Photonicsense
Sense2 / Photonicsense Pty Ltd · Patent Portfolio

Seven patents spanning AI infrastructure, defence & edge perception

A concentrated portfolio addressing three of the highest-spend categories in 2026–2028: AI datacentre thermal & compute bottlenecks, defence autonomy, and edge perception for safety-critical systems. Active monetisation workstream; acquirer targets identified and ranked. Lead asset currently in examination at IP Australia.
Filed patents
7
1 in examination · 6 pending
Risk-weighted value
$400M–$1.2B
Portfolio, base case
Nearest-term grant
DEAN
Cooling / thermal IP
Commercial focus
HBM + Defence
Fastest path to close
Portfolio summary only. This page describes addressable problems and market positioning for each patent. Specific mechanisms, claim structures, worked examples, numerical parameters, and architectural details are held as trade secret and in the filed specifications. A technical deep-dive is available to qualified acquirers under NDA. All filings are owned by Sense2 Pty Ltd; sole inventor: Nancy Joffee.
Lead Asset

DEAN is the spearhead of the portfolio

DEAN is in active examination at IP Australia with a realistic grant path. It addresses the single largest architectural constraint in AI datacentre scaling: thermal density in HBM and co-packaged optics. Every HBM4 stack, every Rubin-class GPU, and every photonic compute platform is thermally bound. The DEAN response package is prepared; grant event expected 12–18 months. First-tier acquirer conversations are sequenced to coincide with examination completion.

The Portfolio

Seven patents ranked by commercial readiness

Tier 1 (near-term close): DEAN, DRONE · Tier 2 (strategic, demo-gated): IOC, VPPNN · Tier 3 (breadth & option value): CIVILIAN, SMC, Dean-PUF. Patents are colour-coded by primary domain: amber = AI infrastructure · blue = defence · purple = physics/compute · cyan = edge perception · green = civilian.
DEAN — Silent Cooling Platform
AU 2025283584 · In examination
Tier 1
Thermal management for AI infrastructure — cooling plus vibration & phase-noise suppression
HBM Cooling AI Infra CPO In examination
Problem addressed

Thermal density is the binding constraint on HBM4/5/6 scaling and on every emerging co-packaged optics (CPO) platform. Existing cooling approaches either cannot reach the heat flux required or introduce vibration that destabilises coherent optical components. DEAN addresses both simultaneously.

Market fit

HBM manufacturers and CPO platform builders are actively procuring. Relevant buyer universe includes memory manufacturers, foundries with packaging divisions, datacentre cooling infrastructure companies, and emerging photonic compute platforms. This is the portfolio's fastest path to a licensing or strategic sale conversation.

Indicative value
$150M – $600M
Grant probability
~65–70%
DRONE — Passive Optical Defence Classification
Filed March 2026 · Defence application
Tier 1
Zero-power, EMP-immune threat classification for directed-energy and counter-UAS systems
Counter-UAS Directed Energy EMP-immune Zero-power
Problem addressed

Counter-UAS and directed-energy weapons require threat triage at speeds and power envelopes that conventional GPU-based classification cannot meet — particularly in contested electromagnetic environments where digital systems are vulnerable. DRONE provides classification without the failure modes of software-based approaches.

Market fit

Procurement window is open now. Ukraine, Gaza, and Iran theatres have accelerated counter-UAS and directed-energy programmes across every NATO-aligned prime contractor and equivalent. Buyer universe includes pure-play EW/photonic defence primes, large diversified primes, and allied-nation defence agencies. Demonstration prototype materially increases negotiating leverage.

Indicative value
$400M – $1.5B
Unlock gate
Optical bench demo
IOC — Inference Safety Architecture
AU 2026200149 · PCT priority window
Tier 2
Real-time safety fallback architecture for autonomous systems and robotics
Robotics Safety Automotive EU AI Act
Problem addressed

Every autonomous platform — vehicles, humanoid robots, industrial robotics, drones — needs a safety layer that cannot be compromised by the AI it supervises. The EU AI Act (fully applicable August 2026) mandates auditable, transparent, and effectively overseeable operation for high-risk AI systems, with non-compliance fines up to 7% of global revenue.

Market fit

Regulatory deadline creates forced-buyer dynamics. Buyer universe includes automotive OEMs and Tier 1 suppliers, humanoid robotics platforms, industrial robotics primes, and edge-AI silicon vendors. PCT conversion is time-critical — global rights outside Australia depend on priority window preservation.

Indicative value
$200M – $800M
Critical deadline
PCT priority
VPPNN — Volumetric Photonic Processor
AU 2025283647 · Filed
Tier 2
Post-digital neural network architecture in a photonic substrate
AI Architecture Optical compute Coherent
Problem addressed

AI compute is becoming energy-bound before it becomes transistor-bound. By 2028–2030, power per datacentre is the binding constraint, not GPUs. Planar photonic approaches (currently attracting billions in venture funding) address part of this problem; volumetric approaches extend the envelope further when coupled with the portfolio's thermal-stability work.

Market fit

Strategic rather than near-term. Photonic AI category has attracted over $1B in venture funding and aggregate valuations approaching $6.5B; this is a validated category, not a speculative one. VPPNN is strongest when paired with DEAN in a combined thermal-plus-compute story. Buyer universe includes photonic AI platform builders and hyperscalers with optical roadmaps.

Indicative value
$200M – $800M
Best paired with
DEAN
CIVILIAN — Passive Optical Classification
Filed March 2026 · Civilian applications
Tier 3
Classification for safety-critical, medical, industrial, and privacy-sensitive applications
Multi-application Edge Privacy
Problem addressed

Safety-critical and privacy-sensitive classification at the edge — autonomous vehicles, medical imaging, industrial inspection, environmental monitoring, surgical guidance, aged-care safety, and similar applications. Each application has distinct regulatory, power, or privacy constraints that software-based classifiers cannot cleanly satisfy.

Market fit

Breadth rather than depth. CIVILIAN is an option-value patent covering many applications with different buyers in each. Most valuable as part of a portfolio sale where an acquirer wants optionality across verticals, rather than as a standalone transaction. Specific high-value sub-applications may be licensed independently.

Indicative value
$100M – $500M
Monetisation path
Portfolio or licence
SMC — Session Memory Controller
AU 2025283642 · Filed
Tier 3
Control-plane acceleration for AI inference infrastructure
AI Infra Inference
Problem addressed

LLM inference at scale is memory-bandwidth bound, not compute bound. The control-plane layer that sits between the application, the memory hierarchy, and the inference engine is a known bottleneck. SMC addresses this layer specifically — distinct from the data-plane KV-cache optimisation work that dominates the current inference-efficiency literature.

Market fit

Crowded field, narrowing window. The broader inference-efficiency category is moving monthly; recent work on KV-cache compression and disaggregated serving validates the category but also raises the bar for novel IP. SMC's value depends on precise positioning of its specific control-plane differentiator. Best monetised as part of a portfolio transaction rather than standalone.

Indicative value
$50M – $200M
Monetisation path
Portfolio bundle
Dean-PUF — Authenticated Photonic Computing
Filed · Security application
Tier 3
Physics-layer authentication gate for holographic AI model weights
Security Physical-layer
Problem addressed

Model-weight integrity and authentication for high-value AI systems. As AI model weights become strategic assets (frontier models, defence-sensitive models, proprietary training runs), physics-layer verification of weight integrity addresses a real emerging problem that software cryptography alone cannot solve.

Market fit

Long-tail option value. Narrower, more speculative market than the other patents. Potential buyers are a small set of defence-AI and frontier-AI operators with specific weight-authentication requirements. Held primarily for portfolio-breadth and as a long-dated option; not prioritised for active outreach.

Indicative value
$20M – $80M
Monetisation path
Long-dated option
Valuation

Portfolio scenario ladder — risk-weighted

Four scenarios ranging from licensing-only floor to full strategic acquisition. Individual patent values reflect realistic near-term acquirer economics and are not cumulative guarantees. Ranges adjust as DEAN advances through examination and as demonstration prototypes are built.
Bear · Licence floor
$180M
Non-exclusive licensing of DEAN and one or two secondary assets. No strategic transaction. Represents the base-rate outcome for filed but un-demonstrated portfolios of similar scope.
Base · Staged outcome
$650M
DEAN grants, one defence buyer licences DRONE, IOC and CIVILIAN licensed into automotive / edge platforms. Core patents retained; long tail of application-specific revenue.
Bull · Strategic sale
$1.2B
Competitive bid environment — defence and AI-infra acquirers both engaged. Demonstration prototypes unlock premium pricing. Achievable with 12–24 months of execution on the current plan.
Max · Full acquisition
$3B+
Convergence acquisition by hyperscaler or prime contractor seeking full freedom-to-operate across the AI-infra + defence-autonomy stack. Requires successful demo execution and favourable macro timing.
Acquirer Landscape

Buyer universe by patent fit

Acquirers are categorised rather than named specifically at this stage to preserve negotiating optionality. Named-target engagement occurs under NDA following mutual interest confirmation.
#1 AI infrastructure platforms
$500M – $2B
Highest technical fit. Major GPU manufacturers, hyperscale cloud providers, and photonic AI platforms have public roadmaps addressing thermal density, inference efficiency, and optical compute — the three areas where this portfolio's AI-infra patents are most directly relevant. These buyers have the largest deal capacity.
Best fit: DEANVPPNNSMC
#2 Defence primes & EW specialists
$400M – $1.5B
Active procurement window. Counter-UAS, directed-energy, and perimeter-security programmes are scaling across allied-nation defence primes. Passive, EMP-immune, zero-power classification is being actively specified in recent RFPs. Deal timelines are faster than commercial because of the current geopolitical backdrop.
Best fit: DRONEIOCCIVILIAN
#3 HBM & packaging manufacturers
$150M – $600M
Thermal ceiling is a known constraint. HBM manufacturers, advanced packaging foundries, and datacentre cooling infrastructure companies all face the same thermal density problem. DEAN is the most direct commercial fit. Shortest time-to-close of any buyer category in the portfolio.
Best fit: DEAN
#4 Automotive & robotics platforms
$200M – $800M
EU AI Act creates forced-buyer dynamics. Automotive OEMs, Tier 1 ADAS suppliers, humanoid robotics platforms, and edge-AI silicon vendors face a fully-applicable regulatory deadline in August 2026. Safety-layer IP with auditability properties addresses a compliance need, not just a performance improvement.
Best fit: IOCCIVILIAN
Execution

Monetisation roadmap

Sequenced to build leverage: grant events unlock demonstrations, demonstrations unlock premium pricing, defence interest creates urgency for AI-infra buyers. Administrative deadlines are the highest-priority items.
Filed
SMC, VPPNN, DEAN, IOC, DRONE, CIVILIAN, Dean-PUF
Seven patents filed across Australian priority applications. Portfolio architecture designed for interlocking commercial value: memory → compute → cooling → safety → classification.
Jan 2026
DEAN examination report received
Examination Report No. 1 issued. Response package (20 claims, amended) prepared. Addresses Section 40 clearance plus novelty/inventive step arguments. Clean response path identified.
Now · Q2 2026
Administrative priority items
Priority actions: DEAN excess-claims fee restoration and response lodgement; IOC PCT filing before 12-month priority window lapses; formalities-direction response on related filing. All administrative items are gating substantive strategy work.
H1 2026
Freedom-to-operate analysis
Formal FTO opinion commissioned against relevant prior art families. Deliverable used in acquirer conversations to pre-empt due-diligence questions and strengthen negotiating position.
H2 2026
DRONE demonstration build
Optical bench demonstration of defence classification embodiment. Patent value without working demo is materially lower than patent-plus-demo value — this step unlocks premium pricing for the Tier 1 defence conversation.
H2 2026
Initial acquirer outreach · Round 1
Defence-first sequence: EW/photonic defence primes, then diversified primes. AI-infra outreach (HBM manufacturers, photonic platforms) timed to coincide with DEAN examination progress. Target: non-binding LOIs by year-end.
2027
DEAN grant event
Expected grant of DEAN as first granted patent in the portfolio. Material re-rating event for the full portfolio valuation. Enables licensing conversations with HBM and CPO buyers in parallel with defence sale process.
2027–2028
Strategic transaction window
Base-case transaction window. Portfolio maturity, demonstration assets, and acquirer procurement cycles align. Bull case: competitive auction completes earlier if defence demonstration wins a Tier 1 evaluation.