If SHIB closes under $0.0000050, the bounce thesis fails. The next stop would be the $0.00000475 to $0.00000490 zone, and a break of that area could open the way toward $0.00000440, the August low.

What happens next: for now, SHIB is sitting on support with no clear trend. The $0.0000050 level will decide the next move.

Zcash Testing the 50-Day AverageZcash is trading near $1,196 after a steady drop from its late-September high of about $1,700. That is a loss of approximately 30%. The price is now sitting right at the blue 50-day moving average, which is the first real test of this pullback. The rise since August was very strong, so a pullback is normal.

ZEC/USDT Chart by TradingViewSupport is at $1,120 (the recent low), then $1,050 to $1,100, where ZEC paused in September. Deeper support is the orange 100-day average near $985 and the black 200-day average near $760. Resistance is at $1,230, $1,290, then $1,350.

Still, the signals are cooling. The RSI has slipped below 50, which suggests momentum has turned weaker. Volume has also decreased over the past few days, so neither side seems fully in control. Today's candle has a long lower wick down to $1,123, so buyers are protecting this area. But the bounce has stayed below $1,230, and the green 20-day average near $1,290 remains above the price.

You Might Also Like

Bulls need a daily close above $1,230, and then above $1,290, to show the pullback is over. That would open the way back toward $1,350 and higher.

If ZEC closes below $1,120, the 50-day average will have failed. The price could then slide to $1,050, and the 100-day average near $985 would be the next big support.

What happens next: the trend is still up on the larger chart, because all the averages point upward. But the short-term trend is down, and ZEC must hold $1,120 to keep the larger trend safe.

Monero Holding a Sharp TestMonero is trading near $528 and holding above the blue 50-day moving average, which sits around $515. Today's candle is small, with a high of $544 and a low of $519. The price is down about 17% from the September peak near $635. That spike was quick and was sold off fast, so the real top area is closer to $590 to $600.

XMR/USDT Chart by TradingViewSupport is at $515 (the 50-day average), then $483 (the recent low), then $472 (the 100-day average). A few days ago, a sharp wick took the price down to about $483. Buyers pushed it back up quickly, which indicates demand below $500. Resistance is at $540 to $545, then $560, then $590.

The trend is still up on the larger chart. All the major averages point higher, and the 200-day line (black, near $430) is far below. Volume has been falling during this pullback. That is a calm sign, because sellers are not pushing hard. But it also means buyers are not rushing in yet.

You Might Also Like

Bulls need a daily close above $545. That would put $560 in play and could start a new push toward $590.

If XMR closes below $515, the 50-day average will have failed. The price could then return to the $483 low, and the 100-day average near $472 would be the next big floor.

What happens next: for now, XMR is in a range between $515 and $545. A break on either side will set the next direction.

Dogecoin Bounce Off $0.0800 Stalls Dogecoin is trading near $0.0845 after a tough drop last week. Today's green candle has a long lower wick down to $0.0800, so buyers defended the round number. The price has bounced about 5% from the low.

DOGE/USDT Chart by TradingViewSupport is at $0.0800 first, then $0.0780, the September low. Resistance is at $0.0862, $0.0877 to $0.0890, then $0.0928.

The problem is the averages above. DOGE is beneath the orange 100-day line (near $0.0862), the blue 50-day line (near $0.0877) and the green 20-day line (near $0.0890). The 200-day line near $0.0928 is the last big barrier. In late September, DOGE failed to hold above it, and the price has slid since. The RSI is below 50, which indicates weak momentum. Volume on today's bounce is modest, which makes the move less reliable until more buyers show up.

You Might Also Like

DOGE needs a daily close above $0.0862 to show the bounce has strength. A close above $0.0890 would be a bigger win, and it could open the way to the 200-day line near $0.0928. Above that, $0.0960 to $0.1000 is the next area.

If DOGE closes beneath $0.0800, the bounce fails. The price could then test $0.0780. A break of that level could send it toward the $0.0700 area, where it traded in July and August.

What happens next: the $0.0800 floor is holding for now. But with every average above the price, DOGE has a lot of work to do before the trend turns.

Stacks token [STX] was trading around the $0.4182 price area on October 13 after a slight touch of $0.4502. Separately, Stacks Labs had reported second Bitcoin staking round, which was full with 500+ BTC and joined by over 6.7M STX.

The filled round shows participation in a program that pairs Bitcoin with locked STX, but it does not show why its price rose.

What the filled staking round means In its October 12 update, Stacks Labs said most of Bond 2 filled through Stacking DAO, with the Xverse pool also participating. The program hopes to give a return to users of 3% on their bitcoin, paid in BTC, with no guarantees on return.

Through Stacking DAO, users receive stBTC [representing the share of their position], and they can use it in financial apps in Stacks [for example, as collateral to borrow USDCx], and they continue to receive staking rewards.

The distinction between participation routes matters because: 

Approved direct participants keep their Bitcoin locked under their own custody. Open pools use sBTC, Bitcoin-backed asset on Stacks, and are not fully self-custodial Bonding STX limits its availability during the lock period, but those tokens are not permanently removed.

STX retreats from $0.45 The daily chart showed STX up 1.83% from its previous daily close at the time of this writing. It remained about 7.1% below the session high, with the trading still going on.

The latest two candles actually topped out at a higher price but quickly came back down. Hence, the $0.45 and $0.46 area stands out as an obvious hurdle.

Source: TradingView Its trading prices have hovered above the key averages after this rise, supporting the rally. Holding above $0.40 will give another opportunity to attempt the resistance.

If that level gives way, traders could start looking at the $0.37–$0.38 price areas. If it trades steadily above $0.46, the outlook for recovery looks good

Another staking round is approaching Bond 3 is expected to launch around November 10, according to Stacks Labs. Anyone wanting to take part directly must hold at least 50 BTC and be approved, but it can also be done through partner pools according to the requirements set by them.

The next round will show how the level of participation has increased or decreased. For STX’s immediate outlook, it remains to be seen whether buyers can defend the $0.40 price area.

Final Summary Bond 2 filled about 500 BTC of capacity, with more than 6.7 million STX bonded alongside it. STX remained above the $0.40 price level, but its pullback from $0.45 means nearby resistance remained unchallenged.

Fluxers! If you have never deployed anything on Flux, this is the article for you. Getting an application running on the decentralized cloud takes minutes, costs from $0.99 a month, and there is more than one way to do it — so you can pick whichever fits how you already work.

Here are four routes, from the easiest to the most hands-on.

Route one: just say what you want
Open the new FluxCloud preview at cloudpreview.app.runonflux.io and type what you want to run, in plain words:

“A WordPress site for a small bakery.”
“A Minecraft server for ten friends.”
“Postgres with an API in front of it.”

The assistant, FluxAI, works out the resources, fetches a live price from the network, and hands you a ready deployment to review. Common workloads fill the form instantly. Check it, adjust anything you like, then sign and pay with your own wallet or a card — the assistant prepares everything, but only you can approve the spend.

This is the best starting point if you know what you want running but not how to describe it to a cloud.

Route two: push your code with Orbit
If your project already lives in a Git repository, Orbit is the fastest way onto Flux. Paste your repository URL into orbit.runonflux.com, set the port your app listens on, and launch.

Orbit detects your framework from more than 100 — Next.js, Vite, Django, Rails, Laravel, Go, Rust, Spring Boot, .NET and many more — builds it, and deploys it. From then on, every push to your branch redeploys automatically, and a build that fails never replaces the version that works.

No Dockerfile, no server to manage. This blog is deployed exactly this way.

Route three: let your AI agent do it
If you work with Claude Code, Cursor, Windsurf or another AI coding tool, add the Flux Cloud MCP server and ask your agent to deploy for you:

npx @runonflux/flux-cloud-mcp

There is no account to create — two keys are the whole identity, and the server can generate them for you. Your agent can price a deployment before you fund anything, and it only spends after you confirm. Ask it what something would cost first; that part is free.

Route four: the full control panel
For complete control over every setting, use the deployment editor in FluxCloud or the new preview. You describe each component directly:

Image — any public Docker image, or a private one with registry credentials.
Resources — CPU, memory and storage per instance.
Instances — how many copies to run across the network.
Ports and domains — what to expose, and your own domain if you have one.
Storage — local, replicated, or primary/standby so your data survives a node going away.
Private mode — encrypt the whole specification so only ArcaneOS nodes can run it.

Or pick a ready-made template
Many popular applications are a single click. The apps hub covers WordPress, n8n, Hermes and OpenClaw, and the games hub covers nineteen game servers, from Terraria and Minecraft to ARK and Factorio. New accounts get the first week of a game server free.

What it costs
Pricing follows the resources you request: CPU, memory and storage, multiplied by the number of instances, with discounts for small applications and for longer subscriptions. Every deployment has a $0.99 minimum, and prices round to a figure ending in .49 or .99, so quotes are easy to read.

A small web app — half a CPU, 500 MB of memory and 5 GB of storage, on three instances — comes to $0.99 a month. Pay by card, or in FLUX for a 5% discount. You always see the price before you commit.

What happens when you deploy
Your signed specification is registered on the Flux network, and nodes around the world pick it up, pull your image and start your instances. Within minutes your app is reachable at its own address — and from then on it runs on independently operated machines, with no single company in the middle.

Want help along the way? The Ask AI button on the Flux documentation answers questions about every part of this, from resource limits to custom domains.

Start now
The quickest first step: open the FluxCloud preview and type what you want to run. Your first app on the decentralized cloud is a sentence away.

Google: Launches Gemini 4 Argon for global developers and enterprises via Gemini API

OP outpaced bitcoin and ether over matched daily and multiweek windows, though Starknet’s STRK posted a much larger four-week rally.

OP outpaced bitcoin and ether over matched daily and multiweek windows, though Starknet’s STRK posted a much larger four-week rally.

Loading this article…

Free newsletters

Good reporting. Delivered.Choose your coverage: crypto and DeFi, the future of finance, or both.

Free. Unsubscribe anytime.

Injective has a new validator, and it comes with a power grid attached.

Gulf Labs, the blockchain and digital-asset investment division of Thailand’s GULF Group, went live as an institutional validator on the Injective network on October 6, 2026. The move gives Injective a direct line into one of Thailand’s largest corporate ecosystems, and it pushes the network’s Asian expansion further beyond South Korea.

What Gulf Labs is actually doing
On a proof-of-stake network like Injective, validators confirm transactions and take part in governance decisions about how the chain evolves. Gulf Labs has bonded approximately 500,000 INJ to support network security and governance.

The partnership reaches beyond running nodes. Gulf Labs and Injective plan to work together on tokenization, stablecoin initiatives, education and infrastructure access within Thailand’s financial ecosystem.

The stated goal is to connect Injective with key Thai sectors, specifically energy, telecom and banking. Those happen to be the industries where GULF Group already operates.

Financial terms were not disclosed. Neither were specific tokenized products, transaction values or timelines.

Why GULF Group’s footprint matters
GULF Group carries a market capitalization of roughly $27 billion. Its business spans more than 24 GW of energy capacity and a telecommunications network with over 46 million mobile subscribers.

Gulf Edge, another part of the GULF ecosystem, is expected to provide data center, cloud and AI resources to support infrastructure development.

The collaboration will also involve Binance TH, which is regulated by Thailand’s Securities and Exchange Commission.

Injective’s broader push into Asia
Thailand is not Injective’s first institutional stop in the region. The network previously worked with POSCO International and LG CNS in South Korea.

As of mid-October 2026, Injective had tokenized more than $1 billion in real-world assets.

What this means for Injective and Thai institutions
For INJ holders, the most concrete change is on the security side. Approximately 500,000 INJ bonded by an institutional operator adds to the stake protecting the network.

The partnership names tokenization and stablecoins as areas to explore, not products already in market. Until specific launches arrive, the commercial impact remains a forecast rather than a result.

Any Thai stablecoin initiative tied to this partnership would need to fit within the framework overseen by the country’s SEC and other authorities, which is where Binance TH’s regulated status could matter.

The signals worth tracking are specific. Look for announcements of actual tokenized products from Thai institutions, any stablecoin pilots involving Binance TH, and whether Gulf Edge infrastructure begins supporting Injective-based applications. Growth in the network’s real-world asset figure beyond the $1 billion mark would also indicate whether Asian corporate partnerships are translating into on-chain activity.

Disclosure: This article was edited by Estefano Gomez. For more information on how we create and review content, see our Editorial Policy.

How Kabosu, a rescued Shiba Inu from Sakura, Japan, went from a 2010 blog photo to the "Doge" meme and the logo of Dogecoin ($DOGE).

Dogecoin's logo traces back to a single snapshot posted on a personal blog in Japan. The dog in the picture, a Shiba Inu named Kabosu, became one of the best-known faces of the internet and later of the crypto market.

From puppy mill to blog postKabosu lived in Sakura, Japan. Her owner, Atsuko Sato, adopted her in 2008, rescuing her from a puppy mill after the business closed down and Kabosu was abandoned with 19 other Shiba dogs. Sato was a kindergarten teacher.

In 2010, two years after adopting Kabosu, Sato took a picture of her pet crossing her paws on the sofa. She posted the image on her blog, from where it spread to the online forum Reddit and became a meme. Meme-makers added rainbow Comic Sans text over Kabosu's side-eyed look and crossed-paws pose, using short, broken phrases like "much wow" or "such happy." The format became known as "Doge."

How Kabosu became the face of $DOGE Dogecoin was introduced on December 6, 2013. Software engineers Billy Markus and Jackson Palmer launched the satirical cryptocurrency as a way to make fun of Bitcoin and the many other cryptocurrencies boasting grand plans to take over the world. Dogecoin features the face of Kabosu from the "doge" meme as its logo and namesake.

The joke coin outlasted its origins. It quickly developed its own online community, reaching a peak market capitalization of over US$85 billion on May 5, 2021. NPR reported that the coin jumped in value years later after Elon Musk, then the world's richest person and a champion of Dogecoin, tweeted about it. In 2023, @elonmusk briefly swapped Twitter's bird logo for the Doge image.

Kabosu's fame also reached beyond crypto. The picture later became an NFT digital artwork that sold for $4 million. Sakura has used the doge image on manhole covers, and a bronze statue of Kabosu was unveiled in 2024, paid for by donations from fans around the world.

Sato said Kabosu died on Friday, May 24, 2024. As a rescue dog, Kabosu's real birthday was unknown, but Sato estimated her age at 18. The @dogecoin account also marked her death, calling her an inspiration.

The story is a reminder of how a casual pet photo, shared without any commercial intent, can end up anchoring a market with a multibillion-dollar peak valuation.

SourcesRolling Stone: Kabosu, Shiba Inu Who Became Face of 'Doge' Meme, Dead at 18

NPR: The Shiba Inu that inspired Dogecoin crypto and countless memes has died

Gulf News (AFP): Japanese dog Kabosu, the face of Dogecoin, dies at 18

The last Aptos team and investor tokens are out

Aptos has quietly put together one of the better months in large-cap crypto. According to The Defiant, the APT token gained 51% over four weeks, a run that left both Bitcoin and Ether trailing.

The rally has a catch. Stablecoin liquidity on the Aptos network has been heading the opposite direction, which raises a fair question about what is actually fueling the move.

The numbers behind the run Price trackers don’t fully agree on the size of the gain. The Defiant puts it at 51%, while other tracking reports show gains of 38.5% to 44% for the four weeks ending October 13, 2026.

Whichever figure you pick, APT beat the two largest crypto assets over the same window. The token traded in a range of roughly $0.82 to $0.85 as of mid-October.

Zoom in to the past seven days and the picture gets less exciting. APT’s weekly return has been nearly flat, which suggests most of the heavy lifting happened earlier in the month.

The less flattering number comes from DefiLlama. The stablecoin market cap it tracks on Aptos fell approximately 17.5% over four weeks, landing at about $910 million.

Stablecoins are the working capital of any blockchain economy. A shrinking pool generally means less dry powder sitting inside the ecosystem, ready to be deployed.

So the token got more expensive while the money sitting on the network got thinner. That divergence is the main tension in this story.

The unlock cliff, now in the rearview mirror The most concrete catalyst landed on October 12, 2026. That date marked the completion of Aptos’s four-year token unlock schedule for early contributors and investors.

With the schedule finished, APT’s monthly emissions dropped by around 60%. Releases fell from approximately 11.3 million tokens per month to roughly 4.5 million.

It also helps explain the timing. A rally into a known supply reduction is a classic pattern, as traders position ahead of the event rather than after it.

A busy stretch for Aptos Labs The unlock was not the only development. Around October 9, 2026, Aptos Labs and the Aptos Foundation announced more than $50 million in commitments aimed at on-chain trading infrastructure and AI-related projects.

On the technical side, Aptos launched an optional encrypted mempool on mainnet on October 12-13, 2026. The feature is designed to protect users from MEV, or maximal extractable value. MEV refers to profits that bots and validators can extract by seeing pending transactions and reordering or front-running them. Encrypting the waiting room means outsiders can’t read your order before it executes.

There is also a proposal on the table for a new execution engine called MonoMove, aimed at improving smart contract performance by up to 55 times. That is a proposal, not a shipped product.

What this means for APT holders and the network The bear case lives in the DefiLlama data. A 17.5% slide in on-chain stablecoins points to liquidity leaving, not arriving. DeFi total value locked on Aptos remains low, which reinforces the concern that price has moved ahead of usage.

The $50 million in commitments toward trading infrastructure looks partly like a response to that exact problem. The encrypted mempool fits the same strategy, since MEV protection is a selling point for active traders choosing where to deploy capital.

The stablecoin figure on DefiLlama is the number to watch. If it stabilizes or recovers while APT holds its gains, the rally has firmer footing. If it keeps falling, the price move starts to look more like a supply-driven trade than a vote of confidence in the ecosystem.

Disclosure: This article was edited by Vivian Nguyen. For more information on how we create and review content, see our Editorial Policy.

Aptos has activated an optional encrypted mempool on its mainnet, offering users the ability to conceal a transaction’s details until a block is finalized. This move aims to improve security and protect traders from front-running strategies, a common concern for network participants managing sensitive transactions.

Aptos introduces encrypted mempoolAptos is a Layer 1 blockchain designed to deliver fast transactions and robust scalability through its unique Move programming language. The network’s newly added feature, based on Aptos Improvement Proposal 144 (AIP-144), allows users to hide which functions their transactions call, along with the inputs involved, until final execution. Transaction senders’ addresses and associated fees, however, remain visible.

The encrypted mempool solution was designed by Aptos Labs cryptographer Rex Fernando. It secures transaction information using a key generated collaboratively by network validators. Each round of network operation produces a new key, which then decrypts transactions only when they are ready to be processed in a block.

Aptos implemented a batch processing technique to keep operation costs low. Validators can unlock all sealed transactions within a block through a single, fixed-size step, regardless of the number of transactions. According to AIP-144, this approach is expected to support over 1,000 transactions per second at launch.

Testing of the encrypted mempool with 50 nodes on Aptos’ consensus system showed an average increase of about 20 milliseconds in block time, while keeping total completion well under one second. A research paper titled “TrX,” submitted to the ACM Conference on Computer and Communications Security (CCS), measured a 27-millisecond increase in the time from transaction proposal to execution—representing roughly a 14% delay compared to conventional processing.

Mini dictionary: Mempool — A “memory pool” that temporarily holds transactions on a blockchain network before they are included in a block. Encrypted mempools shield transaction details from public view until execution, reducing the risk of front-running.

Industry response and Ethereum’s approachEthereum has also considered similar solutions to limit maximal extractable value (MEV) exploits, but its proposals remain under review. The Encrypt the Mempool Coalition, spearheaded by Shutter Network with support from Fairblock and Nillion, has promoted EIP-8184, known as LUCID. This design enables users to commit to a transaction and reveal its contents at a later stage. EIP-8184 was drafted in March 2026 by researchers Anders Elowsson, Justin Florentine, and Julian Ma.

Additionally, Ethereum co-founder Vitalik Buterin presented the concept of a recursive STARK mempool with EIP-8288, a proposal put forward in September. Both options reflect growing interest across major networks in encrypted transaction solutions, though they have not yet been deployed on Ethereum’s mainnet.

The Petra wallet has become the first to integrate Aptos’ encrypted mempool option directly into its transaction signing process. Decibel, a protocol focused on perpetuals trading, also plans to incorporate the feature. Developers have had access to the encrypted mempool via the TypeScript SDK since version 7.0.0.

APT token unlock and supply changesAptos introduced the encrypted mempool as its four-year token unlock schedule concluded. Since launching its mainnet in October 2022, Aptos distributed tokens to core contributors and early backers under a vesting schedule. On October 12, 2026, this initial insider allocation was completed, leaving no new genesis tokens to vest, the Aptos team confirmed on X.

Aptos has ended its genesis token vesting for insiders, with no further unlocks scheduled for core contributors or early investors. The network now shifts focus to community and foundation pools for ongoing distribution.

Following the vesting milestone, Aptos reduced its monthly release of new APT tokens by approximately 60%, dropping from 11.31 million to 4.54 million. Over the next years, the other 4.54 million tokens each month will be allocated to Community and Foundation pools through 2032.

Staking rewards, currently yielding up to 2.6% annually, have become the primary source of new APT tokens, generating around 19.8 million tokens per year. As of September 30, the circulating supply of APT stood at 871.02 million tokens.

MilestoneAmount (APT)NotesPrevious monthly release11.31 millionBefore unlock schedule endedNew monthly release4.54 millionAfter unlock completionAnnual staking rewards19.8 millionEstimated at 2.6% APYCurrent circulating supply (as of Sep 30)871.02 millionReported figure

Robinhood Chain hits a $100B trading milestone

Kaspa's new token standard enters last phase...

In This Article

DATA Network Is Shutting Down: This Is an Absorption, Not a PartnershipThe Supply Overhang the News Cannot FixClaude AI Predicts SUI Levels: Where the Dollar Line Actually Sits

The Sui Foundation has paid $30 million to bring an entire blockchain onto its network, and the market answered by selling SUI back toward $1. Claude AI predicts SUI defends the $1.0017 level through this pullback, because the deal changes what runs on the chain rather than what the token is priced against.

SUI trades at $1.0539, giving it a market cap of $4.344 billion. The 24 hour range tells the more useful story, running from $0.9966 to $1.09, which means the dollar handle has already broken once today and been reclaimed within the session.

EXCLUSIVE: Earn $50 With EdgeX and Enter $300K Prize Draw

DATA Network Is Shutting Down: This Is an Absorption, Not a Partnership
DATA Network is shutting down its own blockchain. Its applications move to Sui and its tokens migrate one to one. The Sui Foundation becomes its largest institutional investor in the process.

What comes with it is measurable. The protocol tracks provenance for AI training data and arrives carrying more than 250 million registered contributions.

The detail that matters most sits at the end of the announcement. Walrus Protocol becomes the default storage layer.

Provenance records are not a one time upload. Every new contribution writes data that has to be stored and retrieved, which means recurring demand rather than a single integration event.

That is the difference between a chain gaining a user and a chain gaining a workload.

Whether it moves SUI is a separate question. The migration is built to raise activity over quarters, and the market prices tokens over days.

DISCOVER: Best Meme Coin ICOs to Invest in 2026

The Supply Overhang the News Cannot Fix
 
Here is why good news keeps failing to stick. SUI has 4.118 billion tokens circulating against a maximum supply of 10 billion. Market cap is $4.344 billion while fully diluted valuation reaches $10.548 billion.

More than half the eventual supply has not reached the market yet. Every rally is priced against tokens that do not exist today but will. The performance table shows the same strain. SUI is up 47.2% over 30 days but down 11.7% over seven and 10% over 14. The September advance is unwinding faster than it built.

Over twelve months the token is down 63%.

Leverage adds to it. Perpetual open interest stands at $1.373 billion against a $4.344 billion market cap, close to a third. That ratio turns ordinary selling into sharp moves, which is how a token touches $0.9966 on a day with genuinely positive news.

Spot turnover reached $670.189 million, healthy enough that this is participation rather than a thin market drifting.

DISCOVER: Aptos Makes Encrypted Mempool Optional on Mainnet

Claude AI Predicts SUI Levels: Where the Dollar Line Actually Sits
SUIUSDT Chart 1D TradingView
The round number is close to a real level, which is rare. AI predicts SUI will be decided at these levels:

The floor: $1.0017. Support through most of 2026 and already tested today at $0.9966. Daily closes above it keep the September structure alive.
The ceiling: $1.4180. The level that capped the May rally and stopped the September push. Roughly 35% above spot.
The downside case: $0.6759. The summer base. Losing $1.0017 removes the next real support until here.

Watch the daily close rather than the wick. SUI dipped under a dollar and recovered within the session, which is a test rather than a break.

A confirmed close below $1.0017 would put the entire 47% monthly gain at risk, since nothing structural sits between there and $0.6759.

The $30 million buys Sui a workload and a storage customer. It does not buy the token a floor, and this week the token is being traded on supply and macro rather than on what is being built.

DISCOVER: 15+ Upcoming Coinbase Listings to Watch in 2026

Follow 99Bitcoins on X For the Latest Market Updates and Subscribe on YouTube For Daily Expert Market Analysis.

#Altcoin News Today

Why you can trust 99Bitcoins

10+ Years

Established in 2013, 99Bitcoin’s team members have been crypto experts since Bitcoin’s Early days.

90hr+

Weekly Research

100k+

Monthly readers

50+

Expert contributors

2000+

Crypto Projects Reviewed

Follow 99Bitcoins on your Google News Feed

Get the latest updates, trends, and insights delivered straight to your fingertips. Subscribe now!

Subscribe now

Alex Ioannou

On-Chain Journalist

Alex is a seasoned cryptocurrency trader and market analyst with over seven years of active experience in the digital asset space. Since entering the markets in 2017, Alex has specialized in identifying emerging "meta" trends and high-volatility narratives. Notably, Alex...
Read More

Read Alex Ioannou Articles

Free Bitcoin Crash Course

Enjoyed by over 100,000 students.

One email a day, 7 days in a row.

Short and educational, guaranteed!

Not financial or tax advice. PANews content is strictly educational and informational and is not investment advice, financial advice, tax advice, legal advice, or a solicitation to buy or sell any digital asset, security, or financial product. Do your own research and consult qualified advisers.

Disclosure. PANews may publish sponsored content, partner content, advertisements, affiliate links, event promotions, and market commentary involving Web3 projects, service providers, or financial products. PANews personnel, contributors, or affiliates may hold digital assets or other interests related to covered topics. See our Terms of Service.

This site is protected by reCAPTCHA.

ARK Invest has graded its own crypto homework from Big Ideas 2026. The verdict: a rough year for prices, but no change to the long-term thesis.

The firm still projects that the digital asset market will reach $28 trillion by 2030. It still sees Bitcoin taking approximately 70% of it, and its 2030 price target remains around $1 million per coin.

What ARK is standing behind The review, which also looks ahead to ARK’s outlook for 2027, is led by Digital Assets Director Lorenzo Valente and CEO Cathie Wood. Their core argument rests on two jobs they believe Bitcoin performs.

The first is digital gold: a scarce asset people hold to preserve wealth. The second is permissionless money: value that can move without a bank or government signing off.

As of mid-October 2026, ARK says it is judging Bitcoin’s performance over a five-year horizon centered on those two roles.

Wood reinforced that framing in September 2026. She reaffirmed that Bitcoin’s primary functions remain intact, even as stablecoins expand what they can be used for.

The numbers ARK adjusted, and the one it didn’t ARK’s model has not stood still. Two adjustments pull in opposite directions, and the net effect is roughly a wash on the headline target.

On the upside, the firm says the total addressable market for digital gold grew by 37%. That expansion followed a 64.5% rise in gold’s market capitalization in 2025.

On the downside, ARK cut its estimate of Bitcoin’s safe-haven penetration in emerging markets by 80%. The reason: stablecoins are spreading rapidly in those regions.

ARK acknowledged that shift in its model. It just did not let it move the 2030 Bitcoin target, which stays at $1 million.

Stablecoins as on-ramp, not replacement ARK expects stablecoin adoption to accelerate the move toward digital money broadly, without fully displacing Bitcoin.

The review also points to other forces shaping the market. Those include the institutionalization of Bitcoin, continued work in decentralized finance (DeFi), and the tokenization of real-world assets such as bonds or property.

Background: a forecast built for the long haul ARK’s Big Ideas report is the firm’s annual roadmap for the technologies it believes will define the coming years. The 2026 edition set out the $28 trillion digital asset market forecast for 2030. It placed Bitcoin at approximately 70% of that total.

ARK has framed its adjustments, including the stablecoin factor in emerging markets, as refinements rather than reversals. The firm says the structural growth of crypto and blockchain technology remains central to how it sees future financial markets developing.

What this means for investors The 80% cut to emerging market safe-haven penetration is a real concession. It shows ARK recognizes that stablecoins are winning a use case once claimed for Bitcoin.

The digital gold half of the thesis leans heavily on gold itself. A 37% bigger addressable market is helpful only if Bitcoin actually captures it, and only if gold’s 2025 surge holds rather than reverses.

Reaching $1 million by 2030 requires the digital asset market to grow into ARK’s $28 trillion projection and Bitcoin to hold its expected dominance. Both are forecasts, not facts, and ARK’s own model shows how individual assumptions can swing sharply.

Disclosure: This article was edited by Vivian Nguyen. For more information on how we create and review content, see our Editorial Policy.

THEO, which operates on the Base network, experienced one of its sharpest declines in history following suspicious token mining. According to data shared by blockchain security company PeckShield, the attacker generated 19 million THEO on the Base network.

The generated tokens were then sold on the market, converting them into 700.8 ETH. At the prices during the transaction, the total value of the Ethereum was estimated to be approximately $1.76 million. It was reported that the attacker transferred the obtained ETH from the Base network to the Ethereum mainnet.

Following the selling pressure, the price of THEO has fallen by approximately 92% in the last 24 hours. The influx of permissionless tokens into circulation has led to a rapid dilution of existing investors’ assets and a significant drain on market liquidity.

While initial data suggests the incident was an attack, it is not yet clear which contract the vulnerability originated from, whether private keys were compromised, and whether the total loss is limited to $1.76 million.

The chart shows the price drop of THEO. *This is not investment advice.

Follow our Telegram and X account now for exclusive news, analytics and on-chain data!

Celestia (TIA) has climbed more than 30% over the past week, reaching its highest level of 2026 as trading volume and open interest rise without a clear single catalyst.

Celestia's native token, $TIA, is rebounding sharply. The asset has gained more than any other major cryptocurrency over the past week apart from @Starknet's $STRK, rising about 34.4% at the time of writing.

The exact trigger is not clear. The official @celestia account has not posted any announcements for nearly two weeks, leaving traders to look to market data and network developments for explanations.

What the data showsAccording to TokenPost, TIA rallied 23% to $0.633 on October 11, its highest level of 2026. The report said open interest rose 45% over 24 hours, while daily trading volume increased 213%. The move came as momentum across the wider altcoin market weakened.

The rally also followed a recovery from lows. TokenPost noted that the token had fallen toward $0.43 in late September before rebounding in October. AMBCrypto said TIA is up 62% over the past month, and that a daily close above the $0.625 swing high would confirm a structural reset, according to its technical analysis.

Possible catalysts and caveatsSeveral developments sit around the move. CoinMarketCap's news summary lists a 3.07 terabits per second benchmark for Celestia Fibre on October 2 and the decision by onchain options exchange Derive to adopt Celestia for data availability for its V3 upgrade on October 1. The same summary notes that mainnet usage of Fibre remains minimal. TokenPost also linked the rally to the Sustainable Blob Economy proposal, described as an economic upgrade for the network.

Analysts remain divided. CoinMarketCap's summary describes social sentiment as a contest between breakout optimism and skepticism over weak onchain fundamentals and low protocol revenue. Traders Union has also flagged a token unlock scheduled for October 30 as a potential source of future supply pressure.

Celestia is a modular blockchain that provides data availability for rollups. Price moves of this size have occurred before without a clear announcement, and the token remains well below its all-time high above $20, according to Kraken's price data.

Sources:

TokenPost: Celestia's TIA Rallies 23% as Price Tests January High
AMBCrypto: Can Celestia hold its 23% rally as altcoins weaken?
CoinMarketCap: Latest Celestia (TIA) news

Starknet launches shadow accounts, enabling users to anonymously link their main wallets with DeFi activities.

According to an official announcement, Starknet has launched Shadow Accounts, allowing users to participate in public DeFi activities via an independent, self-controlled address while keeping the link between this address and their main wallet confidential. Shadow Accounts are independent pseudonymous addresses on Starknet, funded by users’ protected STRK20 balances. The first integration will launch tomorrow. While the balance, transaction history, and application positions of Shadow Accounts remain publicly visible on-chain, STRK20 protects the connection between these accounts and users’ main wallets. This means users do not need to hide their DeFi activities themselves, while reducing the likelihood that external observers can trace those activities back to their main wallets.

2 hours ago

Applied Materials forecasts its services business revenue will grow by more than 20% by 2026, and top $12 billion by 2030.

Applied Materials (AMAT.O) said it targets a non-GAAP gross margin of 54% and operating margin of 40% by 2030. It projects its services business revenue will grow by over 20% in 2026, and exceed $12 billion by 2030. The company plans to expand its R&D labs and manufacturing capacity to more than double current levels by 2030, return 80% to 100% of its free cash flow to shareholders by 2030, and lift its wafer fabrication equipment (WFE) market share by more than 2 percentage points between 2025 and 2030.

2 hours ago

Perspective: Bitcoin investors who accumulated positions at high levels have their cost prices capping upside potential, with selling pressure having notably eased.

Glassnode stated that the cost basis of investors who purchased Bitcoin at high levels continues to cap its upside potential. Following Bitcoin’s rebound to around $88,000, investors with 1 to 2-year holding periods sold off heavily, and this price level is also near the break-even point for investors who have held Bitcoin for 6 to 12 months. Bitcoin’s price subsequently faced resistance below this level, and this selling pressure has now eased significantly.

2 hours ago

21Shares Launches Synthetic Anthropic ETP on the Vienna Stock Exchange

21Shares has launched a synthetic Anthropic ETP on the Vienna Stock Exchange, trading under the ticker CLAU, providing investors with non-leveraged exposure linked to the pre-IPO implied valuation of the artificial intelligence startup. The ETP does not grant investors direct ownership of Anthropic shares, either before or after the firm goes public. The product accesses this exposure via total return swaps (TRS), tracking a reference rate compiled by Kaiko, which is calculated based on Anthropic’s perpetual futures prices on Binance.

2 hours ago

Ripple partners with Stripe-owned Privy to introduce embedded wallet functionality for the XRP Ledger.

Ripple has partnered with Stripe-owned wallet infrastructure firm Privy to integrate support for the XRP Ledger (XRPL) into Privy’s developer platform. The integration eliminates the need for developers to build their own wallet infrastructure, allowing them to add XRPL wallets to their applications with features including wallet creation and management, transaction signing, plus support for XRP, Ripple’s RLUSD stablecoin, and other assets on the network.

2 hours ago

TL;DR: ZetaChain and Verona are running the first study of what people want AI to remember, and what it should never keep. More than 300,000 people have joined Anuma, each with a private AI memory. Verona proves facts about you without exposing the data. The survey takes five minutes: take it here.

ZetaChain is building the Private Memory Layer for AI: user-controlled memory, identity, permissions, and payments that work across AI models, apps, and agents. We believe what your AI knows about you should belong to you, with keys only you hold.

Today, we're excited to announce a research partnership with Verona, the intelligence layer for AI used by 150+ brands including Uber and Nike, and backed by $36M from Multicoin and Circle. Together we are running the first study of what people actually want AI to remember, starting with the more than 300,000 people who have joined Anuma since February. We are also exploring how verified context and private memory can work together in one product.

Memory versus verified contextZetaChain remembers what you tell it. Verona proves what's true about you.

Memory is what an AI keeps: encrypted, portable across models, and under your control. Verified context is what an AI can prove: a fact checked once at its source, with a proof you own that any agent you authorize can reuse without seeing the data.

An AI that remembers you is useful. An AI that can prove what it knows about you is trustworthy. Almost nobody is studying the gap between them.

300,000 people with private AI memoryAnuma launched in February. Since then, more than 320,000 accounts have been created, each one a self-custodial wallet, and Anuma users have made more than 1.3 million AI requests across 38 models spanning chat, image, and video. The live numbers are published openly on ZetaChain Research.

Every one of those accounts has an encrypted memory vault that only the user's key opens. We cannot read it, and neither can the model providers. Your writing style, preferences, projects, and files follow you from one model to the next without starting over, and you decide what stays. Since late September that same memory also powers Nearby, which introduces Anuma users to people near them worth meeting. What an AI remembers now shapes who you meet, not only what it answers.

Why VeronaVerona makes user-verified data portable, private, and programmable. Its Truth Engine uses zero-knowledge proofs across websites, email, passports, and apps to turn a claim into a provable fact that the user owns through a Meta Account and any agent they authorize can reuse. Verify once, reuse everywhere, expose nothing.

More than 150 brands, including Uber and Nike, already build on it, and Verona is backed by $36M from investors including Multicoin and Circle.

Verona is the natural partner for this question because it sits on the other side of the line we want to study. We hold what an AI remembers. Verona holds what an AI can prove. The research asks whether people see that line at all, and where they want it drawn.

The research: What People Actually Want AI to RememberThe question is simple to ask and hard to answer. What do people want an AI to remember about them, and what should it never be allowed to keep? Underneath it sit the questions that decide whether AI memory is trusted: who decides, how you find out what is stored, and what happens when you revoke it. And now that an AI can introduce people to each other, a newer question sits beside them: what would you let an AI share about you to be discoverable, and what should it keep private?

The study covers six themes:

What people want AI to remember

What AI should never retain

Memory portability between models and apps

User visibility, permission, and revocation

The difference between remembered information and verified context

What people would let an AI use to introduce them to someone, and what stays private

We are asking Anuma users, Verona's community, and the broader AI user community. ZetaChain drafted the scope and questions from what we see in Anuma, and both teams reviewed them. Both teams will review the findings together and publish a report in November. Responses are anonymous and reported only in aggregate, and the full question set will be published with the report.

Take partThe survey is anonymous, needs no account, and takes about five minutes. It closes on October 27. Take it here: Take the survey.

Follow @ZetaChain and @verona_dev for the findings, and join the discussion on Discord.

What comes nextThe survey opens today and the report publishes in November, followed by an X Space with both teams and a series of data cards. We are also exploring how Verona's verified context and ZetaChain's private memory work together, so an agent can act on a fact about you without ever holding the data underneath.

About VeronaVerona is the intelligence layer for AI: a network where facts about the real world are proven once, owned by the user, and reusable by any agent they authorize, with the data underneath never exposed. Verona is used by 150+ brands including Uber and Nike and backed by $36M from investors including Multicoin and Circle. Learn more at verona.dev and in the Verona docs.

About ZetaChainZetaChain is building the application layer for private AI: user-controlled memory, identity, permissions, and payments that work across AI models, apps, and agents. Its first consumer product is Anuma, a private multi-model AI app with one memory across every model, joined by more than 300,000 people since February. The next phase is bringing that application layer to Solana and opening it to a broader ecosystem of AI apps and agents.

Follow ZetaChain on X (Twitter) and join the conversation on Discord and Telegram.

Categories

According to an official announcement, Starknet has launched Shadow Accounts, allowing users to participate in public DeFi activities via an independent, self-controlled address while keeping the link between this address and their main wallet confidential. Shadow Accounts are independent pseudonymous addresses on Starknet, funded by users’ protected STRK20 balances. The first integration will launch tomorrow. While the balance, transaction history, and application positions of Shadow Accounts remain publicly visible on-chain, STRK20 protects the connection between these accounts and users’ main wallets. This means users do not need to hide their DeFi activities themselves, while reducing the likelihood that external observers can trace those activities back to their main wallets.

Relevant content

Applied Materials forecasts its services business revenue will grow by more than 20% by 2026, and top $12 billion by 2030.

Applied Materials (AMAT.O) said it targets a non-GAAP gross margin of 54% and operating margin of 40% by 2030. It projects its services business revenue will grow by over 20% in 2026, and exceed $12 billion by 2030. The company plans to expand its R&D labs and manufacturing capacity to more than double current levels by 2030, return 80% to 100% of its free cash flow to shareholders by 2030, and lift its wafer fabrication equipment (WFE) market share by more than 2 percentage points between 2025 and 2030.

2 hours ago

Perspective: Bitcoin investors who accumulated positions at high levels have their cost prices capping upside potential, with selling pressure having notably eased.

Glassnode stated that the cost basis of investors who purchased Bitcoin at high levels continues to cap its upside potential. Following Bitcoin’s rebound to around $88,000, investors with 1 to 2-year holding periods sold off heavily, and this price level is also near the break-even point for investors who have held Bitcoin for 6 to 12 months. Bitcoin’s price subsequently faced resistance below this level, and this selling pressure has now eased significantly.

2 hours ago

21Shares Launches Synthetic Anthropic ETP on the Vienna Stock Exchange

21Shares has launched a synthetic Anthropic ETP on the Vienna Stock Exchange, trading under the ticker CLAU, providing investors with non-leveraged exposure linked to the pre-IPO implied valuation of the artificial intelligence startup. The ETP does not grant investors direct ownership of Anthropic shares, either before or after the firm goes public. The product accesses this exposure via total return swaps (TRS), tracking a reference rate compiled by Kaiko, which is calculated based on Anthropic’s perpetual futures prices on Binance.

2 hours ago

Google: Launches Gemini 4 Argon for global developers and enterprises via Gemini API

Google (GOOG.O) announced the launch of Gemini 4 Argon to developers and enterprises worldwide via the Gemini API. Gemini 4 Argon is accessible through the paid Gemini API and Google AI Ultra subscription service.

2 hours ago

Ripple partners with Stripe-owned Privy to introduce embedded wallet functionality for the XRP Ledger.

Ripple has partnered with Stripe-owned wallet infrastructure firm Privy to integrate support for the XRP Ledger (XRPL) into Privy’s developer platform. The integration eliminates the need for developers to build their own wallet infrastructure, allowing them to add XRPL wallets to their applications with features including wallet creation and management, transaction signing, plus support for XRP, Ripple’s RLUSD stablecoin, and other assets on the network.

2 hours ago

NVIDIA will announce a $2 billion artificial intelligence investment in Europe this week.

NVIDIA will announce a $2 billion artificial intelligence investment in Europe this week.

2 hours ago

Not financial or tax advice. PANews content is strictly educational and informational and is not investment advice, financial advice, tax advice, legal advice, or a solicitation to buy or sell any digital asset, security, or financial product. Do your own research and consult qualified advisers.

Disclosure. PANews may publish sponsored content, partner content, advertisements, affiliate links, event promotions, and market commentary involving Web3 projects, service providers, or financial products. PANews personnel, contributors, or affiliates may hold digital assets or other interests related to covered topics. See our Terms of Service.

This site is protected by reCAPTCHA.

Blur will shut down its website and halt new NFT loans on November 12, just ten days after sister project Blast announced its closure.

Blur to Shut Down Weeks After Blast Announces Closure
Once the dominant NFT marketplace on Ethereum, Blur is shutting down its website and NFT lending services just ten days after its sister project Blast announced its own closure.

What's the Scoop?
Blur Shutting Down: In an announcement Monday, Blur said it will retire its website on November 12, though its underlying NFT marketplace will remain operational onchain. The BLUR token will also retain control of the protocol's fee switch, which determines whether the marketplace can charge trading fees.

NFT Lending Ending: Blur's NFT lending platform, Blend, will stop matching new loans on November 12. Existing borrowers must repay their loans by then or allow them to default, with Blur promising a separate tool afterward to help users withdraw their remaining assets.

Trading Volume Collapses: Launched in 2022, Blur quickly overtook OpenSea, capturing over 70% of Ethereum NFT trading volume in March 2023, when it processed ~$1.43B in trades. By March 2026, that figure had fallen below $7M, a collapse of over 99.5%. Despite billions in trading volume, its zero-fee marketplace generated no trading-fee revenue.

Blast Shutdown: The news comes just ten days after the shutdown of Ethereum L2 Blast, also run by Blur's founder Pacman, citing operating costs exceeding revenue. The network, which once attracted over $2.3B in deposits, has given users until October 26 to withdraw through its regular interface.

David Christopher

733 posts

David is a writer/analyst at Bankless. Prior to joining Bankless, he worked for a series of early-stage crypto startups and on grants from the Ethereum, Solana, and Urbit Foundations. He graduated from Skidmore College in New York. He currently lives in the Midwest and enjoys NFTs, but no longer participates in them.

Starknet launches shadow accounts, enabling users to anonymously link their main wallets with DeFi activities.

According to an official announcement, Starknet has launched Shadow Accounts, allowing users to participate in public DeFi activities via an independent, self-controlled address while keeping the link between this address and their main wallet confidential. Shadow Accounts are independent pseudonymous addresses on Starknet, funded by users’ protected STRK20 balances. The first integration will launch tomorrow. While the balance, transaction history, and application positions of Shadow Accounts remain publicly visible on-chain, STRK20 protects the connection between these accounts and users’ main wallets. This means users do not need to hide their DeFi activities themselves, while reducing the likelihood that external observers can trace those activities back to their main wallets.

1 hours ago

Applied Materials forecasts its services business revenue will grow by more than 20% by 2026, and top $12 billion by 2030.

Applied Materials (AMAT.O) said it targets a non-GAAP gross margin of 54% and operating margin of 40% by 2030. It projects its services business revenue will grow by over 20% in 2026, and exceed $12 billion by 2030. The company plans to expand its R&D labs and manufacturing capacity to more than double current levels by 2030, return 80% to 100% of its free cash flow to shareholders by 2030, and lift its wafer fabrication equipment (WFE) market share by more than 2 percentage points between 2025 and 2030.

1 hours ago

Perspective: Bitcoin investors who accumulated positions at high levels have their cost prices capping upside potential, with selling pressure having notably eased.

Glassnode stated that the cost basis of investors who purchased Bitcoin at high levels continues to cap its upside potential. Following Bitcoin’s rebound to around $88,000, investors with 1 to 2-year holding periods sold off heavily, and this price level is also near the break-even point for investors who have held Bitcoin for 6 to 12 months. Bitcoin’s price subsequently faced resistance below this level, and this selling pressure has now eased significantly.

1 hours ago

21Shares Launches Synthetic Anthropic ETP on the Vienna Stock Exchange

21Shares has launched a synthetic Anthropic ETP on the Vienna Stock Exchange, trading under the ticker CLAU, providing investors with non-leveraged exposure linked to the pre-IPO implied valuation of the artificial intelligence startup. The ETP does not grant investors direct ownership of Anthropic shares, either before or after the firm goes public. The product accesses this exposure via total return swaps (TRS), tracking a reference rate compiled by Kaiko, which is calculated based on Anthropic’s perpetual futures prices on Binance.

1 hours ago

Ripple partners with Stripe-owned Privy to introduce embedded wallet functionality for the XRP Ledger.

Ripple has partnered with Stripe-owned wallet infrastructure firm Privy to integrate support for the XRP Ledger (XRPL) into Privy’s developer platform. The integration eliminates the need for developers to build their own wallet infrastructure, allowing them to add XRPL wallets to their applications with features including wallet creation and management, transaction signing, plus support for XRP, Ripple’s RLUSD stablecoin, and other assets on the network.

1 hours ago

Hyperliquid Assistance Fund Buys 48 Million HYPE for $1.36 Billion

Hyperliquid 3 Min Log Return (Topic 90) brings high-frequency log return predictions to the Allora Forge. Every 3 minutes, inference workers submit predictions over a three-minute time interval for a broad cross-section of the Hyperliquid perp market. The Allora inference network aggregates these into a synthesized prediction for each asset, designed to outperform the individual constituents over time. The Allora whitepaper explains how inference synthesis works.

Directional short-term alpha signals are an integral part of many execution and market making operations. For detailed background on how short-term directional predictions integrate with high-frequency market making systems, see Álvaro Cartea, Sebastian Jaimungal, and José Penalva, Algorithmic and High-Frequency Trading, Cambridge University Press, 2015. Sourcing these predictions through Allora makes this step easier for consumers and enables our community to deliver the secret sauce and earn.

Multi-output Topics
Allora’s recent network upgrade enables multi-output topic types. Instead of one topic per asset, we can now group many assets into a single topic. For topic 90, workers submit labeled arrays, where each element corresponds to a three-minute log return prediction for an asset.

In the Python example, your prediction callback returns a dictionary, and the SDK handles the labeled submission:

predictions = {
"btc": 0.00042,
"eth": -0.00018,
"sol": 0.00073,
}
These are illustrative log returns, not prices or probabilities: 0.00042 is approximately a +0.042% price move. Labels are lowercase asset names such as btc and eth.

Since this is a new blockchain rollout, we’ve capped each submission at 100 asset labels. Aim to submit the 100 assets with the most trailing 30-day volume. Workers will be scored on the top 90. If an asset is unavailable for a round, submit the predictions you do have; one missing asset shouldn’t kill the whole inference.

Data
Allora is now serving and supporting the Hyperliquid universe with one-minute OHLCV candle data through Atlas, our in-house data platform. At any time you can discover the universe through the dedicated current_hl_universe dataset:

curl -fsSG 'https://forge-data.allora.run/api/rows/' \
-H "X-API-Key: $ALLORA_API_KEY" \
--data-urlencode 'dataset_name=current_hl_universe' \
--data-urlencode 'limit=1000' |
jq '[.results[].values
| select(.status == "consumable" and .market_group == "perp")
| .candle_dataset_name] | sort'
This returns Atlas identifiers such as hl_btc_1min and hl_eth_1min. The market_group == "perp" filter keeps us in the native perp universe and excludes XYZ markets.

We’ve also extended the Atlas rows API to receive updates for multiple assets at the same time. That means one HTTP request for the batch, instead of a separate request per ticker. In our example runs, a routine refresh across roughly 178 assets took about 0.4 seconds, including the in-memory cache update. The full inference pass took a few seconds; transaction submission adds its own time.

For example, this requests the latest minute candle for both BTC and ETH:

curl -fsSG 'https://forge-data.allora.run/api/rows/' \
-H "X-API-Key: $ALLORA_API_KEY" \
--data-urlencode 'dataset_names=hl_btc_1min,hl_eth_1min' \
--data-urlencode 'limit=1' \
--data-urlencode 'ordering=-timestamp'
The response groups rows by dataset. Here is a shortened BTC row from an actual response, with metadata and the ETH group omitted:

{
"dataset_id": 1196859,
"timestamp": "2026-10-09T17:19:00Z",
"values": {
"open": 82615, "high": 82623, "low": 82614, "close": 82622,
"volume": 0.08883, "num_trades": 12
}
}
The timestamp is the candle’s opening time. The latest row may still be a partial candle. Increase limit to request more rows per asset, which is useful for refreshing an overlapping slice of recent history.

Builder kit integration
Corresponding to the new Atlas data endpoint, we’ve extended the Atlas data manager with some convenience methods.

AtlasDataManager.discover_hl_universe(): Atlas-specific discovery returning a list of exact dataset names for fresh, consumable native perps.
AtlasDataManager.get_bulk_1min_candles(symbols, limit=10): one multi-dataset HTTP request returning raw minute candles with asset identity preserved.

The ML workflow still works as before. Users configure AlloraMLWorkflow, call backfill() and get_full_feature_target_dataframe(), train models, evaluate through PerformanceEvaluator, and save serialized artifacts for the existing deployment flow.

The workflow already accepts multiple tickers, produces a dataframe indexed by ticker and time, shares historical/live feature extraction, and dispatches target construction through target_type.

Keep target_type="log_return". Existing multi-ticker historical dataset construction supplies the required training data with interval="3m" and target_bars=1. The following short snippet creates data, containing the input features and target for supervised ML:

import os
from pathlib import Path
import pandas as pd
from allora_forge_builder_kit import AtlasDataManager, AlloraMLWorkflow
LOOKBACK = 10 # Configurable number of past three-minute candles.
end = pd.Timestamp.now(tz="UTC").floor("3min")
start = end - pd.DateOffset(months=6)
atlas = AtlasDataManager(
api_key=os.environ["ALLORA_API_KEY"],
base_dir=Path("data"),
interval="3m",
)
universe = atlas.discover_hl_universe()
workflow = AlloraMLWorkflow(
tickers=universe,
target_type="log_return",
interval="3m",
target_bars=1,
number_of_input_bars=LOOKBACK,
data_manager=atlas,
)
workflow.backfill(start=start)
data = workflow.get_full_feature_target_dataframe(start_date=start, end_date=end)
The full example builds this dataset one asset at a time to reduce peak memory, removes unusable rows, and splits it chronologically before training.

Training and live inference
Training uses fixed, non-overlapping three-minute candles. The default input is a window of past normalized candles and the target is the close of the next unobserved candle, expressed as a log return. We pool the rows across tickers to train one model for all assets. The example uses one chronological holdout, keeping the latest 20% of observed timestamps for validation and excluding training labels that resolve after that boundary.

The example gives a light randomized search over LightGBM parameters, using six months of available historical data. In our October 8 run with a 10-candle lookback, that produced about 11.70 million training rows and 3.05 million validation rows across 178 assets. Counts depend on available history and your lookback. Tree-count checkpoints let us compare smaller and larger models without restarting each fit. The top three distinct trials are selected by lowest validation MSE, retrained on all usable data, and exported for live inference.

Our leading model passed five of the seven criteria in the pooled offline validation report, with participation assumed to be full. That’s a starting point, and this is where our community is going to help push the needle. These are validation results used during model selection; live performance is the next test. The example also reports metrics per asset and their equal-weight averages so we can compare that view with pooling all the data together.

Since inference workers are queried asynchronously, the current network scoring adopts the following conventions.

Inference resamples the historical one-minute data to the target start. Let T be the minute boundary starting the evaluation interval; the target ends at T + 3 minutes. Build the input bars backward from T so the final completed three-minute bar ends exactly at T. This aligns the live prediction with our training setup.

Example: the submission window opens at 12:01:27 UTC. The target starts at the preceding minute boundary, 12:01:00, and ends at 12:04:00. At window opening, 27 seconds of that target interval have already elapsed. The approximately 30-second submission window closes around 12:01:57; the prediction horizon still ends at 12:04:00.

The last completed input candle is labeled 11:58, covering [11:58, 12:01). The 12:01-labeled candle is separate: if Atlas has it, its latest close tells us how much the price has already drifted since the target started. The worker adds that realized log return to the model’s prediction. Each asset is checked independently; if its partial candle is missing, the worker assumes zero drift for that asset.

The worker keeps recent minute candles in memory, tops them up before opening the listener, and refreshes them in bulk when queried. Our example waits ten seconds before that refresh to let newer Atlas updates arrive, leaving roughly twenty seconds for prediction and submission. The target boundary stays fixed throughout.

The logic for adjusting raw candles for inference in real time has been designed and implemented in the topic 90 working example, in the Allora Forge Builder Kit.

The example exports the top three models found during hyperparameter search. Then use the WorkerManager for deployment and monitoring:

python notebooks/testnet/topic_90_hyperliquid_3min_logreturn/deploy_managed_example.py \
--model-run notebooks/testnet/topic_90_hyperliquid_3min_logreturn/results/model_search/RUN_TIMESTAMP \
--deploy
Replace RUN_TIMESTAMP with the run directory printed by the training script. WorkerManager allocates one address per model and starts each worker with its own in-memory data buffer. The workers keep running after the launching terminal closes. Use python -m allora_forge_builder_kit.workerctl dashboard to monitor them.

Contributing
The new topic can be found on the Allora Forge. Use the ticker dropdown to switch between assets and explore their predictions, all within the same topic.