RAIN RFID Standards After Gen2V2: File Mode, Arietta and the Bands That Come Next

Standards work is usually presented as a timetable of future features. That framing is wrong for RFID, because the most consequential capabilities are already published and sitting unused. The real question for a deployment in 2026 is not what is coming, but what has been available since 2015 and was never switched on.

Who actually writes the standard

The industry body and the standards body are different organisations, and confusing them leads to bad planning.

• The air interface belongs to the GS1 EPCglobal Gen2 family. The RAIN RFID Alliance does not write it.
• The alliance, founded in 2014 by Google, Impinj, Avery Dennison and Smartrac, has grown to more than 200 members.
• Its work is interoperability certification, use-case promotion and spectrum coordination — the parts that decide whether products from different vendors actually talk to each other.
• Its annual Plugfest is the practical test. The 2025 edition covered 12 chip vendors, 18 reader vendors and 9 antenna vendors, with cross-vendor read success reaching 99.2 % for the first time.
• For a buyer, the useful reading is that certified interoperability is now good enough to rely on — but only for products that actually carry the certification.

Gen2V2 File Mode: the tag becomes a file system

A conventional Gen2 tag gives you one EPC bank and a block of user memory. File Mode changes the data model, not the radio.

• File Mode supports up to sixteen independent files on a single tag.
• Each file carries its own access password and its own permission set, so read and write rights can differ per file.
• One file can hold public product data readable by anyone, while another holds encrypted lot data readable only by a specific partner.
• The same physical tag therefore serves several parties without any of them seeing each other's data.
• This is the feature that makes a tag usable in a multi-tenant supply chain, and it has been in the standard since 2015.

BlockPermalock and the trust boundary

Locking in legacy Gen2 was all-or-nothing: once user memory was locked, it was locked. BlockPermalock introduces granularity, and granularity is what creates trust.

• Permalock applies per 32-bit block, so a tag can be partly immutable and partly writable.
• A locked block is irreversible and readable without a password, which makes it a credible attestation.
• An upstream shipper can lock the lot number so no downstream node can alter it, while keeping its own status blocks writable.
• The audit consequence is significant: tampering becomes detectable rather than merely suspected.
• Most deployments still do not use it, which means they are paying for a security capability and not deploying it.

AES-128 and regulated use cases

Encryption moved RFID from identifier to token, and that shift is what opens regulated sectors.

• Gen2V2 added AES-128 cryptographic instructions alongside the legacy authentication scheme.
• The tag can prove it is genuine and the reader can prove it is authorised, without exchanging a static secret in the clear.
• Pharma traceability, high-value asset custody and anti-counterfeit programmes are the natural first users.
• The limit is operational: key management is now part of the deployment, and most integrators are not set up to run it.
• Plan the key lifecycle before the pilot. A cryptographic tag with a lost key is a very expensive passive label.

Arietta: sub-centimetre positioning from phase

Project Arietta, launched by the alliance in 2024, attacks the problem that has limited RFID in automation: it can tell you a tag is in the room, not where in the room.

• Arietta stays inside the existing UHF band rather than moving to a new spectrum, which is what makes it deployable.
• It uses backscatter phase ranging across multiple fixed readers rather than received signal strength.
• Cooperative localisation resolves the phase difference of one tag's backscatter as seen by several readers simultaneously.
• The theoretical accuracy is around five millimetres, against one to three metres for strength-based localisation.
• The catch is infrastructure: three or more readers with known geometry, tightly synchronised, which is a different cost profile from a portal.

Arietta anti-collision: from a thousand to several thousand tags

The second Arietta proposal addresses the limit that appears in every dense deployment.

• Gen2 uses slotted Aloha, and a single reader tops out at roughly a thousand tags per second in theory.
• Arietta layers subcarrier frequency division on top of the slot structure, creating time and frequency diversity at once.
• The target is several thousand tags identified per second by one reader.
• The scheme also targets ghost reads and missed reads, which are the real cost in a dense room.
• These are research targets, not shipping specifications. Treat vendor claims about Arietta performance as projections until silicon exists.

Spectrum: harmonising Asia-Pacific

Spectrum, not silicon, is what most often blocks a cross-border rollout.

• China uses 920–925 MHz, Japan 916–921 MHz and Korea 917–923 MHz, so a tag designed for one market may not be legal in the next.
• Cross-border logistics operators have had to specify tags and readers covering the full band, which raises cost and reduces choice.
• The alliance is coordinating the three markets toward 920–925 MHz, with legislation expected around 2026.
• Until that lands, multi-country deployments need per-country power and channel configuration, documented in the asset register.
• A reader configured for Europe will not legally run at full power in Asia, and the firmware default is not a compliance strategy.

5.8 GHz: a complement, not a successor

Microwave RFID is often discussed as the next band. It is better understood as a different tool.

• 5.8 GHz offers wider bandwidth, which supports higher positioning accuracy and millimetre-level localisation on automated lines.
• Read range is short, roughly half a metre to one metre, so it is a proximity technology.
• Tags cost far more than UHF inlays, which rules out item-level retail economics.
• The realistic pattern is coexistence: UHF for inventory and logistics, 5.8 GHz for precise station-level control.
• Anyone planning to replace UHF with 5.8 GHz has misunderstood the cost curve.

What integrators should do now

The practical roadmap is mostly about using what exists, then preparing for what is funded.

• Audit which Gen2V2 features your current tags and readers already support, and turn on per-block locking where lot integrity matters.
• Separate public and restricted data into files rather than putting everything in one user-memory block.
• Treat Arietta as a reader and middleware requirement in your 2028 refresh, not as a reason to delay a 2026 project.
• Document the spectrum configuration per country now, so expansion is a configuration change rather than a redesign.
• Ask vendors for Plugfest evidence and certification numbers, not for a specification sheet.
• Keep the tag as the stable element and let readers, middleware and spectrum carry the change.

The pattern across all of these is consistent: the standard has been ahead of deployment practice for a decade. The organisations that gain in the next cycle are the ones that read the published clauses, deploy them, and keep the radio layer boring.