TIGHT Patent Portfolio v3
Three families. Thirteen claims. AUD $300 for the whole of Q4 2026.
TIGHT is a deliberate compression, not a breadth play. The earlier portfolio counted thirty-two internal drafts across a wide surface. It was a worse portfolio: thin claims spread over too many families, at a filing cost that bought less coverage per dollar than a smaller number of well-formed claims in fewer families. Compressing it to three families and thirteen claims roughly tripled the claim density per dollar, and it let the P11 parent absorb three former applications instead of paying to file them separately.
The three families
One parent carries the bulk of the portfolio. Two stay separate because they sit in different layers and have different prior art — folding them in would have weakened all three, not strengthened the total.
P11PARENTL2-L4MMQCR — Multi-vendor Measurement & Quantum Cross-compilation Runtime
Everything that folds into the parent family lands here. P11 was filed at IP Australia on 13 October 2026 and carries the QEC addendum (2 Nov 2026) and the L2 continuation-in-part (13 Nov 2026). P17, P18 and P19 were originally separate applications; they are now folded into a single Q1 2027 continuation-in-part rather than three more filings, because three narrower filings would have cost three times as much and covered less.
P16SeparateL1Cryo-CMOS control and readout
Held separate because it sits in a different layer. P16 claims control electronics and readout at the physical layer, and folding a cryogenic CMOS claim into a compilation-runtime parent would have weakened both. It is a distinct invention with a distinct prior-art base.
T5.4-QECSeparateL7QEC metric normalisation
Held separate because normalisation of QEC metrics is a discovery-layer invention, not an execution-layer one. It is also the invention with the cleanest trigger condition: it becomes separately protectable once the normalisation method is implemented well enough to file, which is a Q1 2027 decision point rather than a Q4 2026 one.
Filing calendar
Every date below is either filed or scheduled. Provisionals first, always.
P11 MMQCR — parent provisional
Filed. Document 22915860013, IP Australia.
P11-QEC addendum
Adds the QEC normalisation claims to the parent family.
P11 L2 continuation-in-part
No incremental fee — the CIP rides on the parent's priority date.
P16 Cryo-CMOS
Separate filing, distinct layer, distinct prior art.
P11 CIP folding in P17, P18, P19
Three former applications become one continuation-in-part.
T5.4-QEC-Normalize
Triggered by the normalisation method reaching a fileable state.
Total for Q4 2026: AUD $300. Total if the two Q1 2027 filings trigger: AUD $500.
Layer coverage
Coverage is concentrated where the method is. L2 is where the invention actually lives, and it is the most protected layer. Overall coverage across L2 to L5 is approximately 75 percent.
P16 carries the control and readout claims. Broker-level hardware abstraction is not claimed.
Cross-vendor compilation, measurement normalisation, and execution scheduling are the family's core.
Vendor dispatch, error budgeting and multi-backend supervision.
The language and its unit system. The compiler implementation itself stays Apache 2.0.
The API surface is claimed where it encodes the orchestration method, not where it is a conventional interface.
Provenance and signing claims only. No general security claims are made.
T5.4-QEC-Normalize carries the metric-normalisation claims once filed.
Filing venue
Everything is filed at IP Australia, at AUD $100 per provisional filed online. The move away from the USPTO was a cost decision, not a quality decision. The same claims in the same sequence would be materially more expensive to file across two offices, and at this stage of the company the money is better spent on evidence generation than on jurisdiction spread.
International filing is deliberately deferred until a claim has an N3 or N4 evidence tier behind it. A provisional filed before the evidence exists is a provisional filed early.
What this portfolio does not cover
Every portfolio has gaps and pretending otherwise is how a portfolio ends up being valued on its line count instead of its enforceability. This one does not claim: general quantum error correction, quantum algorithm correctness, hardware fabrication, cryogenic control electronics beyond P16's scope, calibration algorithms, or any claim on the physics of the underlying devices. It claims a method for orchestrating and auditing across vendors.
It also does not claim the compiler implementation, which stays Apache 2.0 and open. The method claims and the code are deliberately separated, which is what makes the open-source strategy work at all.
Patent Pending
For the full strategy document — the portfolio in context, the seven-layer architecture, the CERN co-development thesis, and the funding pipeline — the canonical brief is the strategy PDF.