US 9990628

I'll search for authoritative information on this patent, including any CAFC 2026 docket activity.

Let me verify the granted claims and check the litigation docket.

Let me pull the granted patent's independent claims directly.

Search / verification statement

I searched for the specific identifier 9990628 (US 9,990,628 B2; application US13/857,905; pre-grant publication US20130297514A1) and for any CAFC 2026 docket activity involving it. I did not auto-correct or normalize the number. Where sources conflict on details, I note the conflict rather than resolving it silently.

1. Bibliographic data — US 9,990,628 B2

Source: https://patents.google.com/patent/US9990628/en and the patent PDF cover page (https://patentimages.storage.googleapis.com/81/b8/c8/9ea9a2ec589ce0/US9990628.pdf).

2. Abstract

"A system and method provide efficient, secure, and highly reliable authentication for transaction processing and/or access control applications. A Personal Digital Key stores one or more profiles (e.g., a biometric profile) in a tamper-proof memory that is acquired in a secure trusted process. Biometric profiles comprise a representation of physical or behavioral characteristics that are uniquely associated with an individual that owns and carries the PDK. The PDK wirelessly transmits the biometric profile over a secure wireless transaction to a Reader for use in a biometric authentication process. The Reader compares the received biometric profile to a biometric input acquired at the point of transaction in order to determine if the transaction should be authorized."

3. Plain-language overview of the claims

Important limitation on this section. I was able to retrieve a verbatim claim recitation for claim 1 (below), but I could not independently retrieve the complete granted claim set of US 9,990,628 B2, nor the total claim count or the full text of every other independent claim. What I can state confidently is limited to: (a) claim 1 as recited, and (b) the fact that the Intel complaint identifies claim 1 as an independent claim being asserted. The specification's description of the two authentication layers is well documented, but the claim scope should be verified against the granted patent document before being relied upon.

What the "two levels" are, per the specification (col. 8–10 region and FIGS. 4–7D):

• Level 1 — Device authentication: begins whenever a PDK enters a Reader's proximity zone. PDK and Reader each validate the other as valid devices (e.g., challenge-response / CHAP) and confirm the PDK holds profile types the Reader requires; this can be the only layer for low-risk transactions (e.g., purchases under a preset amount such as $25).
• Level 2 — Profile authentication: one or more of biometric, PIN, picture (photo), or registry/database authentication, optionally combined. The Reader receives the profile from the PDK over a secure channel and compares it to an input acquired at the point of transaction. A "trigger" (e.g., biometric contact, dispensing gas, scanning items) may be required depending on configuration.

Independent claim 1 (as recited in the US 2013/0297514 A1 publication, which corresponds to application 13/857,905):

"1. A method for electronic authentication comprising: wirelessly receiving a biometric profile from a personal digital key (PDK), wherein the biometric profile is uniquely associated with an individual; receiving a biometric input; comparing the biometric profile to the biometric input; and authorizing a transaction responsive to the biometric profile matching the biometric input."

In plain language: a portable key device (PDK) wirelessly hands over a stored biometric profile tied to one person; the reader captures a live biometric input; the two are compared; and the transaction is approved if they match. Notably, this claim does not on its face recite the "sample-then-full-profile" two-stage comparison described in the specification's FIG. 7A flow, nor the registry lookup of FIG. 7D — those appear to be additional/dependent features in the disclosure.

Caution: The claim text above is from the pre-grant publication of the same application, not from a verified extract of the granted patent. Prosecution amendment could have changed it. I did not obtain an authoritative rendering of the granted claim 1 or of any other independent claims.

4. Asserted independent claims in litigation (separate from the patent text)

Per a third-party complaint analysis of Proxense, LLC v. Intel Corporation, 7:25-cv-00531 (W.D. Tex.):

• US 9,990,628 is asserted, with "at least independent Claim 1."
• The asserted theory characterizes the '628 patent as "methods for biometric transaction processing where a personal digital key performs a local comparison against a stored sample, retrieves a full profile from secure memory for re-validation, and participates in workflows to transmit authorization or transactional data for registry-based validation."
• Accused products: Intel platforms with Platform Trust Technology (PTT) and the Converged Security and Management Engine (CSME).
• The same complaint also asserts related Proxense patents: US 9,405,898; US 10,469,456; US 8,433,919; US 11,182,792; and US 11,551,222.

Sources:

• https://ai-lab.exparte.com/case/dct/txwd/7:25-cv-00531/doc/analysis/1
• https://cdn.pacermonitor.com/public/case/61206235/Proxense,_LLC_v_Intel_Corporation_ (Exhibit F — U.S. Patent No. 9,990,628)
• Case docket: 7:25-cv-00531, W.D. Tex., Midland Division, Judge Alan D. Albright, filed Nov. 14, 2025, status Open.

5. CAFC 2026 dockets

No CAFC 2026 docket involving US 9,990,628 / application 13/857,905 was found. I searched the Federal Circuit scheduled-cases listing and Proxense-related appellate filings. What I did find:

• A Proxense, LLC v. Apple Inc., Microsoft Corporation Federal Circuit appeal whose Appellant's Opening Brief is dated April 3, 2026 (Hecht Partners LLP; David L. Hecht and James A. Zak). That appeal arises from PTAB proceedings IPR2024-00233, IPR2024-01334, and IPR2024-00846 and concerns U.S. Patent No. 8,886,954 — not 9,990,628. Source: https://ipfray.com/wp-content/uploads/2026/04/26-04-03-Proxense-v.-Apple-Microsoft-appellate-opening-brief-1.pdf
• Proxense's Fed. Cir. R. 47.4(a)(5) related-case list in that brief mentions district court matters (Samsung — settled; Google — settled; Microsoft — stayed; Apple — stayed) and three ex parte reexaminations (90/015,052 re US 8,352,730; 90/015,053 re US 9,289,905; 90/015,054 re US 10,698,989). 9,990,628 is not listed there and no CAFC 2026 case number is tied to it.
• The Federal Circuit's June 2026 scheduled-cases listing contains no Proxense case, and no case captioned on 9,990,628.

Caveat: My Federal Circuit docket check relied on publicly indexed listings and third-party aggregators rather than a live PACER/CM-ECF pull of the Federal Circuit docket. A case could exist that is not yet indexed. I cannot state with certainty that no CAFC 2026 appeal touching this patent exists.

6. Explicit uncertainties

1. Full granted claim set not verified. I have verified claim 1 only as it appears in publication US 2013/0297514 A1 (same application 13/857,905). I do not have an authoritative full text of the granted claims, the claim count, or the other independent claims of US 9,990,628 B2.
2. Priority date conflict. Google Patents shows prior-art/priority date 2005-11-30; Unified Patents shows 2005-11-29; the specification's own cross-reference points to a 2007 continuation and 2006 provisionals. Reported as found, unresolved.
3. Family/title overlap. Several distinct Proxense patents share the title "Two-Level Authentication for Secure Transactions" (e.g., US 8,433,919 issued 2013-04-30 and US 11,182,792 issued 2021-11-23 appear in the same Google Patents family listings, and US 11,157,909 relates to the 2006-05-05 family). Do not conflate US 9,990,628 with those.
4. No CAFC 2026 appellate activity confirmed for this patent; the only confirmed litigation asserting it is the W.D. Tex. Intel case (7:25-cv-00531).

I'll research litigation for US 9990628 across patent litigation databases.

Let me dig deeper into cases specifically asserting the '628 patent.

Now let me check other Proxense campaigns for this specific patent, plus any PTAB proceedings.

Summary of findings

Based on searches across Google Patents (patent page itself), RPX Insight, Unified Patents PTAB portal, Docket Alarm, Justia Dockets, PACER Monitor, UniCourt, Law360, and Ex Partes complaint analytics, I identified one (1) litigation matter that asserts U.S. Patent No. 9,990,628. I found no PTAB inter partes review, PGR, or litigation docket entry naming the '628 patent other than the case below.

1. Proxense, LLC v. Intel Corporation — the only confirmed '628 case

Grounding sources:

• Google Patents page for US9990628 itself lists the family litigation link: "US case filed in Texas Western District Court … case/7:25-cv-00531" — https://patents.google.com/patent/US9990628/en
• RPX Insight docket page: "7:25-cv-00531 … W.D. Tex. … Filed: 11/14/2025" — https://insight.rpxcorp.com/litigation/txwdce-1172875230-proxense-v-intel
• PACER Monitor case page showing the complaint and exhibits, including "Exhibit F – U.S. Patent No. 9,990,628" and "Exhibit L – 628 Claim Chart" — https://cdn.pacermonitor.com/public/case/61206235/Proxense,_LLC_v_Intel_Corporation_
• Complaint analysis (secondary source) describing the '628 infringement theory against Intel PTT/CSME — https://ai-lab.exparte.com/case/dct/txwd/7:25-cv-00531/doc/analysis/1

Matters I checked that do not assert the '628 patent

Because Proxense, LLC is a serial litigant, the following Proxense campaigns were reviewed and, in each instance where the patents-in-suit were ascertainable, the '628 patent was not among them:

• Proxense, LLC v. [Microsoft Corp.](/litigations/by-plaintiff/Microsoft%20Corp.), No. 6:23-cv-00319 (W.D. Tex.) — asserts '730, '954, '905, '042, '289, '960 (Microsoft's opening claim construction brief, filed 11/06/2023; Google's invalidity contentions in the parallel case).
• Proxense, LLC v. Google LLC, No. 6:23-cv-00320 (W.D. Tex.) — asserts '730, '954, '905, '042, '289, '960 (Google's Final Invalidity Contentions).
• Proxense, LLC v. Apple, Inc., No. 6:24-cv-00143-ADA (W.D. Tex., Judge Alan D. Albright; filed March 18, 2024) — complaint exhibits list '905, '730, '954, '989, '188, '042. No '628.
• Proxense, LLC v. [LG Electronics Inc.](/litigations/by-plaintiff/LG%20Electronics%20Inc.) et al., No. 6:24-cv-00302 (W.D. Tex.; filed May 31, 2024) — complaint asserts '730, '905, '989, '332, '456, '979, '152 and related Wi-Fi patents. No '628 identified.
• Proxense, LLC v. Hyundai Motor Co. et al., No. 6:25-cv-00016 (W.D. Tex.; filed January 14, 2025) — asserts '960, '332, '640 and related patents. No '628 identified.
• Proxense, Inc. v. Wal-Mart Stores, Inc., No. 1:19-cv-06657 (N.D. Ill.; filed October 7, 2019) — asserts the '672 patent (wireless network synchronization). No '628.
• Proxense, LLC v. Target Corp. (filed September 28, 2020) and other 2020 consumer-retail suits — patents-in-suit not identified as including '628 in any source retrieved.
• Proxense, LLC v. Intel Corp., No. 6:24-cv-00283 (W.D. Tex.; filed May 23, 2024) — a separate, earlier Intel suit asserting Wi-Fi/chipset patents (e.g., '129, '672, '043, '152); the '628 patent is not part of that pleading as far as the retrieved excerpts show. The '628 assertion appears only in the later 7:25-cv-00531 action.

PTAB / post-grant proceedings

No IPR, PGR, or CBM petition against U.S. 9,990,628 was found. The PTAB activity in the Proxense portfolio that surfaced in searches targets sibling patents, e.g.:

• Apple Inc. v. Proxense, LLC, IPR2025-00074 ('289) and IPR2025-00562 ('188);
• Apple Inc. v. Proxense, LLC, IPR2025-00600 ('989 — terminated procedurally);
• Microsoft Corp. v. Proxense, LLC, IPR2024-00573 ('042 — all challenged claims held unpatentable, Final Written Decision, 8/4/2025) and IPR2024-00846 (relating to an ex parte reexamination of the '905);
• Google LLC v. Proxense, LLC, IPR2024-00233/00234 (raised in Google's invalidity contentions in 6:23-cv-00320);
• [Samsung Electronics America, Inc.](/litigations/by-plaintiff/Samsung%20Electronics%20America%2C%20Inc.) v. Proxense, LLC, IPR2021-01438 ('188), IPR2021-01439 ('700), IPR2021-01447 ('905 — institution denied).

Caveats (explicit)

1. Coverage limits. I could not exhaustively verify the patents-in-suit for every Proxense case (e.g., Proxense, LLC v. Samsung Electronics Co. Ltd. et al., No. 6:21-cv-210 (W.D. Tex., Judge Alan D. Albright; filed March 5, 2021) — the patents-in-suit were not disclosed in the docket material I retrieved). The '628 patent's claims are not shown as asserted in that case, but I cannot affirmatively rule it out. Sealed or very recently filed complaints, and any suits filed after my retrieval date, would also not appear.
2. Status currency. The status reported for 7:25-cv-00531 reflects the dockets as retrieved (latest activity shown: complaint, AO 120 notice, and summons request, all dated 11/14/2025). No answer, Markman order, stay, transfer, or judgment was identified. PACER should be checked for current entries.
3. Patent identity. The identifier is interpreted literally as US 9,990,628. It issued June 5, 2018, from Application No. 13/857,905 (filed April 5, 2013), claims priority to November 30, 2005, is titled "Two-Level Authentication for Secure Transactions," is currently assigned to Proxense, LLC, and carries a Google Patents "adjusted expiration" of July 17, 2028.

This is a factual docket survey, not legal advice or a validity/infringement opinion.

I'll start by verifying the canonical list and searching for any PTAB activity on this patent.

Key question: any petition specifically against the '628 patent. Let me search more precisely.

Let me check for any proceeding (IPR/reexam) on the '628 or its direct family members.

Proceedings overview

Zero AIA trial proceedings are on file against US 9,990,628. The USPTO Open Data Portal returns no IPR, PGR, or CBM involving this patent, and targeted web searches surfaced no petition challenging it (I checked by patent number, title, and patent owner). The bottom-line defensive posture is therefore untested-but-exposed, not hardened: no claims of the '628 have ever been canceled — there is nothing to cite as a kill shot — but there is also no PTAB record, no institution denial, and no claim-construction or estoppel estoppel footprint to lean on. Do not tell a client the patent "survived two IPRs." It has survived none, because none were filed. The real story is family-level: Proxense's sibling patents in the same specification family are being shredded at the PTAB (all challenged claims of U.S. 8,646,042 unpatentable as of 2025-08-04), which tells you the prior-art well in this space is deep — and that the '628 is the last major family member not yet tested.

Proceedings against US 9,990,628

None. No entries to list. I am not going to manufacture proceeding numbers to fill this template.

For completeness on the "absence is itself a signal" point: the absence here is not a sign the patent is bulletproof. Proxense has been a serial plaintiff since 2021 (Samsung, Microsoft, Google, Apple, LG, Intel), and petitioners have filed dozens of IPRs against other patents in this family. The '628 simply appears to have been reached late — it was first asserted against Intel Corporation in Proxense, LLC v. Intel Corporation, No. 7:25-cv-00531 (W.D. Tex.), complaint filed 2025-11-14, asserting at least independent claim 1 of the '628 (Exhibit F to the complaint; Exhibit L is the '628 claim chart). See the case link at Unified Patents litigation portal and the W.D. Tex. docket record. That assertion is what would normally trigger an IPR within the § 315(b) one-year window — which I flag as the single most time-sensitive item below.

Family-level PTAB activity (NOT on the '628 — do not conflate)

These are proceedings on different patents that share the same Proxense specification lineage. They are relevant to a defendant only as (a) prior-art/ground templates and (b) evidence of how the Board reads this family's disclosure. None of them creates estoppel, cancellation, or defensive value for the '628 itself.

IPR2024-00573 — Microsoft Corporation (joined by Apple Inc.) v. Proxense, LLC (U.S. 8,646,042 B1)

• Type: Inter Partes Review
• Filed: 2024 (Petition served on Proxense; instituted per Paper 11)
• Status: Final Written Decision — all challenged claims unpatentable
• Judge panel: Thu A. Dang, David C. McKone, Norman H. Beamer
• Petition grounds: Claims 1, 5, 6, 8–11, 13, and 14 challenged under 35 U.S.C. § 103
• Final Written Decision: Issued 2025-08-04; oral hearing held 2025-05-13. Verbatim disposition: "Petitioner has shown by a preponderance of the evidence that claims 1, 5, 6, 8-11, 13, and 14 of the '042 patent are unpatentable under 35 U.S.C. § 103." Claims 1 and 10 are the independent claims; 5, 6, 8, 9 depend from claim 1 and 11, 13, 14 depend from claim 10 — so both independent claims and every challenged dependent claim fell.
• Appeal: Not confirmed in the sources I reviewed.
• Defensive value: This is the family's most important data point. The panel invalidated the hybrid-device claims using art including a "two-level authentication for secure transaction" reference (Bloomberg Law summary, 2025-07-18), i.e., the Board considered the two-level-authentication disclosure itself to be old and obvious-adjacent. Link to the FWD: IPR2024-00573 Final Written Decision.

IPR2024-00846 — Microsoft Corporation v. Proxense, LLC (U.S. 8,886,954 B1)

• Type: Inter Partes Review
• Filed: 2024-04-26
• Status: Final Written Decision (docket shows termination 2025-10-17)
• Judge panel: Thu A. Dang, Kevin F. Turner, David C. McKone (Panel also lists Alan D. Albright)
• Institution decision: Instituted 2024-11-18 under § 314(a) on claims 1–29 (Institution Decision PDF). The panel noted the '954 patent was in three district court cases, including Proxense v. Microsoft, No. 6:23-cv-00319 (W.D. Tex.).
• Defensive value: Confirms the Board will institute on this family despite parallel litigation; no Fintiv shelter available to Proxense here.

IPR2024-00783 / IPR2025-00075 — Google LLC v. Proxense, LLC and Apple Inc. v. Proxense, LLC (U.S. 9,679,289 B1)

• Status: Google's petition instituted 2024-11-18 on claims 1–20 (Institution Decision); Apple's copycat petition in IPR2025-00075 and joinder motion granted; IPR2024-00783 terminated as to Google on 2025-02-05, leaving Apple as sole petitioner.
• Companion: Apple's second, non-joined petition, IPR2025-00074 (filed 2024-11-04), drew a § 314(a)/General Plastic discretionary-denial argument from Proxense in its POPR filed 2025-02-26.
• Defensive value: shows the family-level joinder/copycat strategy — petitioners are reusing the same grounds packages (Giobbi-157 + Giobbi-139 + Dua; Broadcom + Giobbi-157) across the family. Those same packages are the natural starting point for a '628 petition, but also the packages that trigger § 325(d).

IPR2025-00327, -00328, -00329 — Intel Corporation v. Proxense LLC

• Filed: December 2024 (Intel's claim-construction exhibits dated 2024-12-16/2024-12-23)
• Status: Institution denied — Acting Director Stewart, Paper 12, 2025-06-26, under the "settled expectations" doctrine
• Reasoning (verbatim): "the challenged patents have been in force over nine years, creating settled expectations, and Petitioner does not provide any persuasive reasoning why an inter partes review is an appropriate use of Board resources."
• Caveat: I could confirm from public exhibits that these three Intel IPRs track the W.D. Tex. Intel case (6:24-cv-00283) involving U.S. 8,129/8,457,672/9,265,043/8,036,152, but I could not confirm from the sources I reviewed that any of the three challenged the '628. Treat that as unverified. Background: ArentFox Schiff analysis; Finnegan/PTAB blog.
• Defensive value: This is the most important non-merits precedent for a '628 petitioner. The '628 issued 2018-06-05 — eight-plus years in force as of today — squarely inside the "settled expectations" window the Director applied to Intel. Any petition must pre-emptively brief why review is warranted.

Also on the family docket (screening context, not verified claim-level)

• IPR2024-00233 (Google v. Proxense, '954) — instituted; Apple joined via IPR2024-01334.
• IPR2024-00234 (Google v. Proxense) — terminated by adverse judgment 2024-11-01.
• IPR2024-00407 / IPR2024-00405 (Microsoft v. Proxense, '289 and '960) — filed 2024-01-16.
• IPR2024-01326 / -01327 / -01333 / -01335 (Microsoft, Apple v. Proxense, various) — multiple concurrent petitions.
• Ex parte reexaminations on siblings: control nos. 90/015,052 and 90/015,053 (Samsung-filed ex parte requests; Proxense fought the '053 request as an improper "roadmap" serial attack). No '628 reexam was found.
• IPR2021-01438 / -01349 / -01444 / -01447 (Samsung v. Proxense) — the 2021–2022 first wave, at least one institution denied.

Strategic summary

Claim status on the '628: every claim is UNTESTED. Claims 1 through 20 (the patent's full claim set) stand exactly as issued on 2018-06-05. Nothing is canceled, nothing is statutorily disclaimed via IPR, nothing is confirmed patentable by the Board. There is no FWD to quote, no disposition to cite in a § 282 notice, and no claim-level record to mine. What is confirmed is that three independent patents in the same specification family — 8,646,042 (all challenged claims unpatentable, 2025-08-04), plus Final Written Decisions on 8,886,954 and 9,679,289 — were attacked successfully or are under sustained attack, using prior art that the Board treated as squarely on point. The '628 shares its specification and its 2005–2006 priority chain with these patents (it is a continuation of Application 11/744,831, which issued as U.S. 8,433,919, "Two-Level Authentication for Secure Transactions"; U.S. 11,157,909 is in turn a continuation of the '628). Sibling invalidation does not cancel the '628 — but it does establish that a POSITA in 2005–2006 had ready access to the same biometric-PDK/reader architectures, and it hands you a pre-vetted art set.

Estoppel landscape: there is none against the '628. Section 315(e)(2) estoppel attaches only to petitioners who obtained an institution decision — and no one has, as to this patent. Microsoft, Google, Apple, Samsung, and Intel all have estoppel exposure on other Proxense patents they petitioned, but a fresh '628 petition by any of them (or by you) is unencumbered by § 315(e)(2) on the '628. Practically, the binding constraints are instead: (1) § 315(b) — if a party was served with a complaint asserting the '628, the one-year clock is running; and (2) § 325(d) / Fintiv / "settled expectations" — the Board's discretionary-denial machinery, which the Director has recently aimed squarely at exactly this posture (issued 2018, in force eight-plus years, NPE asserting old patents late). One more wrinkle worth scoping: because the '628 was filed 2013-04-05 as a continuation claiming benefit of pre-2013 applications, pre-AIA §§ 102/103 likely govern the claims if every claim finds support in the 2005–2007 priority chain — which changes what qualifies as prior art and should be briefed up front rather than assumed.

Pattern signals. Proxense is an active NPE-style enforcer (Samsung 2021 → Microsoft/Google 2023 → Apple/LG 2024 → Intel 2025) and litigates in the W.D. Tex. (Waco/Austin). It litigates hard at the PTAB too — it filed § 1.183 petitions to try to shut down Samsung's serial ex parte reexam, and it runs discretionary-denial arguments routinely. No defensive aggregator (Unified Patents, RPX Indemnity, etc.) appears in the '628 chain; Unified Patents' portals merely index the patent as Proxense's. The petitioner pool is the corporate defendant pool itself — no third party has taken the validity risk on the '628 yet.

Recommended next steps

1. If a demand letter or complaint cites the '628, there is no invalidity shortcut to hand. You cannot say "claim 1 is dead." The patent is fully intact and has never been before the Board. Treat it as a live, untested asset.

2. Time-critical: check the § 315(b) service date. If you (or a privy) were served with a complaint asserting the '628 — Intel was served on/around the 2025-11-14 complaint in 7:25-cv-00531 — the one-year IPR bar under 35 U.S.C. § 315(b) expires roughly 2026-11-14. As of today, 2026-09-19, that leaves under two months for any barred party to file. Confirm the actual service date from the docket before relying on this; do not treat the filing date as the trigger.

3. Front-load the discretionary-denial fight in any petition. Given the Director's 2025-06-26 decision in Intel Corporation v. Proxense LLC, IPR2025-00327, -00328, -00329, Paper 12, a '628 petition must affirmatively argue against "settled expectations" — by showing a material examiner error, that Proxense has not commercialized/asserted/marked the '628 in your technology space, or a significant change in law bearing on the claims. Also pre-empt Fintiv: the parallel W.D. Tex. case will be at or near trial.

4. Borrow the sibling grounds packages, but improve on them. The petitioner packages used against the hybrid-device patents (Giobbi-157 + Giobbi-139 ± Dua; Broadcom EP 1 536 306 + Giobbi-157) are the obvious templates. Expect § 325(d) pushback for reusing them, so pair them with art the Office has not seen. The '042 FWD at IPR2024-00573 is the best available roadmap for how the Board reads this specification.

5. Verify everything on the record before filing. Live sources in this space are moving quickly (IPR2024-00846 docket shows termination on 2025-10-17; IPR2025-00562 shows activity into March 2026). Before relying on any of the above, pull the current status from PTAB E2E, confirm whether any '628 petition has been filed since this analysis, and check the Federal Circuit docket / CourtListener for appeals of the '042, '954, or '289 Final Written Decisions — a CAFC affirmance on the '042 would materially strengthen an obviousness theory against the '628.

Confidence statement. High confidence that no AIA trial proceeding on US 9,990,628 exists as of the ODP ingest date — the structured block says so and independent searching found none. Medium confidence on the characterization of the family-level proceedings (drawn from secondary aggregators and petition PDFs rather than the full Board records). I could not confirm whether Intel's IPR2025-00327/-00328/-00329 targeted the '628; treat that as unresolved.

I'll research the assignment record and litigation history for US 9990628 before drafting the analysis. Let me run several targeted searches.

I found a related-family assignment abstract. Let me dig for the specific '628 record, the correspondent, and Proxense's history.

Key finding: the correspondent is Paige M. Alsbury of Patent Law Works/Proxense in Salt Lake City. Let me try to pin the exact '628 reel/frame and Proxense's litigation footprint.

I hit the tool-call limit, so here is the analysis based strictly on the records I retrieved. One important caveat up front: I could confirm the conveyance, parties, dates and correspondent for the '628 assignment, but I could not retrieve the exact reel/frame for US 9,990,628 itself. The reel/frame below (034843/0145) is confirmed for the sibling family member US 9,298,905 and is offered as corroboration, not as the '628's own number.

Inventors

• John J. Giobbi, Bend, Oregon — sole named inventor. At the time of filing he was the founder/owner of Proxense, LLC (founded 2001), per contemporaneous press coverage (Bend Bulletin, 2008-05-07) and the patent's own assignee record. The '628 is a continuation of Giobbi's application Ser. No. 11/744,831 (filed 2007-05-05), so the inventive work dates to the 2005–2007 window when Proxense was actively developing the Personal Digital Key (PDK) platform.
• No unusual inventor pattern. Because Giobbi is the assignee's founder, there is no "inventor departs within 12 months of filing" signal. The pattern across the Proxense family is the opposite: a single-inventor portfolio (Giobbi alone on '628, '730, '954, '905, etc.), with multi-inventor exceptions only for the Notary/registration applications (Giobbi + David L. Brown + Fred S. Hirt on the '888/'795 line).

Original assignee

• Proxense, LLC, Bend, Oregon (mailing address of record on the sibling '905 assignment: 689 NW Stonepine Dr., Bend, OR 97701). Same entity is listed as current assignee on Google Patents, i.e. title has never left Proxense.
• Primary line of business: proximity-based secure wireless authentication — the "Personal Digital Key" (PDK) fob plus Receiver/Decoder Circuit (RDC) "Reader," aimed at payment, access control, healthcare and casino tracking. Proxense's own litigation states it "developed... mobile payment technologies" from ~2004–2012.
• Did they ship a product embodying the claims? No evidence of a commercial product. The best contemporaneous trace is a 2008 development/tracking deal with Bally Technologies (casino chips/cash boxes/employees). By the time of the 2021+ litigation the company functions as a licensor/litigant; in the Intel IPR papers (IPR2025-00327) the petitioner characterized the challenged Proxense patents as having "never [been] commercialized, asserted, marked, licensed, or otherwise applied."
• Current status: Operating, not acquired, not dissolved, not bankrupt — but operating as a patent enforcement/holding company, not as a product vendor. Active, wide-ranging assertion campaign (see below).

Assignment timeline

Google Patents' legal-events record for US 9,990,628 shows exactly one recorded assignment, plus the original filing event. The family record (US 9,298,905, same inventor, same assignee) is included because it is the corroborating reel/frame I could actually retrieve.

• 2015-10-15 (recorded) — Reel/frame not retrieved for '628 specifically (Google Patents legal events)

  • Conveyance: Assignment of assignors' interest (recorded as "ASSIGNMENT OF ASSIGNEE INTEREST (SEE DOCUMENT FOR DETAILS)")
  • Assignor: GIOBBI, JOHN J.
  • Assignee: PROXENSE, LLC, Bend, OR
  • Correspondent: not captured for this entry; the sibling family assignment at the same firm/period was filed by Paige M. Alsbury, Patent Law Works/Proxense, 201 S. Main St., Suite 250, Salt Lake City, UT 84111 (see recurrence note below)
  • Context: internal confirmatory assignment — inventor perfection of title in his own company; consistent with a 2015 bulk title-cleanup across the Proxense portfolio ahead of enforcement.

• 2015-01-29 (executed and recorded) / mailed 2015-01-30 — Reel 034843 / Frame 0145 (confirmed for sibling US 9,298,905, "Biometric Personal Data Key (PDK) Authentication")

  • Conveyance: ASSIGNMENT OF ASSIGNORS' INTEREST (SEE DOCUMENT FOR DETAILS)
  • Assignor: GIOBBI, JOHN J. (exec. 2015-01-29)
  • Assignee: PROXENSE, LLC, 689 NW Stonepine Drive, Bend, Oregon 97701
  • Correspondent: PAIGE M. ALSBURY, 201 S. Main Street, Suite 250, Salt Lake City, UT 84111 — recurring correspondent. The same Salt Lake City address appears as the "Patent Law Works/Proxense" agent of record throughout the family (e.g. '905 file history shows filer Paige McDede Alsbury and Greg Sueoka, Patent Law Works/Proxense, 201 South Main Street, Suite 250). This is portfolio-wide outside counsel, not a serial shell-entity assembler, but it is a genuine recurring name.
  • Context: internal reorg / title perfection (inventor → operating company), not a third-party transfer.

• 2013-04-05 — Application filed by Proxense, LLC (Google Patents legal events). Continuation of Ser. No. 11/744,831 (priority 2005-11-30).

• 2018-06-05 — Patent granted to Proxense, LLC.

No assignments recorded after 2015. Title rests with Proxense, LLC. (Sibling family members of the '628 also flow onward to continuation applications during 2018 and 2021 — e.g. US 15/973,089 and US 17/504,231 — which are new filings, not assignments.)

Timeline diagram

timeline
    title Ownership of US 9990628
    2005 : Priority application filed by Giobbi
    2007 : Parent application filed by Giobbi
    2013 : Continuation filed by Proxense LLC
    2015 : Giobbi assigns interest to Proxense
    2018 : Patent issued to Proxense LLC
    2021 : Proxense files first assertion suits
         : Against Samsung and others
    2025 : Proxense asserts 628 against Intel
         : In W D Texas

NPE / troll-pattern signals

1. Shell-entity transfer — not present. The only assignment moves the patent from the founder to his own operating company (Giobbi → Proxense, recorded 2015-10-15; sibling reel 034843/0145, 2015-01-29). There is no transfer to an "IP/Licensing/Holdings/Ventures" vehicle, no registered-agent address, no single-purpose LLC. Proxense is the original and still current owner.

2. Known asserter in the chain — present (as the owner itself). Proxense does not appear on the classic Acacia/Marathon/IV/Wi-LAN lists, but it is now surfaced by both Unified Patents and RPX as a high-frequency plaintiff: the Google Patents page carries Unified Patents litigation data for Proxense LLC v. Intel, 7:25-cv-00531 (W.D. Tex., filed 2025-11-14), and RPX Insight carries the same case. Related campaign: Proxense v. Samsung (6:21-cv-210), v. Microsoft (6:23-cv-00319), v. Google (6:23-cv-00320), v. LG (2024). Proxense is the defendant/patent-owner in IPR2021-01349, -01438, -01447 (Samsung), IPR2024-00232/00233/00234 (Google), IPR2024-00573/00846 (Microsoft & Apple), and IPR2025-00327 (Intel). This is a sustained multi-front monetization campaign.

3. Repeat correspondent across the chain — unclear / weak. The '628 chain has a single link, so there is no intra-chain recurrence. The same correspondent does recur across the portfolio: Paige M. Alsbury / Patent Law Works-Proxense, 201 S. Main St., Suite 250, Salt Lake City is the correspondent on the sibling '905 assignment (reel 034843/0145) and is the prosecuting agent of record family-wide (with Greg Sueoka). One name across a company's own portfolio is normal outside-counsel behavior, not the "one lawyer running many unrelated shells" pattern — so I score this weak despite the recurrence.

4. Cascading transfers — not present. Only one assignment; no chained LLCs; no common-address shell relay.

5. Pre-litigation transfer — not present. The recorded assignment (2015) predates the first suit naming the '628 (Intel, filed 2025-11-14) by roughly a decade. This is not a venue/standing set-up transfer.

6. Bankruptcy fire-sale — not present. No Chapter 7/11, no stalking-horse sale; Proxense is not a distressed divestor.

7. Privateering — not present / unclear. No public-company operating assignor transferring to an NPE to assert on its behalf. Proxense is asserting its own original patents, not another company's.

8. Defensive aggregator — not present. The chain does not terminate at RPX, AST, LOT, Unified or OIN. (Note the contrast surfaced in search: an unrelated RPX/Provenance Asset Group chain did terminate at RPX Corporation — that is a different patent family, not the '628.)

Verdict

NPE — moderate confidence

Justification: The assignment record itself is clean — a single inventor-to-company assignment (Giobbi → Proxense, recorded 2015-10-15; corroborating sibling reel 034843/0145, 2015-01-29) with no shell transfers and title still with the original assignee. The NPE character comes not from the assignment chain but from the owner's conduct: Proxense has no evidence of a commercial product embodying the claims and is running a broad assertion campaign against Intel, Microsoft, Google, Samsung and LG (e.g. the '628 asserted as Exhibit F/L in Proxense v. Intel, 7:25-cv-00531, W.D. Tex., filed 2025-11-14), which is a strong signal but a single signal category, landing it at moderate rather than high confidence.

Verification links:

• Assignment Center: https://assignmentcenter.uspto.gov/ (search "9,990,628" or app 13/857,905)
• Alternate index: https://assignment.uspto.gov/patent/index.html
• Google Patents legal events: https://patents.google.com/patent/US9990628/en
• Unified Patents litigation data: https://portal.unifiedpatents.com/litigation/Texas%20Western%20District%20Court/case/7%3A25-cv-00531
• RPX Insight: https://insight.rpxcorp.com/litigation/txwdce-1172875230-proxense-v-intel

Confidence limitations to flag for the file: (a) the exact reel/frame for the '628 assignment is unverified — confirm on Assignment Center before citing; (b) "no product" is inferred from the complaint's own 2004–2012 development narrative and the Intel IPR petitioner's characterization, not from a financial filing (Proxense is private, so no SEC corroboration is available); (c) the correspondent for the '628 entry specifically was not captured — the Alsbury identification rests on the sibling '905 record.

I'll research the front-page references cited for US 9,990,628 and any USPTO records.

Let me pull the Justia record, which typically reproduces the front-page cited references.

Scope, verification, and one date conflict to flag

Task date conflict (flagged, not resolved silently). Your instructions state "Current Date: April 26, 2026." The previously generated sections of this analysis are dated 2026-09-19, and the header of the source document records a fetch at 2026-09-18T18:45:57Z. I am therefore working from a record as it stood on/around 2026-09-19, while your prompt date is April 26, 2026. Where anything below is time-sensitive (e.g., the § 315(b) window discussed in the prior litigation section), the earlier date controls. I did not normalize either date.

Identifier discipline. I searched literally for 9990628 / US 9,990,628 B2 / application 13/857,905. I did not substitute the sibling patents that share the title "Two-Level Authentication for Secure Transactions" (e.g., US 8,433,919 B2, US 11,157,909 B2) or the parent publication US 2007/0245157 A1. Where those family members appear below, they are labeled as family members, not as the patent you asked about.

1. Headline finding — the deliverable you asked for was not retrievable in this session

Your request was: "Look at each patent citation for 9990628. For each reference, provide: full citation, publication/filing date, brief description, and which claim(s) it potentially anticipates under 35 U.S.C. § 102."

I must state plainly: I did not obtain the granted patent's front-page (56) "References Cited" list for US 9,990,628.

Specifically:

• The authoritative full text of US 9,990,628 supplied to me begins at the Abstract/Description and omits the cover-page "(56) References Cited" block, the Field of Search, and the Examiner/Attorney fields. The Google Patents render I searched reproduces the same omission.
• My searches for the Justia record, the "References Cited" block, and the Primary Examiner field for 9,990,628 returned either the same truncated Google Patents content, or results for other patents entirely (e.g., US 8,500,628, US 8,764,326, US 11,344,413 — unrelated art classes). I ran out of search steps before reaching a page that reproduces the '628 cover page.
• I will not infer or reconstruct a (56) list from memory. Any list presented as "the '628's cited references" without that cover page would be fabrication.

Consequence: everything in Sections 2–5 below is offered as the best-grounded substitute material, with the provenance of each group explicitly labeled so you can tell "cited by the applicant/examiner" from "cited by a third party against a sibling patent." Do not merge these categories.

2. Documents referenced inside the '628 specification (self-references / incorporation by reference)

These are grounded in the authoritative patent text supplied to me (the specification's own cross-references). They are cited by the applicant, but for support and incorporation, not as prior art. Their § 102 significance is the inverse of what you asked: they sit in the '628's own priority chain and therefore are not available as prior art against it if every claim finds support back through the chain.

Note on the parent publication. The '628 is a § 120 continuation of U.S. App. Ser. No. 11/744,831 (filed 2007-05-05), which published as US 2007/0245157 A1 (Giobbi, Brown & Hirt, pub. 2007-10-18). That publication is the '628's own parent document. It is nonetheless a genuine third-party-style § 102/§ 103 reference against the sibling patents (see Section 4) — a fact worth carrying into any '628 invalidity theory, because the '628's own parent is art that the Board has already treated as disclosing the two-level scheme.

3. Front-page reference list I could see — but for the parent publication, not the '628

The Unified Patents page for US 2007/0245157 A1 (the '628's parent publication) exposes a "Patent Art (99)" citation list. I reproduce a representative slice below. Caveat: I could not confirm that these are the (56) references of US 9,990,628. They are the citations associated with the parent publication in that aggregator; the '628 was separately examined and issued in 2018.

(The full aggregator list runs to ~99 entries including US 5,629,980; US 5,892,825; US 6,088,730; US 6,332,614; US 6,336,121; US 6,415,307; US 6,425,084; US 6,550,011; US 6,600,899; US 7,192,924; US 7,249,117; US 2002/0138767; US 2003/0046552; US 2006/0022046; and others. I list the most § 102-salient ones only.)

4. Third-party prior art actually litigated against sibling patents in this family

This is the set with real defensive currency — but note carefully: none of these proceedings is against the '628. They are IPRs against US 8,646,042, US 8,886,954, US 9,679,289 and siblings. Provenance: the PTAB petition/exhibit records identified in the earlier sections (IPR2024-00407, IPR2024-00783/-00784, IPR2024-00573, IPR2024-00846).

5. § 102 analysis — mapping, and three things that complicate it

Verified claim text available. Only claim 1 is available verbatim, and only as recited in pre-grant publication US 2013/0297514 A1 (same application 13/857,905):

"1. A method for electronic authentication comprising: wirelessly receiving a biometric profile from a personal digital key (PDK), wherein the biometric profile is uniquely associated with an individual; receiving a biometric input; comparing the biometric profile to the biometric input; and authorizing a transaction responsive to the biometric profile matching the biometric input."

Earlier sections state the patent has 20 claims (1–20). I did not verify the granted text of claims 2–20, so claim-mapping below must be read as "claim 1 as published, plus the described dependent features whose claim numbers I could not confirm."

Element-by-element exposure of claim 1:

Three threshold issues that determine whether any of the above is actually § 102 art:

1. Pre-AIA vs. AIA. The '628 was filed 2013-04-05 as a continuation claiming benefit back to 2005-11-30 / 2007-05-05. If every claim is entitled to the pre-2013 chain, pre-AIA §§ 102/103 govern, which changes the critical date (invention date vs. filing date), the availability of § 102(b) statutory bars, and the scope of § 102(e)/§ 102(g) art. This should be briefed up front, not assumed — the earlier sections flagged the same point.
2. Anticipation vs. obviousness. Only § 102 was requested. On the record I retrieved, 3.1 (US 4,993,068) and 3.2 (US 6,041,410) are the only references I can defensibly characterize as potential anticipations of claim 1's broad recitation — and even those require the "wireless" and "transaction" limitations to be read permissively. Everything else is properly an obviousness theory (Giobbi-157 ± Dua ± Broadcom EP 1 536 306).
3. Priority integrity is the whole ballgame. The '628's strongest asset is that it is a continuation of 11/744,831 (2007-05-05), which itself is a CIP of 11/292,330 (2005-11-30). If the '628 claims find full written-description support back through that chain, references dated 2007+ (including the parent publication itself, 4.1) are pushed out as § 102 art. Conversely, any claim limitation added in the 2013 continuation without antecedent support opens the door to Giobbi-157 as § 102(b) art. This, not any individual reference, is the decisive question.

6. How to obtain the authoritative (56) list (do this before filing anything)

I cannot substitute for the cover page. In priority order:

1. USPTO PatentCenter — https://patentcenter.uspto.gov — search application 13/857,905, "Patent Docs" → the issued patent PDF; the (56) block is on the front page. This is the controlling source.
2. Patent images / official PDF — the earlier section cited https://patentimages.storage.googleapis.com/81/b8/c8/9ea9a2ec589ce0/US9990628.pdf. The cover page there should carry (56), Field of Search, Primary Examiner, and Attorney/Agent.
3. Google Patents "Citations" tab for https://patents.google.com/patent/US9990628/en — machine-parsed, good for bulk, but verify against (2).
4. Justia — https://patents.justia.com/patent/9990628 — reproduces U.S. Patent Documents, Foreign Patent Documents, and Other References per the cover page; my search did not land on this page.
5. File history (Global Dossier / PatentCenter) for the actual PTO-892 / IDS — this distinguishes applicant-cited from examiner-cited art, which matters for § 325(d) arguments down the road.

7. Explicit uncertainties and contradictions

1. (56) list not retrieved — the core limitation. I have no verified front-page citation list for US 9,990,628. Sections 2–4 are substitutes with labeled provenance. Do not represent them as the '628's cited references.
2. Contradiction flagged — priority date. Google Patents renders 2005-11-30 for the '628; Unified Patents renders 2005-11-29 (and lists parent publication US 2007/0245157 A1's application date as 2007-05-04, vs. the specification's "filed on May 5, 2007"). The '628 specification itself recites the 11/744,831 continuation and five 2006–2007 provisionals (60/798,172, 60/798,843, 60/838,788, 60/824,758, 60/894,608). These dates are inconsistent across sources and I did not resolve them. Given that § 102 turns on the critical date, a one- or two-day discrepancy is material and should be resolved from the certified copy.
3. Contradiction flagged — inventor list. The '628 lists John J. Giobbi as sole inventor. The parent publication US 2007/0245157 A1 lists Giobbi, Brown, and Hirt. Consistent with a continuation narrowing the named inventive entity, but it is a discrepancy worth an inventorship check if the parent is relied on as § 102 art (different inventive entity = the parent is not "by another" only if identical — a continuation with a different inventor set can create § 102(e) exposure).
4. Sibling/family conflation risk (reiterated). US 8,433,919 B2 and US 11,157,909 B2 carry the same title. US 11,157,909 is a continuation of the '628. None of these is US 9,990,628, and none of their (56) lists should be attributed to it.
5. Unverified items. EP 1 536 306 A1's date; the identity of "Giobbi-139"; the examiner identity for the '628. All flagged above rather than filled in.
6. No AIA trial exists on the '628 (per the earlier PTAB section), so no petitioner has ever had to plead a § 102 theory against these claims. The § 102 mapping in Section 5 is my own analysis of public reference disclosures against the published pre-grant claim 1 — it is not a validity opinion and is not grounded in any Board or court construction of the '628's claim terms.

I'll ground this obviousness analysis in the actual prior-art references. Let me verify the claim set and pull the key references.

Critical finding — the granted claim 1 differs from the pre-grant publication. Let me verify the prior-art references and the remaining claim set.

Obviousness Analysis — U.S. Patent No. 9,990,628 B2

"Two-Level Authentication for Secure Transactions" (Proxense, LLC)

35 U.S.C. § 103

Prepared: 2026-04-26 (per the task's stated date; note the file's earlier sections used 2026-09-19 — a ~5-month inconsistency in the working record that is flagged, not silently corrected, because the § 315(b) arithmetic in the earlier sections is date-sensitive.)

0. Threshold: a correction to the claim record that changes the entire analysis

The previously generated "Patent summary" section contains a material error that must be flagged before anything else. That section reproduced claim 1 from the pre-grant publication US 2013/0297514 A1 and then stated, correctly hedging, that the granted claim 1 might differ. It does. It differs completely.

The granted claim 1 of US 9,990,628 B2 (as indexed by RPX Insight at https://insight.rpxcorp.com/patent/[US9990628B2](/patent/US9990628B2)) is:

1. A method comprising:

• receiving, at a biometric reader, a biometric input, a personal digital key (PDK) comprising the biometric reader;
• generating, at the PDK, biometric data based on the biometric input;
• retrieving, from a secured memory element of the PDK, a biometric profile sample comprising biometric information, wherein the biometric profile sample is associated with a biometric profile associated with an individual;
• comparing, at the PDK, the information of the biometric profile sample to the biometric data based on the biometric input;
• subsequent to a determination of a match between the biometric data based on the biometric input and the biometric profile sample, retrieving, from the secured memory element of the PDK, the biometric profile associated with the individual;
• comparing, at the PDK, the biometric data based on the biometric input to the biometric profile;
• transmitting, by the PDK via radio, purchasing means information associated with the PDK, the purchasing means information used by a first remote registry administered by a trusted third-party organization in a first validation, the first validation subsequent to determining that the biometric data based on the biometric input matches the biometric profile, wherein a transaction with a merchant is authorized based on the first validation using the purchasing means information associated with the PDK and wirelessly transmitted subsequent to the comparison between the biometric profile and the biometric data based on the biometric input.

Two consequences:

1. The earlier section's speculation was wrong. It predicted the granted claim might not recite the "sample-then-full-profile" two-stage comparison or the registry lookup. In fact the granted claim recites both, and adds the PDK-resident biometric reader, the PDK-side comparison, the radio transmission of purchasing means information, and validation by a third-party-administered remote registry. This matches (a) the '628 specification's FIG. 7A (steps 708/710 sample comparison → 712/714 full comparison), FIG. 1/7D (validation database 112, Central Registry 114), and ¶¶[0011], [0032]–[0034], [0064], [0071]–[0072]; and (b) the Intel complaint's own characterization of the '628 (Ex Parte complaint analysis, https://ai-lab.exparte.com/case/dct/txwd/7:25-cv-00531/doc/analysis/1).
2. The title is a misnomer as to claim 1. Claim 1 does not recite the "device authentication" first layer at all. Its "two levels" are the sample comparison and the full-profile comparison — plus a registry validation. That matters for obviousness because the claim's actual advance is narrower and more mundane than its title suggests.

Caveat: RPX Insight is a secondary aggregator. The claim set should be confirmed against the printed patent (https://patentimages.storage.googleapis.com/81/b8/c8/9ea9a2ec589ce0/US9990628.pdf, which the cover page confirms is "20 Claims, 8 Drawing Sheets"). Everything below is keyed to claim 1 as recited; dependent claims 2–20 are unverified here.

1. The § 103 framework and the governing statutory scheme

1.1 Which law applies

The '628 issued 2018-06-05 from App. No. 13/857,905, filed April 5, 2013 — after the AIA's March 16, 2013 changeover. But it is a straight § 120 continuation of App. No. 11/744,831 (filed May 5, 2007). Under AIA § 3(n)(1), first-inventor-to-file applies only to applications containing a claim with an effective filing date on or after March 16, 2013. If every '628 claim is supported by the 2005–2007 disclosure chain (11/744,831 → 11/292,330, filed Nov. 30, 2005, plus the five 2006–2007 provisionals), pre-AIA §§ 102/103 govern. This is the earlier "PTAB challenges" section's own conclusion and it is correct — but it must be proven, not assumed, because it drives (i) what qualifies as prior art and (ii) whether § 103(c) common-ownership disqualification is available.

The stakes are stark: pre-AIA § 103(c)(1) bars the use, for obviousness, of subject matter that qualifies as prior art only under § 102(e), (f), or (g), where the subject matter and the claimed invention were commonly owned at the time the invention was made. As shown in § 4.1 below, that provision is the single biggest threat to any attack grounded on the flagship reference from this family's IPRs.

1.2 Legal standard

Obviousness requires the Graham findings (scope/content of the claims, prior art, differences, level of ordinary skill) plus an articulated reasoning with a rational underpinning for combining references. Graham v. John Deere Co., 383 U.S. 1 (1966); KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398, 421 (2007) ("if a technique has been used to improve one device, and a person of ordinary skill in the art would recognize that it would improve similar devices in the same way, using the technique is obvious"); In re Kahn, 441 F.3d 977, 988 (Fed. Cir. 2006). KSR expressly authorizes (a) combining prior-art elements according to known methods to yield predictable results, (b) simple substitution of one known element for another, and (c) "obvious to try."

1.3 Level of ordinary skill in the art (POSITA)

Adopt the uncontested formulation from the family IPRs (Microsoft's IPR2024-00405 Petition, § I.D, https://ai-lab.exparte.com/documents/ptab/ptacts-uspto/[1555152](/patent/1555152)/.../IPR2024-00405_doc_2_exparte.pdf): as of the 2005–2007 window, a POSITA would have a bachelor's degree in computer science, computer engineering, electrical engineering, or a related field, plus at least two years of industry experience in encryption and security, with graduate education substituting for experience and vice versa. This is a low-ish skill bar — favorable to the challenger.

2. The prior art of record on this page

These are the four references the Proxense-family IPRs actually relied on. I use them because the task directs me to the prior art already on this page, and because the Board has already accepted their content and combinability in three sibling proceedings.

Motivation internal to the art (not hindsight): Giobbi-157 and Giobbi-139 share an inventor and the identical problem statement — proximity-based authentication of a carried token, using a portable key plus a reader/decoder, over a short-range wireless link, with remote account validation. The Board accepted exactly this rationale in the sibling trials: the references "are directed to similar operations having similar purposes," "describe similar structures for performing those operations (e.g., PDK and RDCs)," and "combining the techniques… would have worked in the same manner already described in the references." (IPR2024-00405 Petition, § IV, quoting MS-1003 ¶145.) That is a KSR "same field / predictable result" rationale, and it is on the record.

3. Element-by-element analysis of granted claim 1

Notation: [1.0]–[1.7] track the limitations as quoted in § 0.

Result on paper: claim 1 reads on Giobbi-157 essentially in its entirety, with Giobbi-139/Broadcom supplying corroboration for the radio-transmitted "purchasing means information validated by a remote third party" element as an alternative to Giobbi-157's ¶¶[0032]–[0034]/[0075].

4. The four combinations, ranked by realistic viability

4.1 Ground 1 — Giobbi-157 (alone or + Giobbi-139, + Dua) ⚠️ Highest theoretical coverage, but legally fragile

Why it works technically. Giobbi-157 discloses every limitation of claim 1, including the PDK-resident biometric reader, the PDK-side sample→full-profile two-stage comparison, and the third-party-validated purchasing-means transfer. If it were available art, claim 1 would fall on anticipation (§ 102) or, at minimum, obviousness over Giobbi-157 alone (a single reference fully disclosing the claimed method requires no "combination" at all).

Why it is fragile — and why this is the most important point in this memo. Giobbi-157 is the publication of application 11/744,831, the very application of which the '628 is a § 120 continuation. Compare:

• Giobbi-157 front page: "Appl. No.: 11/744,831; Filed: May 5, 2007; Continuation-in-part of application No. 11/292,330, filed on Nov. 30, 2005" — https://www.docketalarm.com/cases/PTAB/IPR2021-01439/SAMSUNG_ELECTRONICS_AMERICA_INC_v._Proxense_LLC/docs/08-26-2021-Petitioner/Exhibit-1004-Giobbi_157.pdf
• '628 front page: "(63) Continuation of application No. 11/744,831, filed on May 5, 2007, now Pat. No. 8,433,919." — https://patentimages.storage.googleapis.com/81/b8/c8/9ea9a2ec589ce0/US9990628.pdf

Consequences:

1. If the '628's claims are entitled to the Nov. 30, 2005 effective filing date, Giobbi-157 (filed May 5, 2007; published Oct. 18, 2007) is not prior art under any subsection — it postdates the invention. Full stop.
2. If the '628 is only entitled to May 5, 2007, Giobbi-157 was filed the same day — and § 102(e) requires a reference "filed… before the invention by the applicant." Same-day filing does not satisfy "before." Giobbi-157 still is not § 102(e) art, and it is not § 102(b) art (published Oct. 18, 2007, well inside any grace period computed from a 2005 or 2007 date).
3. Even if a challenger could force a § 102(e) theory, § 103(c)(1) applies: Giobbi-157's inventive entity (Giobbi + Brown + Hirt) is not identical to the '628's (Giobbi), so it is "by another"; it qualifies as prior art only under § 102(e); and both the reference and the claimed invention were owned by Proxense, LLC at the time the invention was made. Pre-AIA § 103(c)(1) therefore disqualifies Giobbi-157 as § 103 art.

Why the sibling IPRs did not surface this. In IPR2024-00573 (Microsoft v. Proxense, '042), the patent owner's POPR attacked the petition only on § 112(f)/means-plus-function claim construction and never raised § 103(c). The Board observed: "Patent Owner does not present arguments in the Response addressing the specific merits of Petitioner's contentions" and "Petitioner's positions… stand unopposed as a part of this trial." The Giobbi-157 ground succeeded by default, not by overcoming § 103(c). A competent '628 petitioner must assume Proxense has learned this lesson and will raise § 103(c) immediately.

Practical takeaway: do not build a '628 petition on Giobbi-157 as the primary reference. Use it only as a secondary/back-up, and pre-brief the § 102(e)/§ 103(c) fight (e.g., by arguing the '628's effective date, or that some claim lacks § 112 support in the 2005–2007 chain so as to shift the critical date — though note the same-day problem in ¶2 survives that argument).

4.2 Ground 2 — Giobbi-139 + Broadcom (+ Dua) ✅ The most defensible combination

References: Giobbi-139 (US 2004/0255139 A1, published Dec. 16, 2004) and Broadcom (EP 1 536 306 A1, published Jun. 1, 2005).

Availability. Giobbi-139 published more than one year before Nov. 30, 2005, so it is § 102(b) art — not "only" § 102(e)/(f)/(g) art — and therefore § 103(c) does not disqualify it despite the common Proxense ownership. It is squarely available. Broadcom is third-party art, available under § 102(a) (published Jun. 1, 2005, before the Nov. 30, 2005 invention date) if not § 102(b).

Coverage:

• Giobbi-139 supplies: a portable carried key with encrypted data in secured memory; a wireless radio link between key and reader; the device-side comparison of a stored code against a read code, gated on it matching; and remote validation by a third-party "key provider" against a secure account database.
• Broadcom supplies: an RFID token storing credentials and a remote service provider, distinct from the merchant, that authenticates those credentials.

The gap. Neither reference teaches biometrics. That gap must be closed — either by a third reference, or by relying on the well-established state of the biometric-art as of 2005 (fingerprint/PIN/picture/signature verification systems, e.g. Prokoski US 6,850,147; Piosenka US 4,993,068; Hatakeyama US 6,873,975 — all cited on the face of the '628's sibling US 9,542,542). A POSITA combining "carried token + local code comparison + remote credential validation" with then-routine "store a biometric template on the token and compare it locally" is a textbook KSR combination of known elements with predictable results.

Motivation to combine (articulable, record-based):

1. Same field, same problem. Both references address fraud-prevention in transportable-token authentication; the Board in the siblings found them "directed to similar operations having similar purposes" with "similar structures (PDK and RDCs)."
2. Known problem, known fix. The '628's own background (col. 1 of the specification, quoted in the earlier "Patent summary") identifies the exact problem a POSITA would be solving: conventional biometric schemes expose personal biometric data to theft because the template sits in an enrolling party's database and requires repeated enrollment. Giobbi-157's own summary states the goal: "personal biometric information need not be stored in any local or remote storage database and is only stored on the user's own PDK." That is a motivation statement written into this very family — usable without hindsight.
3. Predictable result. Putting the template and the comparison on the carried device was an engineering choice, not an inventive leap: it eliminates a network round-trip, avoids transmitting raw biometrics, and mirrors Giobbi-139's own architecture (secret data lives on the key; the reader only reads and compares).
4. Registry validation is additive and independently known. Broadcom and Giobbi-139 both teach it; Giobbi-157 (to the extent usable at all) describes a Central Registry "administered by a trusted third-party organization" and a credit-card validation database "separate from the merchant." A POSITA would keep it for exactly the anti-fraud reason the '628's FIG. 7D gives.

Chart for Ground 2 (representative):

Weakness to anticipate. Claim 1 requires the same device (the PDK) to contain the biometric reader, hold the sample and the full profile, perform both comparisons, and transmit the purchasing means info for third-party validation. A petitioner must show the POSITA would have unified those functions in one carried token — i.e., that on-token match-on-card plus on-token credential storage was an obvious consolidation. That is arguable but is the soft spot.

4.3 Ground 3 — Giobbi-157 + Giobbi-139 + Dua (the "family template") ⚠️

This is the exact ground set that worked in IPR2024-00573 ('042 FWD, 2025-08-04, all challenged claims unpatentable) and IPR2024-00405 ('960). It is the best-pre-vetted base, because the Board accepted the content and the "similar purpose / similar structure" motivation. But it inherits Ground 1's § 102(e)/§ 103(c) defect for the Giobbi-157 leg. If Proxense raises § 103(c) and the Board applies it, this ground collapses to Giobbi-139 + Dua — under which biopsy coverage is thin unless Dua or another reference supplies biometrics. Verify Dua's actual disclosure before relying on this ground.

4.4 Ground 4 — Broadcom alone ⚠️

Broadcom reaches the "token + stored credentials + third-party service-provider validation" half of claim 1 but does not appear to reach the biometric-capture, sample-comparison, or full-profile-comparison limitations. Use only in combination.

5. Secondary considerations

• No nexus-based secondary considerations are available on this record. The Intel IPR petitioner characterized the challenged Proxense IPRs as art that had "never [been] commercialized, asserted, marked, licensed, or otherwise applied" (IPR2025-00327 papers), and Proxense's own litigation narrative places product development in 2004–2012 with no evidence of a shipped '628-embodying product. Without a product, there is no objective-indicia nexus. If Proxense points to its 2013 licensing-notice correspondence with Intel, the response is the Federal Circuit's rule that licensing unrelated to the claimed invention earns no weight (and a notice letter is not a nexus).
• Unexpected results: none asserted or evidenced in the record; the two-stage sample-then-full comparison yields exactly the expected speed/security tradeoff.
• Long-felt need: the need (avoiding biometric-data custody) was known and articulated in the '628's own background and in Giobbi-157's summary — which undercuts "long-felt but unsolved" rather than supporting it.

6. Defenses Proxense will run, and how to pre-empt them

1. § 103(c)(1) common-ownership disqualification of Giobbi-157. Address head-on; lead with third-party/§ 102(b) art. (§4.1.)
2. § 112(f) claim construction. Proxense's standard play is to argue MPF for family terms ("module," "engine," etc.). It lost that argument in IPR2024-00573 — the panel refused to apply § 112(f) to non-"means" claim language and rejected Proxense's proposed constructions. The '628's claim 1 uses "secured memory element" and "biometric reader"; expect an attempt to convert "element"/"reader" into nonce terms. Note the W.D. Tex. court's construction of "personal digital key" (Proxense v. Microsoft, 6:23-cv-00319, order 2024-05-24: "an operably connected collection of elements including an antenna and a transceiver… and a controller and memory for storing information particular to a user") — which is consistent with an on-token architecture.
3. § 325(d) / "settled expectations" discretionary denial. The Director denied Intel's IPRs in IPR2025-00327/-00328/-00329 (Paper 12, 2025-06-26) because the challenged patents had "been in force over nine years." The '628 issued 2018-06-05 — over eight years in force today. Any petition must affirmatively brief why review is warranted (material examiner error; art never considered).
4. Written description/enablement for Broadcom's token-based credential transmission. The Mapping of Broadcom's "service provider separate from the merchant" was expressly argued and accepted in the sibling petition; reuse the same mapping in the '628 petition but verify Broadcom's text yourself (I could not).
5. Priority/§ 112 challenge backfiring. Attacking the '628's priority to the 2005 provisional does not cure the same-day Giobbi-157 problem (see §4.1 ¶2). Do not build the petition on that move alone.

7. Explicit uncertainties (do not rely on these as verified)

1. The granted claim set beyond claim 1 is unverified. Claim 1 is from RPX Insight (secondary source); the other 19 claims and the independent/dependent breakdown are not established here. Confirm against the printed patent and PAIR before using this analysis.
2. Dua (US 9,042,819) and Broadcom (EP 1 536 306 A1) disclosures are characterized from sibling petitions, not from my own reading of the references. Their specific teachings — and whether each qualifies under § 102(a)/(b)/(e) against the '628's actual effective date — must be verified.
3. The '628's effective priority date is contested and unresolved on this record (Google Patents shows 2005-11-30; Unified Patents shows 2005-11-29; the '628's own § 120 chain runs through 11/744,831 → 11/292,330). Everything about prior-art availability in § 4.1 turns on this. Resolve it first.
4. I could not confirm what the '628's examiner considered. The sibling petitions established that the examiner never considered Giobbi-157/139/Dua/Broadcom in the '042/'960 patents; that does not establish the same for the '628. Pull the '628 file history.
5. § 102(e)/§ 103(c) analysis above is my own legal reasoning, not a holding. The Board never reached it in the siblings because Proxense did not raise it. Treat it as an identified risk, not an adjudicated fact.
6. Note the date inconsistency between this task's "April 26, 2026" and the earlier sections' "2026-09-19." The § 315(b) one-year window discussed in the earlier "Recommended next steps" is date-sensitive; do not mix the two.