US 8448855

I'll verify current docket/litigation status for US 8,448,855 while cross-checking the patent's own record.

Let me check specifically for any 2026 Federal Circuit activity and the current status of the district court cases.

Let me do one final check on recent 2026 appellate activity involving RFCyber.

US Patent 8,448,855 — Summary

Note on sources and scope. I ran live searches against Google Patents/FreePatentsOnline/Justia (the patent's own text), PTAB/docket records, and litigation-intelligence coverage. The patent-number convention below is taken literally from the record; I did not normalize any identifiers.

Bibliographic data

⚠️ Minor discrepancy to flag: Google Patents lists the priority date as 2006-09-24, filing date 2012-02-18, issue date 2013-05-28. A third-party aggregator (Unified Patents) lists 2006-09-23 / 2012-02-17 / 2013-05-27 (one day earlier across the board). The face-of-patent text reproduced in court exhibits confirms filed Feb. 18, 2012, issued May 28, 2013, which I treat as authoritative.

Abstract (verbatim)

"Techniques for funding an electronic purse (e-purse) are disclosed. According to one aspect of the invention, a mechanism is provided to enable a portable device to conduct transactions over an open network with a payment server without compromising security. In one embodiment, a device is loaded with an e-purse manager. The e-purse manager is configured to manage various transactions and functions as a mechanism to access an e-purse therein. The e-purse is funded by interactions among the e-purse manager, a payment server and a financial institution (its server) that maintains an account therefor."

Plain-language claim overview

The patent has 17 claims with two independent claims: claim 1 (device side) and claim 9 (server side). All others (2–8, 10–17) depend from one of these.

Independent Claim 1 — "Funding from the device's point of view."
The portable device must be an NFC-enabled device containing a card module. The method:

1. Receive a PIN from the user;
2. After the PIN is verified, a midlet (an app on the device) sends a request to an e-purse applet (residing in the card module);
3. The e-purse applet composes a response to that request;
4. The e-purse applet sends the response over a wireless network to a server that administers the e-purse; that server verifies the response against an account at a financial institution and, if verification succeeds, initiates a fund transfer request to the institution;
5. Receive the commands back from the server; and
6. The emulator on the device updates a transaction log after the e-purse applet verifies the authenticity of those commands.

Critically, all of this is only covered where the e-purse was previously personalized by two steps:

• establishing an initial security channel between the card module and an e-purse SAM external to the card module to install and personalize the e-purse applet, and
• creating a second security channel "on top of" the initial security channel to protect subsequent operations — with all later transactions conducted over that second channel.

Independent Claim 9 — "Funding from the server's point of view."
This is the mirror-image method performed by the server side:

1. The server receives a request from the portable device (where the request was composed by an e-purse applet after the applet received an initial request from the midlet and after a PIN was entered and verified);
2. The server verifies the request against a bank account across a network;
3. If verified, the server initiates a fund transfer request with the financial institution administering the e-purse;
4. The server sends commands to the portable device that cause an emulator to update a transaction log — but here the authenticity of the commands is verified by the midlet in the device (note: claim 1 says the applet verifies; claim 9 says the midlet verifies);
5. Same personalization predicate as claim 1: an initial security channel between card module and external SAM, and a security channel layered on top of it for subsequent emulator operations.

Representative dependent claims:

• 2 / 11: card module is a SmartMX (SMX) module pre-loaded with the emulator for storing secured values.
• 3 / 12: e-purse built on Global Platform, both e-purse keys and card access keys personalized into a tag.
• 4 / 13: security via a card manager; the e-purse applet acts as gatekeeper.
• 5 / 14: e-purse implemented in the emulator; 6 / 15 adds access via transformed passwords based on access keys.
• 7 / 16: commands are network messages containing APDU commands, which are extracted from the messages; 8 / 17: commands include the financial institution's response.

Post-grant and litigation status (as verified)

• IPRs (all failed against the '855):
  • IPR2021-00954 (Google LLC / Google Payment Corp.) — terminated by settlement before institution (joint motion granted Oct. 2021).
  • IPR2021-00978 (Unified Patents) — institution denied on the merits.
  • IPR2022-01241 (Apple Inc.) — institution denied (Paper No. 7, Jan. 23, 2023).
• Ex parte reexaminations — claims survived intact, twice:
  • C1 certificate, effective April 28, 2023 — all original claims 1–17 confirmed with no amendments.
  • C2 certificate, effective March 7, 2025 — claims again confirmed.
• District court campaigns: '855 asserted against Google, Samsung, and LG (E.D. Tex., 2:20-cv-00274/00335/00336); Apple (W.D. Tex. 6:21-cv-00916); Visa (W.D. Tex. 6:22-cv-00697, dismissed by joint motion Feb. 2024); and a 2024 E.D. Tex. wave against Costco, Kroger, Shell (2:24-cv-00549, dismissed with prejudice), Starbucks, Volkswagen and Walmart.
• Maintenance fees: 4th (2016), 8th (2020), and 12th year with late surcharge (Jan. 28, 2025) all paid. Google Patents projects anticipated expiration Sept. 24, 2026 (i.e., 20 years from the 2006 priority).

⚠️ CAFC 2026 docket — what I could and could not confirm

I could not find any Federal Circuit 2026 docket or opinion specifically addressing US 8,448,855. My searches of Federal Circuit activity involving RFCyber returned appeals concerning related but different patents:

• Fed. Cir. 2023-2418 — RFCyber v. Coke Morgan Stewart (Intervenor), decided Aug. 14, 2025, affirming PTAB's unpatentability of US 9,240,009 (the '009 patent), a continuation-in-part of the '218 family.
• Fed. Cir. 24-1612 and 24-1614 — Rule 36 affirmances involving US 10,600,046 and US 11,018,724, respectively (again, not '855).

Uncertainty statement: I do not have authoritative confirmation of a 2026 CAFC appeal for the '855 patent specifically. It is possible the '855 patent has not been the subject of a Federal Circuit appeal (consistent with all IPRs against it being denied or settled, and the reexaminations confirming the claims), or a 2026 docket may exist that my search did not surface. The related family patents (notably the '009) have been invalidated on appeal, which matters for the family's overall value even though the '855 itself remains confirmed — a point worth distinguishing carefully in any analysis of this asset.

I'll research the litigation history for US Patent 8448855 using live sources.

Let me dig deeper into the specific cases involving the '855 patent.

Litigation Report — U.S. Patent No. 8,448,855

Patent: U.S. 8,448,855 B1, "Method and Apparatus for Funding an Electronic Purse" (app. 13/400,038; filed Feb. 18, 2012; granted May 28, 2013; priority Sept. 24, 2006 via continuation of 11/534,653, now U.S. 8,118,218). Owner/enforcer: RFCyber Corp. (copyright records / IPR filings list the patent as assigned to Shenzhen RFCyber Asset Management, LLP, with RFCyber Corp. as exclusive licensee). Continuation sibling: U.S. 9,189,787.

Note on scope: The '855 patent is one of a five-patent RFCyber contactless-payment family (the '218, '855, '787, '009, and '046 patents). RFCyber has generally asserted them as a group against each defendant. I identified each case below from the record; where I could not confirm that the '855 patent specifically was asserted (as opposed to a sibling), I say so. I did not auto-correct any case numbers or identifiers.

A. District Court Cases

Cases where the '855 patent is confirmed as asserted

Cases involving the RFCyber family where I could NOT confirm the '855 patent specifically

The Google Patents family record (family ID 39223872) additionally lists the following as family litigation, but I could not verify from the retrieved records that the '855 patent specifically (as opposed to the '218, '787, '009, or '046 patents) was asserted in each:

• RFCyber Corp. v. Apple Inc., 6:21-cv-00916 — W.D. Tex. (Judge Albright), filed ~Sept. 2021. Apple's IPR IPR2022-01241 did target the '855 patent (see below), and RFCyber's Visa briefing refers to "a pending case against Apple before this Court." Whether 6:21-cv-00916 itself asserted the '855 patent is not confirmed by the records retrieved.
• RFCyber Corp. v. Apple Inc., 1:23-cv-00661-ADA — W.D. Tex., filed 2023. Referenced in RFCyber's Visa briefing as a pending Apple case involving the same family.
• 6:23-cv-00708 — W.D. Tex., filed 2023. Defendant and asserted patents not identified in the records retrieved.
• 2:24-cv-00548, 2:24-cv-00550, 2:24-cv-00551, 2:24-cv-00576 — E.D. Tex., filed 2024. Defendants and asserted patents not identified in the records retrieved. These appear to be part of the same mid-2024 RFCyber campaign as the Costco and Shell suits.

B. PTAB (Inter Partes Review) Proceedings

Samsung filed companion IPRs (IPR2021-00979, -00980, -00981) against the sibling '218, '787, and '009 patents; the '855 petition (‑00978) was denied institution, as was Apple's '855 petition.

C. Post-Grant Reexaminations

• Ex parte reexamination filed Jan. 14, 2022 (third-party requester associated with Greenberg Traurig); reexamination certificate C1 issued — decision confirms the claims, effective Apr. 28, 2023.
• Second ex parte reexamination filed July 17, 2023; reexamination certificate C2 issued — decision confirms the claims, effective Mar. 7, 2025.
• (The 2025 Law360 item "Fed. Circ. Won't Revive RFCyber Mobile Data Patent," Aug. 14, 2025, affirms a PTAB obviousness finding; based on the record retrieved this appears to concern a different RFCyber patent, not the '855 patent. I could not confirm which sibling patent it addresses.)

D. Summary

• Confirmed '855-asserting litigation (6 district court cases): Google (2:20-cv-00274), Samsung (2:20-cv-00335), LG (2:20-cv-00336), Visa (6:22-cv-00697), Costco (2:24-cv-00546), Shell (2:24-cv-00549).
• Outcomes: No case reached a merits judgment on the '855 patent. LG and Shell were dismissed with prejudice; Costco without prejudice; Visa dismissed by joint motion; Google/Samsung resolved (settlement indicated).
• PTAB: Three IPRs against '855 — Google (settled pre-institution), Samsung (institution denied), Apple (institution denied). No institution on the '855 patent.
• Patent status: Active; the '855 patent survived two ex parte reexaminations with claims confirmed (certificates C1 and C2). Anticipated expiration is docketed as Sept. 24, 2026.
• Unverified: The specific defendants/patents in 6:23-cv-00708 and in E.D. Tex. 2:24-cv-00548, ‑00550, ‑00551, and ‑00576, and whether the two Apple district cases asserted the '855 patent specifically.

Sources: Google Patents record for US8448855B1 (family litigation list); Unified Patents litigation/PTAB portal (portal.unifiedpatents.com) case pages; PACER dockets via docketalarm.com (e.g., 2:20-cv-00274, 2:20-cv-00335, 2:20-cv-00336, 6:22-cv-00697, 2:24-cv-00546); PTAB IPR papers (IPR2021-00978 Paper 10; IPR2022-01241 Paper 7); USPTO AO 120 filings; PatSnap and RPX litigation summaries.

If you want, I can pull the remaining unidentified 2024 E.D. Tex. defendants (‑00548, ‑00550, ‑00551, ‑00576) and the 6:23-cv-00708 defendant directly from PACER/Unified Patents to complete the table.

I'll research these proceedings before writing. Let me start with the PTAB dockets.

Now let me dig into each proceeding's specifics.

Let me get the institution decisions and grounds.

Important framing note before the report

The structured "PTAB proceedings on file" block I was handed says the USPTO ODP returns no AIA trial proceedings for US 8,448,855. That is wrong for this patent. The authoritative patent record for US 8,448,855 B1 (Google Patents, which mirrors the ODP/PTAB litigation metadata) lists three AIA trial proceeding numbers against this patent, and I independently corroborated each one against Docket Alarm dockets and PTAB papers served on the OCR mirrors:

I therefore depart from the "no PTAB activity" default and flag it as an ODP ingest gap. What I could not retrieve within my research budget is the verbatim text of the two institution decisions — every factual statement below is sourced; every gap is labeled as a gap rather than filled in.

Proceedings overview

Total: 3 AIA trials. Zero claims invalidated. Zero Final Written Decisions. Zero Federal Circuit appeals. One settled pre-institution (Google), two institution denials on the merits (Samsung, Apple). Bottom line for a defendant: the '855 patent has never been substantively tested at the PTAB and has emerged un-narrowed — all 17 claims stand as issued. But it went three-for-three at the Board only because no panel ever reached the merits, and it has since been twice re-confirmed in ex parte reexamination (reexam certificates C1 and C2). If the patent owner's demand letter leans on a PTAB record, there is no PTAB record to lean on — but conversely, you get no free claim cancellations either. Combined with a term that appears to have run on 2026-09-24, the practical exposure is back damages only.

IPR2021-00978 — Samsung Electronics America, Inc. v. RFCyber Corp.

• Type: Inter Partes Review (35 U.S.C. §§ 311–319)
• Filed: 2021-06-08
• Status: Not Instituted — Merits (structured data); Docket Alarm reflects Institution Denied, terminated 2021-12-14.
• Judge panel: Administrative Patent Judges Kevin W. Cherry, Kristi L. R. Sawert, and Patrick R. Scanlon (Docket Alarm docket, https://www.docketalarm.com/cases/PTAB/IPR2021-00978/Samsung_Electronics_America_Inc._v._RFCyber_Corp/). Note the panel is identical to the panel that handled Google's RFCyber set — the Board grouped the RFCyber campaign before one trio.
• Petition grounds: Not retrieved. I could not pull the petition or the decision paper, so I am not going to name art or statutory subsections. The '855 claims RFCyber was actually asserting against Samsung in the parallel E.D. Tex. case (No. 2:20-cv-00274-JRG) were claims 1–6, 10, and 12 (RFCyber's own submission, quoted in the court's Claim Construction Memorandum and Order), and the one-year § 315(b) clock means the petition would have targeted those or a superset.
• Institution decision: Denied, 2021-12-14. The structured record classifies the denial as merits-based ("Not Instituted – Merits"), as opposed to a discretionary § 314(a)/§ 325(d) denial or a settlement dismissal — i.e., the Board was not persuaded the petition established a reasonable likelihood of prevailing on at least one challenged claim. I could not verify the panel's verbatim reasoning; treat the "merits" label as the metadata's characterization and pull the paper from PTAB E2E (https://ptab.uspto.gov) before relying on it. For context, this denial issued during the Fintiv discretionary-denial era with trial in E.D. Tex. then set for March 2022.
• Final Written Decision: None. No FWD was ever issued. No claim was canceled, confirmed, or held patentable in this proceeding.
• Settlement / termination: Not a settlement — the petition was dismissed on non-institution.
• Appeal: None. There is nothing to appeal where institution is denied, and § 314(d) makes the non-institution decision non-appealable.
• Defensive value: A prior merits-based refusal to institute is a modest quality signal you can cite to the patent owner ("your own petition-targeted claims survived a merits screen by Samsung"), but it is not a validity adjudication and carries no estoppel — Samsung is free to run the identical art in district court.

IPR2022-01241 — Apple Inc. v. Shenzhen RFCyber Asset Management, LLP

• Type: Inter Partes Review
• Filed: 2022-07-20
• Status: Not Instituted — Merits (structured data); Docket Alarm reflects Institution Denied, terminated 2023-01-23.
• Judge panel: Administrative Patent Judges Josiah C. Cocks, Kevin W. Cherry, and Patrick R. Scanlon (https://www.docketalarm.com/cases/PTAB/IPR2022-01241/Apple_Inc._v._Shenzhen_RFCyber_Asset_Management_LLP/).
• Patent owner of record: Shenzhen RFCyber Asset Management, LLP — not RFCyber Corp. This matters. The '855 patent was assigned to Shenzhen RFCyber on 2014-02-08 (USPTO assignment reel/frame 032177/0743), and RFCyber Corp.'s standing to sue on the '218 and '855 patents was actively litigated in E.D. Tex. (Samsung's motion to dismiss for lack of constitutional standing, Dkt. 81; RFCyber's opposition at Dkt. 92). A defendant today should confirm the current chain of title before conceding the plaintiff's standing — note that the "Current Assignee" field on the patent record now also lists Rich House Global Tech Ltd.
• Petition grounds: Not retrieved. Same caveat as above — no art, no statutory subsections asserted here.
• Institution decision: Denied, 2023-01-23, again classified in the structured record as merits-based ("Not Instituted – Merits"). Verbatim reasoning not retrieved.
• Final Written Decision: None.
• Settlement / termination: Not a settlement; dismissal on non-institution. (Do not confuse this with Apple's separate PGR2022-00003 v. RFCyber Corp., which was denied institution on 2022-03-24 and relates to a different RFCyber patent family member — not the '855 patent.)
• Appeal: None.
• Defensive value: Same posture as Samsung — two separate sophisticated petitioners put the '855 patent in front of the Board and got nowhere on the merits, but neither produced an invalidity judgment. Cite it as a screening signal, not as a defense.

IPR2021-00954 — Google LLC v. RFCyber Corp.

• Type: Inter Partes Review
• Filed: 2021-05-19
• Status: Settlement (structured data) — terminated prior to institution by Board order granting the parties' joint motion under 35 U.S.C. § 317 and 37 C.F.R. § 42.74.
• Judge panel: Administrative Patent Judges Patrick R. Scanlon, Kevin W. Cherry, and Kristi L. R. Sawert (opinion author). The termination order covered four companion Google petitions in one order: IPR2021-00954 ('855), IPR2021-00955 ('787), IPR2021-00956 ('009), and IPR2021-00957 ('218).
• Petition grounds: Moot and never adjudicated. The Board expressly noted the proceedings were "in a preliminary stage, and no decisions on whether to institute any trial have been issued." No claims were specified in an institution decision because none issued.
• Institution decision: No institution decision — case dismissed before the deadline. The Board's reasoning was purely procedural: "The Board generally expects that a case 'will terminate after the filing of a settlement agreement, unless the Board has already decided the merits,'" citing the Office Patent Trial Practice Guide, 77 Fed. Reg. 48,756, 48,768 (Aug. 14, 2012) and 37 C.F.R. § 42.72.
• Final Written Decision: None.
• Settlement / termination: Joint motion filed 2021-10-19. The Board granted termination and dismissed Google's petitions under 37 C.F.R. §§ 42.5(a) and 42.71(a). The parties simultaneously settled the parallel E.D. Tex. litigation, RFCyber Corp. v. Google LLC and Google Payment Corp., No. 2:20-cv-00274-JRG, and represented they "do not anticipate further litigation between them concerning" the challenged patents. A true, unredacted copy of the settlement and license agreement was filed as Ex. 1040 and, on joint request, was granted business-confidential treatment under 35 U.S.C. § 317(b) and 37 C.F.R. § 42.74(c) — so the terms (including any license scope or royalty) are not public. Google's counsel: Haynes & Boone LLP (Andrew S. Ehmke, Michael Parsons, Vinu Raj). RFCyber's counsel: Fabricant LLP (Vincent J. Rubino, III; Peter Lambrianakos; Alfred R. Fabricant; et al.).
• Appeal: None.
• Defensive value: The lowest-information proceeding of the three. No merits signal, no estoppel, and a confidential license. Two things to check: (i) whether you are a Google affiliate or privy (the license may run to Google and its privies only), and (ii) whether the '855 patent is encumbered by that license in a way that affects standing or damages. Absent that, Google's settlement tells you only that Google preferred a license to a fight.

Strategic summary

Claim-level status of US 8,448,855. All 17 claims — independent claims 1 and 9, and dependents 2–8 and 10–17 — are UNSUSTAINED-AND-UNCANCELED: no claim has been CANCELED by the PTAB, and no claim has been HELD PATENTABLE by the PTAB either, because no IPR ever instituted. There are no "surviving claims" in the IPR sense, because nothing was ever narrowed. What did happen is two ex parte reexaminations, both of which came out for the patent owner: per the patent's legal-events record, a request filed 2022-01-14 produced reexamination certificate C1 (effective 2023-04-28, "Reexamination decision confirms claims"), and a second request filed 2023-07-17 produced certificate C2 (effective 2025-03-07, "Reexamination decision confirms claims"). Secondary aggregator data (PatSnap) lists reexamination control numbers 90/015,260 and 90/020,144, the latter filed by counsel at Greenberg Traurig — treat the control numbers and requester identity as unconfirmed, since I could not verify them against the official reexamination record.

Estoppel landscape. This is the cleanest part of the picture: no IPR estoppel exists on this patent. Under 35 U.S.C. § 315(e)(2), estoppel binds a petitioner only in an IPR "that results in a final written decision." None of IPR2021-00954, -00978, or -01241 reached one, so Google, Samsung, and Apple are not estopped and may raise the same § 102/§ 103 grounds in the district court. Likewise, a new defendant faces no estoppel-forward constraint and is free to run any prior-art patent or printed publication under §§ 102/103. The real constraint on new art is § 325(d) / Becton, Dickinson: the reexamination requester's references for the '855 patent were squarely before the examiner and the claims were twice confirmed, so art that was "previously presented to the Office" invites discretionary denial. The opening is non-cumulative art the examiner never saw, plus § 112 and § 101 theories (which are unavailable in IPR entirely). Note also the § 315(b) one-year bar: defendants served in the 2020–2024 Texas campaigns are already time-barred from filing, and any newly served defendant must petition within one year of service.

Pattern signals. (1) Google filed a four-petition barrage across the RFCyber family ('855, '787, '009, '218) on the same day and then bought its way out of all four plus the district court case in one settlement — a coordinated, well-funded campaign, not a lone defendant. (2) Samsung and Apple each filed separately and both were shut out on the merits. (3) No defensive aggregator IPR: the Google petition was filed by Google and Haynes & Boone, not by Unified Patents; Unified appears in the record only as the source of the litigation metadata, not as a petitioner. If a third party wanted the '855 patent tested, it had to use ex parte reexamination (the 2022 and 2023 requests) rather than an AIA trial. (4) No Federal Circuit appeal exists on this patent — nothing to appeal. (5) The patent owner, RFCyber/Fabricant LLP, was indeed aggressive on the enforcement side, running a broad campaign across E.D. Tex. and W.D. Tex. (including 2:20-cv-00274, 2:20-cv-00335, 2:20-cv-00336, 2:24-cv-00546 through 2:24-cv-00551, 2:24-cv-00576, 6:21-cv-00916, 6:22-cv-00697, 6:23-cv-00708, 1:23-cv-00661), and at least one 2024 case was dismissed without prejudice (2:24-cv-00546 v. Costco, filed 2024-07-18, closed 2024-10-23) — the tell-tale footprint of a license-and-dismiss monetization model.

The timing point that dominates everything. The structured record lists anticipated expiration 2026-09-24 — twenty years from the 2006-09-24 filing date of parent application 11/534,653, with the 12th-year maintenance fee paid 2025-01-28. That date is four days before today's date (2026-09-28). If that term holds, the '855 patent is expired, which caps your exposure at back damages inside the 35 U.S.C. § 286 six-year lookback (i.e., infringement from roughly 2020-09-28 forward) and eliminates prospective injunctive relief. Verify this before building a defense around it: confirm via USPTO PatentCenter whether any patent term adjustment extended the term past 2026-09-24, and note that the "Active" legal-status flag on the patent record reflects the last maintenance-fee payment, not remaining term. Related family members carry longer stated terms (the '218 patent is listed as active through 2029-08-16), so do not assume the family is dead just because the '855 term has run.

Recommended next steps

• Pull the two denial decisions verbatim. I did not retrieve the text of the IPR2021-00978 (2021-12-14) or IPR2022-01241 (2023-01-23) institution decisions, so no claim number, reference, or § 102/§ 103/§ 112 ground in this report should be attributed to those panels. Get Paper 10 (or equivalent) from PTAB E2E: https://ptab.uspto.gov — docket links: https://portal.unifiedpatents.com/ptab/case/IPR2021-00978, https://portal.unifiedpatents.com/ptab/case/IPR2022-01241, https://portal.unifiedpatents.com/ptab/case/IPR2021-00954. This matters because a merits-based denial is a much stronger defensive signal than a Fintiv discretionary denial, and the metadata's "Merits" label is the only thing supporting the former.
• Confirm the term. Run PatentCenter for US 8,448,855 to check PTA and the maintenance-fee history. If expiration is confirmed at 2026-09-24, your damages case is a § 286 back-damages case, and any demand letter implying ongoing infringement is overreaching.
• Fix the ODP gap. The ODP-derived "no AIA trial proceedings" finding is incorrect for this patent. Re-query with the patent number rather than the application number (13/400,038) and re-check whether the three proceedings were delisted or orphaned in the ingest.
• Map the reexamination record. Order the file histories for the two ex parte reexaminations (C1 and C2) from PatentCenter. Those references are the § 325(d) "already before the Office" set — i.e., exactly the art you don't want to rely on in a future IPR. Everything not in those two files is your white space.
• Check the Google license. The Google settlement-and-license agreement (Ex. 1040 in IPR2021-00954) is business-confidential and undisclosed. If you are a Google customer, OEM, or privy, exhaust your license defense before spending on invalidity.
• Confirm chain of title and standing. The '855 patent was assigned to Shenzhen RFCyber Asset Management, LLP in 2014, while RFCyber Corp. asserted it as exclusive licensee; the "Current Assignee" field now also lists Rich House Global Tech Ltd. Samsung already litigated (and RFCyber had to brief) the Article III standing question on the sibling '218/'787/'046 patents. That attack is worth re-running on this patent.
• Milestones: none. There are no pending proceedings, no institution-decision deadlines, no oral hearings, and no FWD due dates, because all three petitions were dismissed or denied before institution and no party appealed. If the patent is expired, a fresh IPR is both time-barred in most scenarios under § 315(b) and strategically pointless.
• Sanity check on entity involvement: I found no PTAB petition on the '855 patent filed by Unified Patents or any other defensive aggregator. Unified's name appears in the structured metadata only as a data source, not as a petitioner. Any representation to the contrary should be verified against the three docket numbers above.

I'll research the assignment record for US 8,448,855, building on the bibliographic and litigation findings already established.

Strong initial leads. Let me dig into the correspondent of record on those reel/frames and check for later recordings.

Let me pin down the correspondent of record, the Chinese entity, and the parties' current status.

Assignment & Ownership Analysis — US 8,448,855 B1

Cross-reference note: This builds on the previously generated summary. That section reported the two 2014-02-08 legal events from Google Patents and the post-2020 litigation-entity confusion. This section reconstructs the chain with reel/frame specificity and adds the Chinese corporate-registry evidence for the assignee. One item in the prior section is now confirmed more precisely: the prior section said only that assignees were "RFCyber Corp." and, later, "Shenzhen RFCyber Asset Management LLP"; the litigation record now shows these were distinct legal persons (California "RFCyber Corporation" vs. the Texas plaintiff "RFCyber Corp."), which matters for standing and ownership analysis.

Inventors

All four residences are in the Fremont/Milpitas/San Jose corridor, matching RFCyber Corp.'s then-address at 4160 Technology Drive, Suite A, Fremont, CA 94538 (per RFCyber product collateral). The application's assignee-data form (SB/29) named Rich House Global Technology Ltd., Shenzhen, China and RFCyber Corp., Fremont, CA, confirming the inventors were working for the RFCyber/Rich House group at filing.

Unusual pattern — present, but not the usual one. The inventors' assignment to RFCyber Corp. was executed 2006-09-23 — one day before the 2006-09-24 priority filing — but was not recorded at USPTO until 2014-02-08 (Google Patents legal event, Reel 032177/0739, effective date 20060923). That is a ~7.4-year recording lag, and the recordation lands on the same day as the transfer to the Chinese entity. This is not an "inventors flee within 12 months" pattern; it is a chain-creation-at-transfer pattern: the foundational inventor→company assignment was papered onto the record only at the moment the company needed a clean chain to move the asset offshore. Note also that the four inventors are the same four across the entire family ('218, '855, '787, '009, '046), i.e., a single small founding team, not a rotating R&D staff — consistent with the "inventor-backed plaintiff" characterization RPX has applied to RFCyber (RPX Insight, July 26, 2024).

Original assignee

Assignees on the face of the issued patent (both): Rich House Global Technology Ltd., Shenzhen (CN) and RFCyber Corp., Fremont, CA. Confirmed verbatim in the patent PDF attached as Exhibit A to the 2024 complaints (RFCyber Corp v. Costco Wholesale Corp., E.D. Tex. 2:24-cv-00546, Dkt. 1-1, filed 2024-07-18).

• Rich House Global Technology Ltd. — Shenzhen entity, apparently an affiliate of the RFCyber group. Concrete link: the Shenzhen partnership that later took the '855 was registered with the contact e-mail portia.zhu@richhouseglobal.com (企知道/企查 registry extraction for 深圳市可秉资产管理合伙企业（有限合伙）). I could not verify this company's current operating status, capitalization, or product activity — state it explicitly as unverified.
• RFCyber Corp. (Fremont, CA) — an operating company. It shipped a real product line: "Mobile Financial Service" (ePurse based on PBOC 2.0), "Trusted Service Management Platform," and "Mobile POS," with claimed deployments in Breda, The Netherlands (Social Care e-Voucher, Aug. 2007) and the Chunghwa Telecom / China Trust / Visa / Nokia NFC smart-poster program in Taipei (Sept. 2007). This is a genuine practicing entity at the relevant time — the patents were not born in a shell.
• Current status: Not bankrupt or dissolved on any record I found. RFCyber collateral and product PDFs remain online at rfcybercorp.com; RFCyber Corp. appears as a real party-in-interest in the 2021–2023 PTAB proceedings (e.g., IPR2022-01240 mandatory notices, Aug. 10, 2022, identifying "Shenzhen RFCyber Asset Management, LLP" and "RFCyber Corp." as RPIs); a Texas affiliate, RFCyber Corp., is the plaintiff in the 2020–2024 campaigns and per its Rule 7.1 disclosure identifies RFCyber Corporation as its corporate parent (2:24-cv-00546, Dkt. 5). The Texas plaintiff has been described in court filings as a Texas corporation with a place of business at 600 Columbus Avenue, Suite 106, Waco, Texas (quoted in RFCyber Corp. v. Apple Inc., W.D. Tex. 6:21-cv-00916, Dkt. 149 at 3) and in press coverage as "of Plano, Texas" (The Recorder, July 23, 2024).

Assignment timeline

Two recorded assignments. Both were recorded on the same day, 2014-02-08, on the same reel (032177), on consecutive frames — a single filer recording a two-step chain.

• 2006-09-23 (executed) / recorded 2014-02-08 — Reel 032177/0739

  • Conveyance: Assignment
  • Assignor: Koh, Liang Seng; Cho, Futong; Pan, Hsin; and others (i.e., all four named inventors)
  • Assignee: RFCyber Corporation (Fremont, CA)
  • Correspondent: NOT DETERMINED. I could not retrieve the correspondent field for this recording. See "Limitations" below.
  • Context: Internal founder/employee→company assignment (the original, normal first link), but recorded nunc pro tunc 7.4 years late — simultaneous with the offshore transfer below.

• 2014-01-27 (executed) / recorded 2014-02-08 — Reel 032177/0743

  • Conveyance: Assignment
  • Assignor: RFCyber Corporation (Fremont, CA) — "RFCyber Holding"
  • Assignee: Shenzhen RFCyber Asset Management, LLP (China)
  • Correspondent: NOT DETERMINED. Same caveat; because both recordings sit on one reel, four frames apart, filed the same day on the same two-step chain, a single-filer inference is reasonable but is not proof — I will not name a correspondent without the record.
  • Context: Transfer-to-holding-vehicle as a capital contribution. The consideration was intellectual property contributed in kind: Chinese registry data (Tianyancha / 企知道, publicity filed 2022-05-09) show RFCyber Corp. subscribed 24.3M RMB of the partnership on 2013-11-13 and made the contribution paid-in on 2014-01-13, with 出资方式 = 知识产权 ("intellectual property"), holding 99.7947%; the only other partner is an individual, 沈娇, at 5万元. Registry: USCC 914403000834215126, registered capital 2435万元, address 深圳市龙华区民治街道樟坑社区青创城A栋3-A5 (Shenzhen Longhua), listed industry "retail." The USPTO effective date (2014-01-27) sits 14 days after the paid-in IP contribution — the recordation followed the equity contribution.

Subsequent transfers — documented in litigation, but I found no corresponding USPTO assignment records:

• 2020-07-30 — Shenzhen RFCyber granted RFCyber Corp. (TX) an exclusive license to the '218 and '855 patents "as well as all patent applications, patents to be issued pursuant thereto, divisional, continuations, continuations-in-part, reissues, substitutes, and extensions thereof." (Plaintiff's Response in Opposition to Samsung's Motion to Dismiss, 2:20-cv-00274, Dkt. 92, Ex. 6.)
• 2020-07-30 — RFCyber Corporation (CA) assigned the '787, '009, and '046 (not '218/'855, which it no longer owned) to RFCyber Corp. (TX). (Same brief, Exs. 5–6.)
• 2021-07-30 — Shenzhen RFCyber made a confirmatory assignment to its exclusive licensee RFCyber Corp. (TX) "for the avoidance of any doubt." (Same brief.) Note the brief describes this confirmatory assignment as covering "the patents-at-issue" = '787/'009/'046; for '218 and '855 the parties relied on the exclusive license, which is why litigation over the '855 chain turns on license scope rather than a recorded assignment.

Google Patents' legal-events table for the '855 stops at the two 2014 records, and its current-assignee field still reads Shenzhen RFCyber Asset Management LLP and Rich House Global Tech Ltd. — consistent with no post-2014 recorded assignment against this patent.

Timeline diagram

timeline
    title Ownership of US 8448855
    2006 : Filed by RFCyber inventors
    2013 : Patent issued May 28
    2014 : IP contributed to Shenzhen RFCyber
         : Reels 032177 0739 and 032177 0743 recorded
    2020 : Exclusive license to RFCyber Corp Texas
         : First suits against Google Samsung LG
    2021 : Confirmatory assignment from Shenzhen
    2023 : Reexam C1 confirms claims
    2024 : Retail campaign filed against six brands
    2025 : Reexam C2 confirms claims

NPE / troll-pattern signals

1. Shell-entity transfer — PRESENT (with an affiliate caveat).
Reel 032177/0743 (executed 2014-01-27, recorded 2014-02-08) moved the asset from the operating company to Shenzhen RFCyber Asset Management, LLP — a Shenzhen limited partnership holding vehicle whose only economic partner is the transferor itself (99.7947% by IP contribution) alongside a single 5万元 individual partner. Evidence it is a non-operating holdco: registered in Longhua, Shenzhen; industry classification "retail" with software-copyright and patent holdings (企知道 reports 15 patents); its registry contact is a richhouseglobal.com address shared with Rich House Global Technology Ltd.; no US products, no US address, and the operating/commercial function stayed with the California entity. Caveat I am obliged to state: this is an intra-family affiliate transfer, not a sale to an unrelated third-party troll, so it satisfies the name/structure/no-products prongs but not the arm's-length prong.

2. Known asserter in the chain — PRESENT (moderate).
None of the enumerated lists (Acacia, Marathon, IV, IPNav, Wi-LAN, Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, DGC, Spangenberg) appears in this chain — RFCyber is a home-grown asserter, not a portfolio buyer. However, the rubric also asks about entities "surfaced by Unified Patents or RPX as a high-frequency plaintiff," and RFCyber qualifies on both: the '855 carries Unified Patents litigation-portal entries and three PTAB filings (IPR2021-00954, IPR2021-00978, IPR2022-01241), and RPX Insight covered the July 2024 filing wave as "the latest defendants to be sued in the August 2020 litigation campaign launched by inventor-backed RFCyber Corp." (RPX, July 26, 2024) and maintains an RPX Insight docket page for 2:24-cv-00546. Assessed as moderate/weak because the entity is a practicing-origin company rather than a classic third-party NPE.

3. Repeat correspondent across the chain — UNCLEAR / NOT VERIFIABLE on the evidence I could reach.
I could not read the Assignment Center correspondent field for either reel 032177/0739 or 032177/0743. What I can document as recurring counsel (a different thing, and I will not conflate them):

• Fabricant LLP (411 Theodore Fremd Avenue, Suite 206 South, Rye, NY 10580) appears repeatedly across the entire RFCyber campaign: PTAB mandatory notices for IPR2022-01240 (Aug. 10, 2022, listing Vincent J. Rubino III, Peter Lambrianakos, Alfred R. Fabricant, Enrique W. Iturralde, Richard Cowell), the updated power of attorney in IPR2022-00412 (Feb. 2, 2023, adding Jacob Ostling), and as filing counsel on the 2024 complaints (signature "Fabricant, Alfred" on the Costco complaint, 2:24-cv-00546, filed 2024-07-18), with attorney appearances for Lambrianakos, Rubino, Cowell and Ostling. That is genuine recurrence — but it is litigation counsel, not the assignment correspondent, and the rubric warns that a single appearance is not a finding. I therefore record signal 3 as unclear, and flag that resolving it requires pulling the correspondent field from https://assignmentcenter.uspto.gov/.
• Prosecution-side correspondence of record on the parent application 11/534,653 was Silicon Valley Patent Agency, 7394 Wildflower Way, Cupertino, CA 95014 (attorney docket RFID-081), per the 2008 filing receipt reproduced in Samsung's FRD exhibit (IPR2021-00981, Ex. 1002). Caveat: that exhibit is noisy OCR and contains other garbled correspondent strings ("HNL Patents"); I rely only on the clean Silicon Valley Patent Agency address block.

4. Cascading transfers — UNCLEAR (leaning not present on the record).
The recorded chain is only two links, both recorded the same day on one reel (032177/0739 and /0743), which is a simultaneous chain-completion, not a rolling series of shell-to-shell hops. The later steps (2020-07-30 exclusive license, 2021-07-30 confirmatory assignment) are documented in briefs but appear unrecorded, so they cannot be measured against the "<24 months through chained LLCs" test from the assignment record. The real cascading pattern here is ownership/license bifurcation across three similarly named entities in two countries (RFCyber Corporation CA → Shenzhen RFCyber Asset Management LLP → exclusive license back to RFCyber Corp. TX), which required the plaintiff to litigate standing in 2021 — a chain-structure red flag even though it is not a classic LLC cascade.

5. Pre-litigation transfer — PRESENT.
The first complaints naming the '855 were filed 2020-08-21 (2:20-cv-00274, E.D. Tex.) and 2020-10-16 (2:20-cv-00335 and 2:20-cv-00336, E.D. Tex.). The exclusive license and chain-cleaning agreements were executed 2020-07-30 — 22 days before the first '855 complaint. A second chain-cleaning act (confirmatory assignment, 2021-07-30) followed Samsung's 2021-07-16 motion to dismiss for lack of standing (2:20-cv-00274, Dkt. 81), i.e., the chain was actively repaired during litigation to defeat a standing challenge. Both acts are pre/during-litigation chain arrangement.

6. Bankruptcy fire-sale — NOT PRESENT.
No Chapter 7/11 record, no trustee sale, no docket evidence of an insolvency proceeding for RFCyber Corporation, RFCyber Corp., Rich House Global Technology Ltd., or Shenzhen RFCyber Asset Management LLP. Maintenance fees were paid on schedule (4th yr 2016, 8th yr 2020, 12th yr with late surcharge 2025-01-28); reexamination certificates C1 (2023-04-28) and C2 (2025-03-07) were actively prosecuted. A party in bankruptcy does not spend six figures defending two ex parte reexaminations.

7. Privateering — NOT PRESENT (as classically defined), with a self-dealing note.
Classic privateering is an operating company funding an NPE to sue competitors; here the operating originator (RFCyber Corporation, CA) transferred to its own Chinese holding LP and then took an exclusive license back into another of its own affiliates (RFCyber Corp., TX). There is no third-party financier operating at arm's length and no SEC-filed operating company behind the assertion. The economic direction is reversed relative to the textbook pattern. The substantive concern is self-directed monetization plus a litigation vehicle, which is captured under signals 1 and 5 rather than 7.

8. Defensive aggregator — NOT PRESENT.
No chain step ends at RPX, AST, LOT Network, Unified Patents, or OIN. Important distinction: Unified Patents appears in this patent's history only as an IPR petitioner (IPR2021-00978, not instituted) and as a litigation-data publisher — being challenged by a defensive aggregator is the opposite of being acquired by one. The patent has been repeatedly asserted, not neutralized. (The '855 remains live; conversely, the family's '009 patent was invalidated on appeal in RFCyber v. Coke Morgan Stewart, Fed. Cir. 2023-2418, decided 2025-08-14 — a family-value distinction the prior section already flagged.)

Verdict

NPE — high confidence.

Two strong signals support this. (i) Shell/holding-vehicle transfer, Reel 032177/0743 (executed 2014-01-27, recorded 2014-02-08): the '855 was moved out of the operating originator and paid into a Shenzhen limited partnership as an in-kind IP capital contribution (24.3M RMB, 99.7947% of the partnership, paid-in 2014-01-13 per Chinese registry), with the same reel also carrying the nunc-pro-tunc recording of the 2006 inventor assignment (Reel 032177/0739, effective 2006-09-23) — i.e., the record was built at the moment of transfer. (ii) Pre-litigation chain arrangement: the exclusive license to the Texas plaintiff was executed 2020-07-30, 22 days before the first '855 suit (2:20-cv-00274, filed 2020-08-21), and a confirmatory assignment followed on 2021-07-30 in direct response to Samsung's standing motion. Reinforcing context: RPX calls the campaign "inventor-backed RFCyber Corp.," and the 2024 wave (2:24-cv-00546/-00548/-00549/-00550/-00551/-00575) targets retail/consumer-app defendants — Costco, Kroger, Shell, Starbucks, Walmart, Volkswagen — that are not competitors in any market RFCyber serves, which is the substantive NPE hallmark rather than a competitor-vs-competitor dispute. The verdict would drop to moderate if one insists on treating the 2014 transfer as an arm's-length-looking intra-family reorganization, and signal 3 (repeat assignment correspondent) remains unresolved pending a direct Assignment Center pull.

Limitations and verification path (stated plainly)

• The assignment correspondents of record are NOT in this analysis — I could not retrieve them, and I decline to infer a name from the reel number, the filing date, or the litigation firm. Verify at https://assignmentcenter.uspto.gov/ (search "8448855" or Reel 032177) and at the Google Patents legal-events tab: https://patents.google.com/patent/[US8448855B1](/patent/US8448855B1)/en#legalEvents.
• I did not find USPTO recording numbers for the 2020-07-30 exclusive license, the 2020-07-30 assignment, or the 2021-07-30 confirmatory assignment. They may exist unrecorded, recorded under the sibling patents only, or recorded after the last Google Patents refresh. Confirm by searching each of '218, '855, '787, '009, '046 in Assignment Center.
• The Chinese registry figures (24.3M RMB, 99.7947%, 知识产权 contribution, USCC 914403000834215126) come from third-party aggregations of the National Enterprise Credit Information Publicity System (Tianyancha, 企知道, publicity filed 2022-05-09) — treat as strong but not primary-source verification of the IP-as-capital-contribution story. The USPTO effective date (2014-01-27) is independent corroboration of the timing.

I'll research the cited prior art for US 8,448,855. Let me pull the reference records from authoritative sources.

Prior-Art Analysis — U.S. Patent No. 8,448,855 B1

Patent searched (literal ID, not normalized): US 8,448,855 B1 — "Method and apparatus for funding an electronic purse," App. No. 13/400,038, filed Feb. 18, 2012, granted May 28, 2013, priority Sept. 24, 2006 (continuation of 11/534,653, now US 8,118,218).

⚠️ Two threshold flags before the analysis

1. Date contradiction (must be flagged per your rules). The task header says the current date is April 26, 2026; the authoritative patent record I fetched is timestamped 2026-09-28. The previously generated "Patent summary" section states the docketed expiration is Sept. 24, 2026 — i.e., on or before the 2026-09-28 fetch date. If the fetch date governs, the '855 term has run. If the April 26, 2026 task date governs, it has not. I am not auto-resolving this; I state both. Google Patents' projection (20 years from Sept. 24, 2006) is an assumption, not a legal conclusion, and the maintenance-fee record shows the 12th-year fee paid (Jan. 28, 2025, late surcharge), so no fee-based lapse is indicated.

2. Which citations are "the" '855 citations. The '855 record identifies exactly four U.S. patent citations ("Patent Citations (4)"). Separately, the same page lists three "Family Cites Families" references that were cited in related family members (e.g., the '218), not necessarily in the '855 prosecution itself. I treat the four as the '855's citations and present the three family cites separately and clearly labeled.

A. The four cited references of record

For each: full citation, dates, description, and the claims it could potentially touch under 35 U.S.C. § 102. Because the '855 has a Sept. 24, 2006 effective filing date and was filed Feb. 18, 2012 — before the AIA first-to-file provisions took effect (March 16, 2013) — pre-AIA §§ 102(a)/(b)/(e) govern.

1. US 2003/0145205 A1 — Branko Sarcanin, "Method and system for a virtual safe"

Description. Discloses an electronic-commerce "VirtualSAFE" system using virtual smart cards. A transaction server contains a security module (a "Crypto-Engine," CEV), a virtual smart-card database, and — the key disclosure — "an emulator for emulating a smart card, the emulator for receiving smart card commands and processing the commands in conjunction with the virtual smart card database and the security module," plus a "virtual card reader module" relaying smart-card commands to the emulator. Card records store purchase key and load key identifiers, balance, BIN, etc. It expressly describes loading/initiating value onto a (virtual or physical) smart card over an open network such as the Internet, with a bank server, an initiation server with a security module, and existing clearing/settlement. It also describes the smart card as possibly embedded in "a personal digital assistant, telephone, or some other form" ([0004]).

Anticipation analysis — which claims it could reach:

• Cannot anticipate independent claim 1 or claim 9 as a whole. Claim 1 expressly requires an NFC-enabled portable device including a card module housing an e-purse applet, a midlet that sends a request to that applet, and personalization via an e-purse SAM external to the card module plus a second security channel created "on top of" the initial channel. Sarcanin's emulator sits on a remote transaction server, not in the user's portable smart-card module, and it discloses no external e-purse SAM or layered-channel personalization. Its own text even frames the server-side emulation as a way to dispense with physical cards and readers — the opposite architecture.
• Potentially relevant to: claim 5 / claim 14 ("the e-purse is implemented in the emulator") and the key vocabulary underlying claims 3–6 / 12–15 (load key, purchase key, balance, transformed keys/passwords to access card data). Sarcanin's server-side emulator is the closest cited-art teaching of an "emulator" handling card commands, but it is not the claimed in-device emulator.
• Realistic role: § 102 art for narrow dependent claims at best; principally § 103 art.

Source: https://patents.google.com/patent/US20030145205

2. US 2002/0145632 A1 — Shimon Shmueli, "Portable interface for computing"

Description. Discloses a portable memory "key" (USB key, smart card, or wireless transponder) that automatically executes software ("keylets" — Java applets) on a host computer once inserted/detected. It includes an authentication routine (user name/password or biometric indicia compared against encrypted data stored on the key), an encrypted "e-wallet"/account manager storing credit/debit card numbers, billing/shipping addresses, and password management, and file-system emulation on the host.

Anticipation analysis — which claims it could reach:

• Cannot anticipate claim 1 or claim 9. No e-purse applet, no card module with an emulator, no NFC enabled device requirement, no e-purse SAM, no layered security-channel personalization, and no server-side funding flow against a bank account. Its "keylet" is an applet that runs on a host computer, not the claimed e-purse applet inside a smart-card module, and its "e-wallet" is a stored-form filler, not a fundable e-purse.
• Potentially relevant to: the general concepts recited in dependent claims about access keys / transformed passwords (claims 6 / 15) and user authentication; useful as background for "midlet/executable application on a portable device" terminology, but not anticipatory.

Source: https://patents.google.com/patent/US20020145632A1

3. US 2005/0222961 A1 — Philippe Staib et al., "System and method of facilitating contactless payment transactions across different payment systems using a common mobile device acting as a stored value device"

Description — the most material cited reference. Discloses a mobile telephone with an NFC module and its own secure memory area storing "stored value (digital cash)" — i.e., an electronic purse — plus a mobile application developed with J2ME/MIDP 2.0 (the specification repeatedly uses "MIDlet suites," security domains, and X.509-signed MIDlets). The mobile application emulates transmission standards and data-exchange formats of contactless payment systems. Critically, it discloses remote charge/recharge (funding) of the stored value over GPRS/Internet without a dedicated terminal: the user selects "recharge," the mobile application transmits the instruction to a service operator computer; the service operator queries the user's bank to confirm the registered account and obtains a recharge approval code and confirms with a wallet operator; a PIN is entered by the user and validated by the bank; the service operator then "updates the mobile application with the value of the recharge approved by the user's bank," and the bank transfers money to the wallet operator's account (steps 120–144; FIG. 9 Visa 3-D Secure flow for card-funded recharge). The mobile device stores transaction details in a transaction log and updates the remaining stored value in secure memory.

This is the reference that maps most closely to the funding steps of claim 1 — PIN entry/verification, a mobile (MIDlet) application initiating a request, transmission over a wireless network to a server, server verification against a bank account, initiation of a fund-transfer request to the financial institution, and return of a value-update to the device. It was evidently the lead applied art against this family (Google used it as Ex. 1005 in IPR2021-00957 against sibling US 8,118,218).

Anticipation analysis — which claims it could reach:

• Likely does NOT anticipate claim 1 or claim 9 in full. Two claim-1 elements remain untaught: (i) the e-purse "applet" residing in a "card module" with a "midlet" that sends its request to that applet and the applet composing the response; and (ii) the personalization predicate — an initial security channel between the card module and an e-purse SAM external to the card module, and a second security channel layered "on top of" the initial channel, over which "any subsequent transactions with the e-purse are conducted." Staib's secure element/NFC chip, MIDP security domains, and X.509 secure channel are adjacent but are not the recited SAM-plus-superimposed-channel personalization.
• Potentially anticipatory / strongly relevant to the dependent claims, e.g.: claim 5/14 (e-purse implemented in the emulator/secure element), claim 6/15 (access-key-based access), the NFC-enabled mobile device context of claim 1/10, and the transaction-log updating and APDU-adjacent command handling concepts behind claims 7–8 / 16–17.
• Realistic role: primary § 103 reference (and possibly § 102 for the funding-flow-dependent claims), which is exactly how it was deployed in the PTAB.

Source: https://patents.google.com/patent/US20050222961A1

4. US 2006/0171383 A1 — Alexander Davydov, "Handling incoming data"

Description. Concerns buffering incoming data for a MIDlet ("target application") that is not yet running, using middleware/the Application Management System (AMS) and J2ME MIDP 2.0 Push Registry. It addresses push technology, inbound connection types (socket, datagram, WMA/SMS), and launching a MIDlet in response to an inbound connection.

Anticipation analysis — which claims it could reach:

• Anticipates none of the substantive claims. It discloses no e-purse, no stored value, no card module/emulator, no SAM, no funding, and no security-channel personalization. Its only arguable bearing is terminological — it evidences that a "MIDlet" was a well-known J2ME/MIDP application component managed by an AMS with a push registry, useful background for construing "midlet embedded in the portable device" in claims 1 and 9 (and for the "midlet" recited as the verifier in claim 9).
• Realistic role: § 102(a)/(e) background art only; not an anticipation reference for any claim.

Source: https://patents.google.com/patent/US20060171383

B. "Family Cites Families" references (NOT the '855's own citations)

The '855 record lists these under "Family Cites Families (3)", meaning they appear as citations in related family members. I present them for completeness but do not treat them as the '855's cited prior art — and none is a viable § 102 anticipatory reference against the '855's independent claims.

Source (record): https://patents.google.com/patent/US8448855/en (Family Cites Families section)

C. Bottom line

1. No reference cited on the '855 — and, on this record, none of the four — is a clean § 102 anticipatory reference against independent claims 1 or 9. Both independents are gated by a personalization predicate that is unusual and specific: an initial security channel between the card module and an e-purse SAM external to the card module, plus a second security channel created "on top of" the initial channel, with all subsequent e-purse transactions conducted over the second channel. None of Sarcanin, Shmueli, Staib, or Davydov discloses that SAM-plus-superimposed-channel structure.
2. Staib (US 2005/0222961 A1) is the closest art on the funding element and maps substantially to the PIN → mobile application → wireless request → server verifies bank account → fund transfer request → return/update flow of claims 1 and 9. It is the reference most likely to support a § 103 combination (or § 102 against narrow dependent claims), which is consistent with its use in the family's IPRs.
3. Sarcanin (US 2003/0145205 A1) supplies the "emulator receiving smart-card commands + security module + virtual card database" teaching — relevant to dependent claims 5/14 and the key vocabulary of 3–6/12–15 — but places the emulator server-side, not in the NFC portable device.
4. Shmueli (US 2002/0145632 A1) and Davydov (US 2006/0171383 A1) are, respectively, an encrypted portable "e-wallet/keylet" system and a MIDlet buffering/push mechanism; neither reaches the e-purse funding claims.
5. Prosecution/contested-proceeding corroboration: the '855 survived two ex parte reexaminations with claims confirmed unamended (certificates C1, effective Apr. 28, 2023, and C2, effective Mar. 7, 2025) and saw institution denied in every IPR filed against it (Google IPR2021-00954 — settled pre-institution; Samsung IPR2021-00978; Apple IPR2022-01241). That outcome is consistent with this analysis: the cited art, including Staib, does not squarely disclose the claimed SAM-based, two-channel personalization of the in-device e-purse.

Caveats on confidence. (a) I did not retrieve the actual examiner's Form 892/IDS or the office actions; "Patent Citations (4)" reflects citations as recorded in the Google Patents machine transcription, and additional references cited only in the parent (11/534,653) are not shown for the '855. (b) I have not verified the grant status of every cited publication from the USPTO PatentCenter directly; the publication↔grant correspondences above (US 2003/0145205→US 6,941,285; US 2006/0171383→US 7,707,291) are drawn from the records retrieved and should be confirmed against the face of the documents if it matters. (c) No US Federal Circuit appeal specifically addressing the '855 was located; related family patents have been invalidated on appeal, but that is a different asset.

Sources: Google Patents records for US 8,448,855 B1, US 2003/0145205 A1, US 2002/0145632 A1, US 2005/0222961 A1, and US 2006/0171383 A1; PTAB exhibits (e.g., Ex. 1005 in IPR2021-00957); Unified Patents PTAB/litigation portal; USPTO reexamination certificates C1 and C2 (as recorded on the '855 page).

I'll ground this analysis in the prior-art references listed on the patent's own page (the four "Patent Citations," three "Family Cites Families," and note the "Cited By" items are post-priority). Let me confirm what the key references actually teach before mapping them to the claims.

Let me check what prior-art grounds the actual IPR petitioners asserted against the '855 (helpful to explain why those combinations were rejected), and confirm the two additional family-level references.

Obviousness Analysis — U.S. Patent No. 8,448,855 under 35 U.S.C. § 103

Scope note / ground rules. This analysis uses only the references listed in the Prior Art section of the patent's own Google Patents page (the four "Patent Citations," the three "Family Cites Families," and the "prior art keywords"). The "Cited By (2)" and "Families Citing this Family (49)" entries are forward citations that post‑date the patent and are therefore not § 102/§ 103 prior art; I exclude them except as context. I do not normalize any publication numbers. Where the record contradicts my assumptions, I flag it rather than auto‑correct.

Which law applies. The '855 has an effective filing date of Sept. 24, 2006 (continuation of 11/534,653). It was filed Feb. 18, 2012 — before the AIA's March 16, 2013 first‑inventor‑to‑file cutoff — and its claims have a pre‑2013 effective filing date. Pre‑AIA 35 U.S.C. § 103(a) therefore governs, applied under the Graham v. John Deere framework as refined by KSR Int'l v. Teleflex (motivation + reasonable expectation of success, predictable variation).

Minor date discrepancy (carried over from the summary, still relevant to § 102 dates). Unified Patents lists the whole chain one day earlier (priority 2006‑09‑23, filing 2012‑02‑17, grant 2013‑05‑27, expiry 2026‑09‑23); Google Patents uses 2006‑09‑24 / 2012‑02‑18 / 2013‑05‑28 / 2026‑09‑24. This does not change any § 103 outcome below — all references were published well over a year before either date.

1. The prior‑art universe (from the page's Prior Art section)

§ 102 timing check: Sarcanin (2003), Shmueli (2002), Staib (2005), Davydov (2006‑08‑03), Banksys (1996), Beepcard (2003) all published more than one year before the 2006‑09‑24 priority date. EP 1 961 153's PCT was filed 2005‑12‑15 and published in English, designating the US — available as § 102(e) art as of its international filing date, i.e., before the priority date. All seven qualify.

2. The claim limitations that must be met

Independent claim 1 (device side) and claim 9 (server side) share the same architecture. Parsed into elements:

• (A) NFC‑enabled portable device containing a card module.
• (B) PIN entered and verified.
• (C) Midlet sends a request to an e‑purse applet; the applet composes a response.
• (D) The applet sends the response over a wireless network to a server that verifies it against an account at a financial institution and, on success, initiates a fund‑transfer request.
• (E) Receiving commands back from the server (claim 7/16: network messages carrying APDU commands, extracted).
• (F) An emulator on the device updates a transaction log after authenticity is verified (claim 1: by the applet; claim 9: by the midlet).
• (G) — the crux: the e‑purse must have been personalized by (G1) establishing an initial security channel between the card module and an e‑purse SAM external to the card module to install/personalize the applet, and (G2) creating a second security channel "on top of" the initial security channel to protect subsequent operations, with all later transactions conducted over that second channel.

Dependent claims add: SmartMX (SMX) card module (2/11); Global Platform (GP) build, both e‑purse and card‑access keys personalized (3/12); card‑manager security with the applet as gatekeeper (4/13); e‑purse implemented in the emulator (5/14); transformed passwords based on access keys (6/15); APDU extraction (7/16); commands include the FI's response (8/17).

3. Proposed obviousness combinations

Combination A — Staib + Sarcanin + Davydov (primary; the likely Petitioner‑style ground)

Motivation to combine (KSR): All three are in the same field (contactless/mobile stored‑value payment and smart‑card personalization) and address the same problem the '855's own Background identifies — enabling an e‑purse to transact over an open network without compromising security. Staib supplies the express motivation: the admitted prior‑art drawback that stored value "can only be recharged using dedicated hardware," and the stated desire to recharge "anywhere without requiring the user to be near a dedicated hardware terminal." A POSITA seeking network‑based top‑up of a contactless stored‑value device would naturally look to a secure transaction server with a card emulator and security module (Sarcanin) and to standard mobile middleware for delivering inbound network data to an on‑device application (Davydov). The combination is a predictable arrangement of known elements with a reasonable expectation of success (all rely on ISO 7816 APDUs / smart‑card security, which the '855 itself treats as routine).

Combination B — Staib + EP 1 961 153 (Nokia) + Sarcanin (strongest on elements D–F)

Nokia/Nyström directly teaches the "receive commands from the server and hand them to the contactless secure element" mechanism: the terminal's interconnectivity component detects network messages destined for the secure storage subsystem and supplies them to a secure memory controller that executes the instructions. This maps almost one‑to‑one onto claim 1's (E) receive commands and the plumbing to the emulator (F), and provides a clean motivation: remote over‑the‑air management of a contactless secure element (Nokia's stated object is "to enable a cardholder as well as a service provider to remotely exercise active and/or passive control"). Combined with Staib (stored‑value funding over the network) and Sarcanin (security module + emulator + fund/verify server), the Combination‑A chart is fully populated, with Nokia strengthening the OTA‑command path.

Combination C — add for the dependent claims

• 4/13 (card manager / applet as gatekeeper): Banksys BE 1008699 (selective access to a security system) + the GP/card‑manager framework that the '855's own specification admits is prior art ("a card manager module is configured by using the smart card security framework … Global Platform 2.1 … performs the card manager functionality"). Using a known card manager to gate access to an applet is the predictable use of that admitted tool.
• 6/15 (transformed passwords based on access keys): combining Sarcanin's security module with the Mifare/contactless key‑derivation techniques the specification treats as known (MIFARE is described as "the most widely installed contactless smart card technology in the world").
• 2/11 (SMX), 3/12 (GP), 5/14, 7/16 (APDU extraction), 8/17: These recite selection of a known secure element/OS and routine data handling (extracting APDUs from network messages is the ordinary reader/host function in ISO 7816 systems). Courts generally treat such platform selections and data‑plumbing as obvious absent unexpected results.

4. Where the obviousness case is weak — and why it matters

Two limitations are not squarely taught by anything in the listed prior art, and they are exactly where the real‑world challenges failed:

1. Element (G): the layered security‑channel personalization. The claims require an initial channel between the card module and an external e‑purse SAM to install/personalize the applet, and then a second channel "on top of" the initial channel for subsequent operations. The listed art teaches personalization (Sarcanin) and remote secure messaging (Nokia), but neither teaches the two‑tier, on‑top‑of relationship characteristic of a Global Platform security domain / Secure Channel Protocol layered over an install channel. Building (G2) requires importing Global Platform itself as the "known art" — the patent's own specification leans on GP 2.1, but none of the four required "Patent Citations" or three family cites is a GP reference. A Petitioner must therefore bridge this gap with a separate GP/SCP reference or expert testimony, not with the page's prior art alone.
2. Claim 9's inverse responsibility allocation. Claim 9 says commands are verified by the midlet, whereas claim 1 says the applet verifies. A combination that teaches verification by the secure element (Sarcanin/Nokia) does not, without more, disclose midlet‑side verification — a genuine claim‑drafting distinction that narrows the server‑side claim.

The record corroborates the gap

• Three IPRs against the '855 — all denied institution: Google's IPR2021‑00954 (terminated on settlement pre‑institution), Samsung's IPR2021‑00978 (institution denied on the merits, Paper 10, Dec. 14, 2021), and Apple's IPR2022‑01241 (institution denied, Paper 7, Jan. 23, 2023). Petitioners evidently did not establish a reasonable likelihood of success with the art available to them.
• Two ex parte reexaminations confirmed all claims: certificate C1 (effective Apr. 28, 2023) and certificate C2 (effective Mar. 7, 2025), with no amendments. A reexamination confirmation is a merits determination by the examiner over the art of record — stronger evidence against a § 103 challenge than a mere institution denial.
• Caveat, per the operating rules: these outcomes do not prove the claims are non‑obvious as a matter of law. Institution denial is a threshold "reasonable likelihood" standard; a different combination or different art (e.g., a GP reference or the RFCyber family's own intervening art) could still support a § 103 theory. I could not independently confirm from the retrieved records the precise reference sets the petitioners used in IPR2021‑00978 and IPR2022‑01241; the only firm datapoint is that Staib was used against the parent '218 in Google's IPR2021‑00957 (as Ex. GOOG‑1005). That strongly suggests Staib is the primary intended reference, consistent with Combination A above.

5. Bottom line

• Strongest prima facie § 103 case: Staib + Sarcanin + Davydov (with EP 1 961 153 added to firm up the OTA command path). This combination appears to render claim 1 and the platform/plumbing dependent claims (2–8, and their 11–17 counterparts) obvious: Staib teaches the NFC stored‑value device and network recharge; Sarcanin teaches the card‑emulator + security module + fund‑verifying transaction server and personalization; Davydov teaches the midlet/agent architecture for delivering network data to an on‑device applet; Nokia teaches handing OTA network messages to the secure element.
• Likely non‑obvious hook: element (G) — the layered, on‑top‑of security‑channel personalization between the card module and an external e‑purse SAM — is not disclosed by the listed art and would require importing Global Platform/SCP as background knowledge. Claim 9's midlet‑verifies allocation is a secondary distinction.
• Consistency with the record: the two reexamination confirmations (C1, C2) and three denied IPRs indicate that, on the art of record, the claims have not been held obvious — a proposition an accused infringer must overcome with art or testimony beyond what this page lists.
• Explicit uncertainties: (i) I could not confirm the exact ground/reference sets in IPR2021‑00978 and IPR2022‑01241; (ii) forward citations ("Cited By," "Families Citing") are not prior art and were excluded; (iii) the one‑day date discrepancy between Google Patents and Unified Patents does not affect any § 103 conclusion.