PLEXADI INTERNATIONAL

Deep-tech · Sveden · Diazezoù diazez 12

Diazezadur diazez-12 evit mediaoù, stokañ ha treuzdougen.

Plexadi International a ziorren ar gwiskad matematikel dindan ar rummad diazezoù niverel da zont — kodekoù, reizhiadoù restroù ha protokoloù treuzdougen savet en un doare orin war ar matematikoù daoudezimal (diazez-12) kentoc'h eget war harzoù teorikel ar savadur binarel hengounel.

11+
Brevedoù degaset
4
Barregezhioù
2026
Deiziad pennañ
12
Diazez matematikel

Universal glyph transport

Information. In another form.

😊✦◈😎🌐♥☀🙂🔑🤝🔐✓😊💜
01 / Server02 / Network03 / End user

Smileys, emoticons, symbols and colour become a visual language for information in transit. Concept illustration—not a live network capture or a disclosure of the protocol.

Dizoleiñ an teknologiezh

Ar rakvarn

Ur c'henemglev e oa an binvioù, n'eo ket un dibenn.

Pep live eus an diazezoù niverel a-vremañ — kodekoù, reizhiadoù restroù, treuzdougen, kriptografiezh — a hêrezh strishadurioù ur reizhiad niverel diazez-2 dibabet evit an elektronegezh er bloavezhioù 1940. Pinvidikoc'h eo matematikoù ar bazenn-12 dre vras ​​: muioc'h a rannerien, darempredoù harmonikel glanoc'h, simetriezhioù naturel a rank ar binvioù simulañ a-gostez.

Plexadi a sevel ar gwiskad ma vez implijet ar perzhioù-se en un doare orin — adalek ar c'hodek a enroll ar ster kentoc'h eget ar pikseloù, betek ar reizhiad restroù a indez en daoudezimal, betek an dorn-ha-dorn a glot gant ar framm harmonikel. An disoc'h a zo un diazez n'eo ket buanoc'h hepken, met reizh war an dachenn matematikel.

Diazez matematikel

12 : 2

1 / 3 · 3 : 1
BASE 12= 0.412
012345678910113 : 1
12 % d = 0 1 · 2 · 3 · 4 · 6 · 12
BASE 2↻ 0.(01)2
013 : 1
2 % d = 0 1 · 2

Teknologiezh

Seizh piler, ur fondasion matematikel.

/ 01

P6-Reizhiad restroù orin Base-12

Roadennoù enrollet, indekset ha chomlec'hiet en un doare orin e daoudezimal — gant perzhioù frammadurel na c'haller ket adtapout en ur gwiskad binarel orin. Arveret e vez evit ar stokañ, ar vemor hag an E/S.

/ 02

Treuzdougen glifoù hollek

Un treuzdougen gwelet-kentañ a gas ne vern pe restr, protokol pe stream dre ne vern peseurt skramm, kamera pe chadenn optikel — eus ar skramm d'ar c'hamera d'ar skignañ d'an ingest aer-gapped.

/ 03

P4-Treuzdougen binarel-gwelet

Ur protokol hiron a dremen hep kudennoù etre treuzdougen glifoù binarel ha gwelus, o vont en-dro d'an doareoù rouedad digredus, enebet pe aer-gapet hep koll gwirionded an treuzdougen.

/ 04

Prennañ ar sinal harmonikel

Ur skoazell-dorn kriptografikel diazezet war darempredoù harmonikel ar bazenn-12 — resis ouzh gorreadoù tagañ eskemm alc'hwezioù hengounel ha dereat evit an treuzdougen sur eus ar rummad da zont.

/ +

Savadurezh all

Gwarezet eo ar gwiskadoù ouzhpenn — en o zouez ur c'hodek media semantek, un is-strollad metaroadennoù, ur protokol gwiriañ, ur chadenn glifoù liesmorfek hag un enframmadur OS orin — evel arz a-raok ha hegerz dindan an dizoloadenn etrezo.

Perc'henniezh speredel

Ur portfolio glan, ijinour nemetañ.

An holl brevedoù degaset a ziskouez Robert Dave Myrland evel an ijinour nemetañ — ur post IP ral ha dister evit un embregerezh deep-tech. Gant Brann AB e vez meret an teulioù. Diazezet eo an testenioù kriptografikel a-raok evit ijinadennoù all war levrioù foran digemm.

TitlTeknologiezhNiv. goulennDeiziad pennañ/ PRVGoulennoù
P4-Treuzdougen binarel-gweletTransportSE 2630545-82026/08/1730
P4-S — Semantic Routing & ReconstructionSemantic reconstructionSE 2630818-92026-11-1151
P5-Treuzdougen restroù hollekContainerSE 2630546-62026/08/1720
P6-Reizhiad restroù orin Base-12SanailhañSE 2630547-42026/08/1720
P7-GlyphLockCryptologySE 2630541-72026/08/1622
P8-BC12 — Relational State-Space ComputingComputingSE 2630817-12026-11-1011
P9 — Base 0Mathematics2026-11-116

P4-S · P8-BC12 · P9: PRV receipt dates, not a confirmation of earlier priority or patent grant.

Emañ an teulioù ouzhpenn — en o zouez liveoù savadurioù diazez-12 all ha fazioù etrebroadel dre USPTO ha barregezhioù all — o vont war-raok hag a vo meneget amañ pa vo resevet niverennoù ar goulennoù. Testenn klemm munut a c'haller kaout evit ar strolladoù barrek dindan an dizoloadenn etrezo.

En e sav

Ijinour nemetañ

An holl brevedoù a zo bet lakaet gant un ijinour anvet hepken — n'eus ket a amjestregezh war ar c'hen-ijinour, n'eus ket a zisputoù perc'henniezh a-is.

Degaset dre Brann AB

Unan eus embregerezhioù brevedoù pennañ ar rannvro nordik a ra war-dro an holl deulioù, an enebiezhioù hag an holl fazioù etrebroadel.

Arz kentañ kriptografikel

An testenioù a-raok a zo staget ouzh levrioù foran digemm, gwiriekaet en un doare dizalc'h da bep mare.

Kaozeadennoù oberiant

E-barzh kaozeadennoù oberiant gant kevelerien bras war dachenn ar skingomz hag ar mediaoù. Muioc'h a ditouroù a c'haller kaout dindan NDA.

Plexenc research · Test report v1.2 · 10 August 2026

8K video. Two benchmarks.

Plexadi’s supplied report compares a synthetic 8K UHD clip with a real NASA “Moving Water” 8K DCI clip. These are reported results, not independently reproduced tests or a NASA endorsement.

Animated summary of the report · 58 s · Figures as reported, not independently reproduced

A · Synthetic 8K UHD

7680 × 4320 · 60 fps · 180 frames · 3.00 s

489 B

Reported semantic stream, video + audio

Glyph-PNG transport
296,091 B
H.264 MP4 reference
10,176,686 B

H.264 reference generated with x264, medium preset.

B · NASA-source 8K DCI

8192 × 4096 · 23.976 fps · 71 frames · 2.96 s

445 B

Reported semantic stream, video + audio

Glyph-PNG transport
251,733 B
H.264 MP4 reference
16,643,985 B

H.264 reference is the original NASA-source MP4, not re-encoded.

100% semantic-stream integrity ≠ lossless original video

The report states byte-perfect recovery of the transmitted semantic descriptions in both tests. It explicitly does not claim pixel-perfect or byte-identical recovery of the original media. Video and audio reconstruction depend on an AI receiver; source-image quality was not established by these byte counts.

Raw-video sizes are equivalent uncompressed RGB footprints, not files transmitted by Plexenc. Semantic-stream sizes exclude the AI receiver and other reconstruction resources. These results concern Plexenc semantic media research, not proof of P4-Binary-Visual Transport’s deterministic original-byte transport or the security of P7-GlyphLock.

Read the full test report

PDF · 4 pages · Source code and benchmark data available under NDA

Enterprise & banking security

P7-GlyphLock & B12PPP

A base-12 cryptology research direction for financial institutions and security-sensitive enterprises.

Beyond another layer of patches.

Banks and financial institutions devote substantial resources to patching, monitoring, key management, and maintaining layered security infrastructure. Legacy systems and complex integrations can make that work costly and continuous.

Plexadi positions P7-GlyphLock and B12PPP as a different approach to protected communication, exploring base-12 cryptology and session-specific symbolic exchange. The aim is to address weaknesses in sensitive exchanges at the protocol level—not merely add another security layer.

Where the 1200× estimate comes from

Layers an attacker must solve before cryptanalysis begins

Conventional cipher comparisons assume an attacker already holds clean, well-delimited packets under a known protocol. B12PPP is designed so that an interceptor must first overcome several compounding obstacles.

Unknown boundaries

Slices vary in size, so an interceptor cannot tell where one handshake ends and the next begins.

40–90 slices of unequal size

Ordering alone grows factorially: 40 slices allow about 8 × 10⁴⁷ orderings, 90 slices about 1.5 × 10¹³⁸.

Undisclosed base conversions

Correctly ordered slices still need the right sequence of base and symbolic conversions to become meaningful data. These conversions include proprietary mathematical bases invented by Plexadi and held as company trade secrets, so an attacker cannot simply test known number bases.

A moving target

The semantic language rotates every 60 seconds, or sooner when unusual line latency is detected, making captured state quickly obsolete.

Mixed conversion pipelines

Slices of uneven share, such as 3.3%, 5.9% or 1.2% of the whole, do not all pass through the same conversion order. Solving most slices still leaves the remainder unsolved, and reconstruction requires every slice. Any partial progress expires when the semantic language rotates.

These compounding layers are the reasoning behind the AI-generated comparison. It remains a design rationale, not a proof: independent cryptanalysis is still required, and no system is shown to be impossible to break.

Protect sensitive exchanges

A proposed fit for confidential messaging and data exchange between participating endpoints. No elimination of vulnerabilities or immunity to attacks is demonstrated.

Evaluate before deployment

Independent cryptanalysis, implementation audits, interoperability testing, and regulatory assessment are needed before use in banking or enterprise systems.

Potential cost savings

The ambition is to reduce operational complexity and recurring security overhead. Large savings remain a proposed benefit, not a measured result; integration, audit, and maintenance costs must be included in any comparison.

The economics of security

What could a 15% reduction mean?

Gartner’s February 2026 forecast projected worldwide information-security spending of $244 billion in 2026—across all industries, not just banks.[1]This is a dated forecast, not a measured global banking total.

Bank of America’s CEO reported more than $1 billion in annual cybersecurity spending in June 2021.[2]A June 2026 trade-press article estimates JPMorgan Chase’s annual cybersecurity operations spending at $2.5–$3 billion; that figure is not bank-confirmed in the sources reviewed.[3]

The proposed $35–$45 billion annual banking and financial-services total could not be verified from the Gartner or GlobalData sources reviewed. It is used below only as a Plexadi-supplied scenario assumption, not an attributed industry finding or a census of every bank worldwide. Financial services includes more than banks.

Global financial-sector scenario

$5.25–$6.75billion/year

$35–$45 billion assumed eligible annual spending ×15%.

Single large-bank scenario

$150–$375million/year

$1–$2.5 billion assumed eligible annual spending ×15%.

Using the separately reported $2.5–$3 billion baseline instead gives $375–$450million per year at 15%, not $150–$375 million.

These are hypothetical gross savings, not measured P7-GlyphLock or B12PPP results. The 15% reduction is an assumption, not a forecast. Actual savings require evidence that eligible infrastructure costs can be reduced without weakening security or compliance.

Formula: eligible annual spending × 0.15. Applying it to an entire cybersecurity budget assumes every dollar is eligible; a cryptology change alone does not establish that. Net savings must subtract licensing, migration, integration, independent audits, training, and ongoing operating costs.

Data centres & server energy

Plexadi energy-efficiency research

Less electricity for the same useful work?

Electricity is a major operating consideration for data centres and server infrastructure. Plexadi is interested in whether its technology could reduce the energy required to deliver equivalent computing and data-handling workloads.

Lower energy use, if demonstrated, could help reduce electricity costs and operational emissions. Financial and environmental benefits depend on actual consumption, local electricity prices, and the energy mix.

Equivalent workloads

Compare the same completed work, output quality, throughput, and latency on a defined hardware and software baseline. Base-12 representation alone does not demonstrate lower electricity use.

Measure actual energy

Validation requires repeatable energy measurements, such as kWh per completed workload, including conversion and integration overhead—not just mathematical operation counts.

Define the boundary

Server energy and whole-facility electricity are different measures. Cooling, networking, storage, utilisation, and facility overhead must be accounted for before claiming data-centre-wide savings.

Markets

The P-Series covers 12 industries.

From critical infrastructure to consumer platforms, the P-Series patent family applies wherever data must be transported, stored, or secured.

Cybersecurity

Base-12 cryptography and harmonic signal lock for next-generation defense.

SOC & MSSP

Security operations centers and managed security providers gain structurally stronger transport and storage layers.

Energy & Power

Resilient data transport for grid operators, utilities, and critical power infrastructure.

Telecom & Digital Infrastructure

Universal glyph transport across carriers, data centers, and degraded network conditions.

Healthcare & Life Sciences

Secure, verifiable transport and storage for patient data, imaging, and research.

Transport & Logistics

Tamper-evident data exchange across fleets, ports, and supply chains.

Public Sector

Sovereign-grade cryptography and storage for government and public administration.

Manufacturing & Industry

Air-gapped ingest and optical transport for industrial and OT environments.

Retail & E-commerce

Secure transactions and efficient media delivery at consumer scale.

Cloud & SaaS Providers

Base-12 native storage and transport layers for hyperscale and edge platforms.

Broadcasting

Semantic codecs and visual transport for broadcast, streaming, and post-production.

Defence & Space

Defence organisations, defence contractors and space exploration: smaller representations for bandwidth-constrained links and remote operations.

Defence contractors · Space exploration

Less to transmit. More reach.

Remote defence operations and space exploration can depend on links with limited bandwidth, intermittent coverage or long delays. The P-Series explores sending smaller representations instead of repeatedly transferring complete static files, with the intended content reconstructed at a compatible receiver.

Less data over the link

Compact representations may reduce the amount of data that must cross a constrained network, compared with sending complete static files.

Designed for poor coverage

Smaller payloads could reduce transfer time and the risk of stalled loading when bandwidth is scarce. They cannot restore a lost connection or guarantee uninterrupted delivery.

Reconstruction at the destination

A compatible receiver reconstructs the represented content. Exact file recovery and semantic reconstruction are different capabilities; the required fidelity must be established for each use case.

Potential applications, not validated defence or space deployments. Benefits depend on the data, receiver resources and link conditions. Semantic reconstruction does not establish byte-identical recovery of original static files or media; independent testing is needed for mission-critical use.

SDK under development — seven patents pending.

Serious customers and integration partners are welcome to contact us for early access and licensing discussions.

Contact us

Aotreoù

Aotreoù diuzet war pemp tachenn.

Mediaoù ha skignañ

Kodekoù semantek ha treuzdougen evit ar skignañ, ar skignañ hag ar mont en-dro goude ar produiñ.

Sanailhañ

Diazez-12 treuzfollennoù reizhiad restroù orin evit an embregerezh, ar c'hoad hag ar stokañ bord.

Protokolloù treuzdougen

Treuzdougen glifoù hollek evit an doareoù rouedad enebour, distrujet ha digor-aer.

Reizhiadoù optikel

Treuzdougen optikel eus ar skramm d'ar c'hamera hag eus ar c'hamera d'ar stokañ dre implijoù greantel.

Taol-dorn kriptografikel

Protokolloù prennañ sinal harmonikel evit eskemm alc'hwezioù sur en diazezoù nevez.

Partners

Working with select partners.