AIX Global Innovations

First to achieve verified FTQC and quantum compute on raw NISC QPU hardware. Seed IQ brings governed execution to quantum, data centers, logistics, and more #AI

[AIX Global Innovations @AIX_SeedIQ]

Quantum Now Has a Path to Scale. Seed IQ Intelligence+Quantum

➡️ Ran on live @IBM #quantum hardware
➡️ Under real noise conditions
➡️ Held system-level fidelity at ~0.969, while preserving coherence & entanglement with 2 bell pairs across 3 logical qubits

[AlphaSense @AlphaSenseInc]

The next phase of AI transformation is about what your company learns from every decision it makes, and who owns that learning.

Enterprise value is the sum of a company's decisions, and in the agentic era, the most valuable asset a firm can build isn't a model. It's the loop that captures context, evidence, outcomes and patterns from every decision. Own that loop, and the company compounds. Route it through a single vendor, and your edge becomes a commodity your competitors can buy.

@cackerso, SVP of Product at AlphaSense, breaks down how leaders should think about architecting the agentic enterprise, and where to own versus partner. Read the full article: https://t.co/NxMWv7eLlF

[Denise Holt @deniseholt_]

🔴 ARC-AGI 3 Benchmark Seed IQ UPDATE: 💡 ✅ 8/8 games now, ✅ 60 levels, ✅ 2674 total actions ... ✅ 100% overall score - and look closely at the second image here of our Seed IQ scorecard... We are actually scoring 115% 😯 on most levels. And all of this at 2-3x the human baseline. Superhuman performance across all ARC Prize challenges! 🥳 🥳 🚀
(If you're wondering why we are not on the Leaderboard, it's because we have proprietary IP and the rules state you have to turn over codebase and commercialization rights to be included.)

🔗LIVE Scorecard link: https://t.co/4RVl0PVGGU
...click around in it to see more details and replays. This scorecard is displayed directly from the ARC Prize website.

Seed IQ uses zero tokens, no RL, no central control. This is topological perception under bounded autonomy. Active Inference combined with physics-driven dynamics.
Seed IQ is not getting there by memorizing examples, scaling a foundation model, or brute forcing action sequences.

It is improving because it is getting better at inferring the priors of the environment, or the hidden structure that makes the game solvable in the first place. Those priors are the invariances, constraints, symmetries, affordances, object relations, boundary conditions, and transition rules that define what actions are admissible and what paths can actually close.. Once those priors are inferred correctly, the search space collapses. The system no longer has to explore like RL or sample like a neural network.. It can identify the governing structure of the task and move through the admissible solution manifold directly.

This is why the performance is getting both faster and more deterministic. (Superhuman level) Seed IQ is not just playing better. It is perceiving the structure underneath the game better.

#AIX #SeedIQ #ARC3 #ARCAGI3 #Quantum #DataCenters #EnergySystems
#AIXGlobalInnovations
@GregKamradt

[Lightmatter @LightmatterCo]

Moore's Law is dead — @theanalognick breaks down why optics are the future of AI computing in this @tomshardware piece: Copper hit its limit. 2027-28 will be huge for near-package and co-packaged optics. Lasers are the next bottleneck.

Lightmatter is scaling its solutions through open standards and through @nvidia NVLink Fusion ecosystem as an optical interconnect partner.

@OpenComputePrj #photonics #optics #artificialintelligence

https://t.co/0Qjt6WGw4M

[AIX Global Innovations @AIX_SeedIQ]

Data normalization and collaboration infrastructure platform @Narrative_IO lunched a major expansion of the Narrative Marketplace, evolving it from a place to find and license data into a composable hub for everything a modern data & AI strategy needs https://t.co/dzM1wiJpN4 #AI

[Pep Invest @Pep_Invest]

Lightmatter recently demonstrated the world’s first bidirectional 16-wavelength optical link over a single fiber.

To build this ultra-fast connection for AI supercomputers, @LightmatterCo needs a physical laser source that can emit exactly 16 different wavelengths of light simultaneously, with extreme precision and stability.

However, Lightmatter does not manufacture these lasers themselves.

$SIVE is currently the only commercial laser manufacturer in the world that has officially developed, publicly validated, and delivered 16-wavelength Distributed Feedback (DFB) laser arrays in mass production.

They originally developed this specific technology for their partnership with @AyarLabs .

Because $SIVE is the only company offering this rare 16-wavelength laser technology, and Lightmatter is the only company building a 16-wavelength optical chip network, the industry conclusion is clear: Lightmatter must be using $SIVE technology.

[Lightmatter @LightmatterCo]

Lightmatter is in the spotlight in @IEEESpectrum!
The latest feature explores a critical turning point in data center architecture as the industry rapidly transitions to optical networking inside the rack.

In the article our VP of Product, Roy Kim, breaks down how Lightmatter is solving the copper bottleneck. While traditional methods rely on interfacing optical chiplets to the edges of processors, we’re 3D-integrating silicon photonics as the interconnect fabric inside the package. With Passage photonic interposers, we allow processors to stack vertically—bypassing the edge-routing chokepoints that limit communications bandwidth.

The packaging hurdles of the past have been cleared. Today, scale-up networking also requires dense, high-performance external laser solutions. Lightmatter is tackling this head-on by integrating lasers directly onto silicon, driving unprecedented bandwidth density and scalability. Read the full article: https://t.co/3HqQcyryKK

[Lightmatter @LightmatterCo]

Lightmatter was just named one of the 10 coolest semiconductor startups of 2026 by @CRN. We're proud to be recognized alongside other innovators pushing the boundaries of AI infrastructure and tech hardware — see page 7 for our overview. Thanks @markharanas for the spotlight. @theanalognick
https://t.co/ZAbw1VDCRy

[Denise Holt @deniseholt_]

I met Charles Bonnaire of Equinix in London last week at The Economist’s Commercialising Quantum event. This is such a great example of the difference between Active Inference AI and deep learning! He really does understand. 🙌🏻🥳 … This might help some if you understand, too

For clarity, Seed IQ is our AI engine (our product - it uses active inference combined with physics-driven dynamics), and our company is AIX Global. (He mixed that part up.)…

🚀🙌🏻

Link to his full post here: https://t.co/ZpwRhcjgNr
@AIX_SeedIQ
#SeedIQ #Quantum #ActiveInference #DataCenters #EnergySystems

[AIX Global Innovations @AIX_SeedIQ]

@elonmusk 😳

[AlphaSense @AlphaSenseInc]

$MU CEO Sanjay Mehrotra says their large customers include data center and the large and medium customer agreements go across data center, consumer, and automotive markets. Large agreements generally have a price band with a floor and a ceiling established at CQ2 price levels, which are reflected in their FQ3 results and FQ4 guidance. Those price levels provide for unprecedented levels of profitability.

"Our large customers include data center and the large and medium customers that you mentioned, of course, including our smaller customers, they do go across data center, consumer, and automotive markets. We have provided you color on the large agreements have, of course, generally have a ceiling price and have a price band which has a floor as well as a ceiling, and the ceiling is established at the CQ2 price levels. Of course, you know that CQ2 price levels are reflected in our FQ3 results as well as FQ4 guidance, and they provide for unprecedented levels of profitability. Those price bands also provide for floor prices, where the gross margins are well above the peaks at any past cycle in company's history.

The large agreements that we mentioned, these are multi-year agreements, and they provide us tremendous visibility to demand customer commitments, and they, of course, come with the financial commitments, including cash deposits, that Mark elaborated on earlier."

[AlphaSense @AlphaSenseInc]

$MU Earnings:

- Revenue of $41.46 billion versus $23.86 billion for the prior quarter and $9.30 billion for the same period last year
- GAAP net income of $28.24 billion, or $24.67 per diluted share
- Non-GAAP net income of $28.86 billion, or $25.11 per diluted share
- Operating cash flow of $25.39 billion versus $11.90 billion for the prior quarter and $4.61 billion for the same period last year

“Micron’s record fiscal Q3 financial results and even stronger outlook for Q4 reflect the strategic value of memory in the AI era,” said Sanjay Mehrotra, Chairman, President and CEO of Micron Technology. “Micron is investing at record levels in technology, products and supply to address our customers’ rapidly growing demand. We believe our multi-year Strategic Customer Agreements will significantly enhance the durability and predictability of Micron’s strong financial performance.”

[NIXXY @nixxyofficial]

NIXXY's Next Chapter: Tachyon9 Secures Offtake Agreement Supporting $156 Million Annual Revenue Platform and Up to 1 GW of Grid-Independent AI Infrastructure Powered by Abundant North Dakota Natural Gas

Learn how NIXXY is involved here: https://t.co/o7YOxO2C2p

[Denise Holt @deniseholt_]

Not often you see phone booths even in London anymore ☺️ … Heading back to California today after an awesome week at The Economist’s Commercialising Quantum Event. They put me in touch with their PR team to use the event to announce our quantum breakthrough, so I had about 8 journalists from their media team reach out to me directly for interviews, including an interview by WIRED magazine at the event yesterday! ☺️ … Our press release on Monday was syndicated out and published to almost 500 media outlets. And I made direct connections with all of the major quantum hardware companies here - IBM, Microsoft, Rigetti, IonQ, QuEra, TreQ, and more - with follow up meetings coming because our recent FTQC breakthrough enables all quantum hardware to become useful for compute now, rather than years from now (which has been the current narrative/expectation).

I met so many wonderful people in this industry over the past few days. I’m truly grateful for this experience. Especially to Christopher Bishop for the invite and all the amazing introductions, and the The Economist for welcoming me to their stellar event and facilitating so much media coverage.

All in all, this week could not have gone better! I am so excited. When we first looked at quantum at the beginning of this year, we were just looking at it as a proof case to showcase how Seed IQ could govern even the most complex systems. We didn’t set out to achieve the holy grail of quantum computing - what everyone still thinks is 3-5+ years away, and yet, here we are… I guess we are a quantum company now! LoL

And this is only one vertical for us. Seed IQ is an adaptive multiagent autonomous control (AMAC) engine for governing the execution of all complex systems, so there are an unlimited number of other verticals and industries that it will prove to be just as effective in.

Things are only just beginning…. But what a coming out party this has been. 🥳🥳🥳🙏🏻

Denis O. @AIX_SeedIQ
#SeedIQ #quantum #activeinferenceAI

AIX Quantum Paper:
https://t.co/q6pwbAzVMX

Notebook: https://t.co/M5YFeY5ITw

[Denise Holt @deniseholt_]

🔴 Here's a walkthrough of AIX’s Seed IQ™ AMAC engine solving the preliminary first three ARC AGI 3 games — where complexity was increased just days before the official launch. (On March 21st, they substantially updated the complexity in gameplay. The developer toolkit has been avail. since the end of January, and we started looking at it last week. There was already a substantial “human-level” baseline from all the actual humans who have been playing these games since last summer, and there was a “verified” and “unverified” leaderboard for agents who had been attempting to play since late January)

➡️ This is where things get interesting...

🔸 Before the complexity increase, (last week) we achieved:
• 100% Score across 3 games / 20 levels
• 536 actions (human-level performance – the system placed our score under the top 3 human scores until that point). Denis O.posted screenshots of this last week.

🔸 Since the complexity increased last Saturday, we still achieved:
• 95% overall score across all three games
• 681 actions (still human-level performance given the best human score is noted to be 546 actions)
• Solved all levels 20/20 of the three available games, including LS20, in which the complexity difficulty skyrocketed.

▪️ At the same time, other agents attempting these games degraded sharply. The top baseline agent, Stochastic Goose, dropped from 12.58% Score with 255,964 actions and 18 levels & only 2 games completed, to today displaying only a 0.25% Score (and it doesn’t show how many actions) after the complexity increase. All leading LLM-based systems (GPT 5.4, Gemini 3.1, Grok4, and Claude Opus) are effectively at zero on these tasks:
· GPT 5.4 Score: 0.3%
· Gemini 3.1 Score: 0.2%
· Claude Opus: 0.2%
· Grok 4 Score: 0.0%

This is not about incremental improvement.

🔸 🔸 What you’re seeing in this video is a different class of intelligence.

Seed IQ™ is not trained to recognize patterns from past data. It constructs a world model in real time, using topological perception & adaptive multi-agent dynamics to continuously update its understanding and strategy. (Which we demonstrate and explain in this video.)

When the environment changes, it doesn’t retrain. It adapts.

When a strategy fails, it doesn’t stall. It recomputes and reroutes.
And as complexity increases, it maintains coherence rather than degrading.

➡️ That’s the difference. … And it is a stark difference.

All other agents attempting these games were already struggling, and with the substantial increase in complexity over the weekend, they nose-dived in performance. Seed IQ™ was not only able to adapt to the complexity increase, but learned to solve these games while maintaining human-level performance.

If you’re interested in where AI moves beyond prediction and into real-time adaptive execution in complex systems, this is worth watching.

(These results were achieved prior to today's official launch, and we are in communication with the ARC Prize folks to negotiate being able to participate through official evaluation without revealing our proprietary IP. We hope they will find a way to enable/allow this.)

Check it out on YouTube here: 🔗https://t.co/eCOyr3huJB

#AIX #SeedIQ #ARCAGI3 #ActiveInferenceAI #MultiAgentSystems #ARCPrize

[Peter Bordes @PeterBordes]

The AI infrastructure Neo Clouds ranked by NTM est revenue growth.

Next to join the group are $NIXX $BRUN

$NBIS $IREN $WULF $CRWV $CIFR $APLD

[Exascale Labs @ExascaleLabs22]

AI data centers are projected to triple electricity demand in the next few years.

Power + cooling = is the new bottleneck. Traditional DCs simply can’t scale fast enough.

The market is exploring: modular, liquid-cooled AI infrastructure designed around power, cooling, and deployment speed.

At Exascale Labs, we spend a lot of time thinking about that physical layer.

[Peter Bordes @PeterBordes]

Big first $2.34b rev deal and like investing at pre IPO venture valuations in $NBIS $IREN $CRWV

$NIXX Tachyon9 data center campus signs $2.34 billion AI infrastructure deal with Yotta Data Services.
*$2.34b revenue initial 100 MW deployment of 1 GW

https://t.co/MijgC5OmT1 #AI

[Trajectory AI @Trajectory_AI]

Quantum innovator @AIX_SeedIQ achieved FTQC on IBM QPUs, making governed quantum compute possible. Enabling what the quantum industry has long considered the threshold for useful #quantum computing, then advanced beyond it into real molecular computation. https://t.co/GHkHftcTTx

[Denise Holt @deniseholt_]

🔴 BREAKING NEWS: AIX Global Innovations announces quantum breakthrough: In April 2026, Seed IQ independently achieved FTQC on rented public access IBM QPUs, making quantum compute possible today, not 3–5 years out.

➡️ As I arrive in London for Europe’s biggest annual quantum conference: The Economist’s Commercialising Quantum event, I’m so pleased to share this news just released this morning.

AIX achieved the holy grail of quantum computing: full Fault-Tolerant Quantum Compute. In a process that began on April 9, unfolding over the past two months, AIX’s Seed IQ held all 150 qubits of the register - essentially the entire processor IBM made available - as a single entangled logical system at near-perfect fidelity, with zero recorded logical errors. Crossed the line of fault-tolerance, then composed the full FTQC primitive pipeline, computing ground state molecules well within chemical accuracy (spectroscopic and wave number accuracy)  - on public, cloud QPU hardware anyone can rent, publishing a 103 page report on the entire progression of the 8 weeks of achievements.

Read the full Press Release here:

🔗 https://t.co/xyvRNMtkgJ

So excited to be here in London. Looking forward to a stellar event!!

@AIX_SeedIQ @TheEconomist

#quantum #FTQC #seediq #AIX #ActiveInferenceAI

[AIX Global Innovations @AIX_SeedIQ]

Meet our CEO @deniseholt_ at @econ_enterprise Commercialising Quantum Global, and learn how we’re the first to achieve verified FTQC and quantum compute on raw NISC QPU hardware. Seed IQ brings governed execution to quantum, data centers, logistics & more. #EconQuantum #Quantum