US 8566164

I'll search for authoritative information on US 8566164, including any 2026 CAFC docket activity.

Let me get the exact claim text and check for 2026 appellate activity.

Let me confirm the independent claim text for 8566164 specifically and check whether it appears in any 2026 CAFC appeal.

US 8,566,164 B2 — Summary

Bibliographic data (per the patent front matter / Google Patents)

Related family/reference: the specification states the disclosure "may be related to U.S. non‑provisional application Ser. No. 11/736,544 filed Apr. 17, 2007" (incorporated by reference), which issued as US 7,861,260. Continuation filings traced from 11/968,117 include US 12/981,925 (which issued as US 8,595,069), which in turn spawned further continuations.

Abstract (verbatim)

"In a computer-implemented method, a selected online advertisement is automatically directed to, or online activity is automatically tracked and recorded from, an online user interface device which corresponds, at the time the advertisement is directed or the activity is tracked, to a first online access identifier. The advertisement is directed or the activity is tracked based on information from a user profile that references or includes a first set-top box identifier. That information is derived at least in part from automatically collected data related to user behavior with respect to at least one television advertisement delivered to a set-top box, which set-top box corresponds, at the time of that user behavior, to the first set-top box identifier. The first online access identifier is associated with the first set-top box identifier, by being referenced or included in the user profile, by being associated in a database, or otherwise."

Plain-language overview of the independent claim(s)

Independent claim 1 (method claim — text as quoted in W.D. Tex. litigation pleadings):

"1. A method implemented using a programmed hardware computer system coupled to the Internet, the method comprising:
(a) with the computer system automatically causing, in response to online activity from a first online user interface device, a first selected online advertisement to be directed to the first online user interface device, which online user interface device corresponds, at a time the first selected online advertisement is directed, to a first online access identifier;
(b) wherein the first selected online advertisement is selected based on information from a first user profile, which first user profile references or includes a first set-top box identifier that the computer system has electronically associated with the first online access identifier;
(c) wherein the information in part (b) is derived at least in part from automatically collected first data related to user behavior with respect to at least one television advertisement delivered to a set-top box, which set-top box corresponds, at a time of that user behavior, to the first set-top box identifier;
(d) wherein the first set-top box identifier and the first online access identifier are associated without using personally identifiable information …"

Plain language: a computer system server watches what a TV set‑top box did with a television ad (e.g., the viewer watched or interacted with it), keeps a user profile keyed to that set‑top box identifier, links that set‑top box identifier to the household's online access identifier (e.g., an IP address or cookie), and then — triggered by online activity on the household's computer/device — automatically picks and delivers a tailored online ad to that device based on the TV‑ad‑derived profile information. Critically, the set‑top box identifier and online access identifier are associated without using personally identifiable information.

Uncertainty flags (important)

• Full claim set not authoritatively verified here. The claim‑1 text above comes from an infringement exhibit quoting the '164 patent in the W.D. Tex. case; the final clause (d) is truncated in the retrieved source (it continues after "…the set-top box"), so I cannot state its exact completion with confidence. I did not retrieve the complete, verbatim official claim listing (all independent and dependent claims) from USPTO PatentCenter/Patent Full‑Text in this session, so I cannot confirm how many independent claims there are or whether any independent claim is directed to the "tracking and recording online activity" alternative recited in the abstract. Treat the "each independent claim" enumeration as limited to claim 1 unless independently verified.
• Assignee: the record shows Intent IQ, LLC as current assignee, with AlmondNet, Inc. appearing historically (and AlmondNet, Inc. plus Intent IQ, LLC appear together as co‑plaintiffs/patent owners in later litigation). Do not treat "Intent IQ LLC" alone as the sole owner without checking the assignment record.

Litigation / PTAB activity on THIS patent (as listed on the patent record)

• PTAB IPR2022‑01506 — filed, not instituted.
• PTAB IPR2022‑00953 — filed, not instituted.
• W.D. Tex.: 6:21‑cv‑00891; 6:21‑cv‑00876; 6:21‑cv‑00731 (AlmondNet/Intent IQ campaigns asserting 8,566,164 among other patents).
• D. Del.: 1:25‑cv‑00406; 1:24‑cv‑01259.
• N.D. Cal.: 4:22‑cv‑07515.

CAFC 2026 dockets

I searched specifically for US 8,566,164 at the Federal Circuit and did not find a 2026 CAFC docket that affirmatively identifies this patent. I therefore cannot confirm any 2026 appellate proceeding on the '164 patent; state any such assertion as unverified. For transparency, the 2026 Federal Circuit activity I did find involving the same patent owner concerns other members of the Intent IQ / AlmondNet portfolio, not confirmed to be the '164 patent:

• AlmondNet, Inc. v. Meta Platforms, Inc., No. 24‑1834 — nonprecedential judgment affirmed (Fed. Cir. R. 36), April 13, 2026; appeal from PTAB IPR2022‑01436 (a docket number that does not match the '164 patent's IPRs listed above). https://law.justia.com/cases/federal/appellate-courts/cafc/24-1834/24-1834-2026-04-13.html
• AlmondNet, Inc. v. Amazon.com, Inc., No. 25‑1967 — appeal voluntarily dismissed (FRAP 42(b)); mandate issued April 6, 2026.
• Meta Platforms, Inc. v. AlmondNet, Inc., No. 25‑1662 — consolidated with 25‑1638 (filed 2025).
• Intent IQ, LLC v. Viant Technology LLC, Nos. 26‑1701 and 26‑1702 — filed April 20, 2026, appeals from PTAB (26‑1701 originates from IPR2025‑00128). Patent(s) at issue not confirmed to be 8,566,164.

Bottom line: US 8,566,164 is an Intent IQ (originally AlmondNet‑assigned) patent, filed 2007‑12‑31 and issued 2013‑10‑22, directed to cross‑media targeting in which a profile built from a viewer's behavior toward a television ad is used to automatically select and deliver a targeted online ad to the same household's online device, with the set‑top‑box identifier and online access identifier linked non‑PII. It has been asserted in district‑court litigation and challenged (unsuccessfully at institution) in two 2022 IPRs; I found no 2026 CAFC docket specifically tied to this patent, and I flag that the complete independent‑claim set was not authoritatively retrieved.

I'll research litigation involving US 8566164. Let me start with searches across patent litigation sources.

Let me dig into the specific cases and PTAB proceedings identified.

Litigation Involving U.S. Patent No. 8,566,164 ("the '164 patent")

Patent: U.S. 8,566,164 B2 — "Targeted online advertisements based on viewing or interacting with television advertisements"

• Inventors: Roy Shkedi; Ronen Shlomo
• Original/current assignee: Intent IQ, LLC (with related patents held by AlmondNet, Inc.)
• Filed: Dec. 31, 2007; Granted: Oct. 22, 2013; Adjusted expiration Jan. 28, 2032
• Note: Intent IQ and AlmondNet are affiliates — AlmondNet is the parent, and the two entities frequently appear as co-plaintiffs.

Below is what I could verify. Where I could not confirm a detail, I say so rather than guessing.

District Court Litigation

1. AlmondNet, Inc. & Intent IQ, LLC v. Roku, Inc. (W.D. Tex.)

• Plaintiffs: AlmondNet, Inc.; Intent IQ, LLC
• Defendant: Roku, Inc.
• Jurisdiction: U.S. District Court, Western District of Texas, Waco Division (Judge Alan D. Albright)
• Case No.: 6:21-cv-00731-ADA
• Filed: July 15, 2021
• Patents asserted included: '639, '586, '164, '069, '398
• Status/outcome: Stalled by Roku's first-to-file motion; the case was stayed pending resolution of the parallel Delaware declaratory-judgment action (1:21-cv-01035). It was ultimately closed by joint dismissal on June 20, 2025 (Joint Motion to Dismiss, Dkt. 28; Order for Dismissal, Dkt. 30).

2. AlmondNet, Inc. v. Roku, Inc. (W.D. Tex. → D. Del.)

• Plaintiff: AlmondNet, Inc.
• Defendant: Roku, Inc.
• Jurisdiction (original): W.D. Tex., Waco Division (Judge Albright)
• Case No.: 6:21-cv-00876-ADA
• Filed: Aug. 20, 2021
• Status/outcome: Transferred to the District of Delaware as Case No. 1:22-cv-01540-MN, then consolidated into the lead Delaware case 1:21-cv-01035 (MN).

3. Roku, Inc. v. AlmondNet, Inc. & Intent IQ, LLC (D. Del.) — Declaratory Judgment / Counterclaim

• Plaintiff: Roku, Inc. (DJ action)
• Defendants/Counter-Claimants: AlmondNet, Inc.; Intent IQ, LLC
• Jurisdiction: U.S. District Court for the District of Delaware (Judge Maryellen Noreika)
• Case No.: 1:21-cv-01035-MN (consolidated lead case, incorporating 1:22-cv-01540)
• Filed: July 15, 2021
• '164 relevance: Appears in the consolidated patent list; AlmondNet/Intent IQ counterclaimed for infringement.
• Status/outcome: Docket flags show "CLOSED, LEAD, PATENT." Substantial activity through 2025 (e.g., Dkt. 244, Apr. 22, 2025; Dkt. 252, June 20, 2025), including briefing on jurisdiction over "DJ-only"/"dropped" patent claims. Both sides' claims were ultimately dismissed in June 2025 alongside the Texas case.

4. AlmondNet, Inc. & Intent IQ, LLC v. Samsung Electronics Co., Ltd. et al. (W.D. Tex. → N.D. Cal.)

• Plaintiffs: AlmondNet, Inc.; Intent IQ, LLC
• Defendants: [Samsung Electronics Co.](/litigations/by-defendant/Samsung%20Electronics%20Co.), Ltd.; Samsung Electronics America, Inc.; AdGear Technologies Inc.
• Jurisdiction (original): W.D. Tex., Waco Division (Judge Albright)
• Case No.: 6:21-cv-00891-ADA
• Filed: Aug. 26, 2021
• Patents asserted included: '307, '822, '582, '164 (held by Intent IQ), '139
• Status/outcome: Transferred to the Northern District of California (Sealed Order granting motion to transfer, Nov. 28, 2022, Dkt. 65/66) — see next entry.

5. AlmondNet, Inc. et al. v. Samsung Electronics Co., Ltd. et al. (N.D. Cal.)

• Plaintiffs: AlmondNet, Inc.; Intent IQ, LLC
• Defendants: Samsung Electronics Co., Ltd.; Samsung Electronics America, Inc.; AdGear Technologies Inc.
• Jurisdiction: U.S. District Court, Northern District of California
• Case No.: 4:22-cv-07515
• Filed: Nov. 2022 (post-transfer)
• Status/outcome: This is the post-transfer home of the W.D. Tex. 6:21-cv-00891 action. I could not confirm a final disposition from the sources reviewed.

6. AlmondNet, Inc. et al. v. Mediaocean LLC et al. (D. Del.)

• Plaintiffs: AlmondNet, Inc.; Intent IQ, LLC
• Defendant(s): Mediaocean LLC (and related entities per case caption)
• Jurisdiction: U.S. District Court for the District of Delaware (Judge Maryellen Noreika)
• Case No.: 1:25-cv-00406-MN
• Filed: April 1, 2025
• Patents asserted included: 10,321,198; 11,949,962; 8,566,164; 8,595,069; 8,677,398
• Status/outcome: Pending (open) as of the most recent data reviewed.

7. Delaware Case No. 1:24-cv-01259 (D. Del.)

• Jurisdiction: D. Del.
• Status: Listed in the Google Patents "family has litigation" data for the '164 family, but I could not confirm the parties, filing date, or outcome from the sources retrieved. Flagging as unverified.

Related consolidated D. Del. actions (lead case 1:24-cv-00831-MN, Judge Noreika): LiveIntent (1:24-cv-00831), Connatix (1:24-cv-01296), Zeta (1:25-cv-00579), Criteo (1:25-cv-01266), and Samba TV (1:25-cv-01084). Per the July 31, 2026 consolidation order these were consolidated for all pretrial purposes. Whether the '164 patent specifically is asserted in each of these is not fully confirmed from my sources; the '164-family patents are asserted across the portfolio of Intent IQ/AlmondNet suits generally.

PTAB (Inter Partes Review) Proceedings

Both petitions directed to the '164 patent were denied institution ("Not Instituted – Merits"):

Per the Google Patents family page, both are marked "Not Instituted - Merits," with links at:

• https://portal.unifiedpatents.com/ptab/case/IPR2022-01506
• https://portal.unifiedpatents.com/ptab/case/IPR2022-00953

Caveat: I could not conclusively identify the petitioner(s) for these two '164-specific IPRs from the sources retrieved. (Related IPRs in this portfolio were filed by Meta Platforms (IPR2022-00773), Roku (IPR2022-01236), and Microsoft (IPR2022-01420), but those are directed at other patents in the family, not the '164 patent, so I do not attribute them to the '164 proceedings.)

Summary Table of '164 Litigation

Notes on Confidence and Limitations

• Confirmed: The case numbers, jurisdictions, filing dates, and party identities for Items 1–6, and the existence of the two denied '164 IPRs, are supported by court filings (AO 120 forms), PTAB dockets, Docket Alarm, and the Google Patents litigation records fetched from the patent page.
• Not confirmed: (a) Petitioners for IPR2022-00953 and IPR2022-01506; (b) the parties/outcome in D. Del. 1:24-cv-01259; (c) the current final disposition of the N.D. Cal. Samsung/AdGear action; and (d) whether the '164 patent is asserted in every one of the 2024–2026 consolidated Delaware ad-tech suits.
• Because this patent is held by a non-practicing-entity-style monetization portfolio (AlmondNet/Intent IQ) that has litigated broadly, additional, more recent suits may exist that did not surface in these searches. I recommend confirming current status via PACER/Unified Patents Docket and the PTAB Portal directly.

Primary grounding sources:

• Google Patents US8566164B2 (litigation family data): https://patents.google.com/patent/US8566164/en
• Unified Patents patent page: https://portal.unifiedpatents.com/patents/patent/8566164
• PTAB Portal IPR dockets: https://portal.unifiedpatents.com/ptab/case/IPR2022-01506 and https://portal.unifiedpatents.com/ptab/case/IPR2022-00953
• Docket Alarm case records for W.D. Tex. 6:21-cv-00731, 6:21-cv-00876, 6:21-cv-00891; D. Del. 1:21-cv-01035, 1:25-cv-00406; N.D. Cal. 4:22-cv-07515.

I'll start by reconciling the structured ODP block against the litigation data embedded in the patent's own front matter, then research each proceeding.

Key lead: IPR2022-01506 appears to be a Samsung petition against the '164 patent. Let me pin down both proceedings.

Two proceedings confirmed by name, petitioner, and disposition. Now I need filing dates, panels, grounds, and whether anything else hit the '164 patent.

PTAB Proceedings — US 8,566,164 ("Targeted online advertisements based on viewing or interacting with television advertisements")

Threshold note on the source data (read this first). The structured "PTAB proceedings on file" block asserts that the USPTO ODP API returns no AIA trial proceedings for this patent. That is contradicted by the patent's own front matter in the authoritative full text, which lists two PTAB cases under "Family has litigation":

• PTAB case IPR2022-01506 filed (Not Instituted - Merits) — https://portal.unifiedpatents.com/ptab/case/IPR2022-01506
• PTAB case IPR2022-00953 filed (Not Instituted - Merits) — https://portal.unifiedpatents.com/ptab/case/IPR2022-00953

I am treating the ODP "no proceedings" result as a stale/incomplete ingest and flagging both proceedings, per your instruction to flag anything web search surfaces that ODP hasn't indexed. Both are independently corroborated by (a) the Patexia docket for IPR2022-00953, (b) petition exhibits filed in later, unrelated IPRs that recite the disposition of each proceeding, and (c) the Google Patents front matter. I have not verified either against PTAB E2E directly.

Proceedings overview

Two AIA trial proceedings are on file against US 8,566,164 — IPR2022-00953 (Roku, Inc.) and IPR2022-01506 (Samsung Electronics Co., Ltd.) — and both were denied institution, leaving the patent entirely intact with no claim ever canceled, no Final Written Decision issued, and all claims (including independent claim 1) still live and wholly untested on the merits. Defensively, this is a "patent owner prevailed, twice, at the institution stage" posture: the two most obvious prior-art attacks were screened out on the merits with no estoppel attaching, which means a defendant today inherits an un-exhausted prior-art space but also gains no free invalid-claims argument — you will have to build a materially better petition than Roku's or Samsung's, from scratch.

IPR2022-00953 — Roku, Inc. v. Intent IQ, LLC

• Type: Inter Partes Review (35 U.S.C. §§ 311–319)
• Filed: 2022-05-24
• Status: Not Instituted – Merits (Google Patents front matter, verbatim). Plain English: the Board considered the petition on the merits and refused to open trial.
• Judge panel: Not identified in the sources I retrieved. The docket confirms a panel-issued Paper 8 (Institution Decision: Deny), but the APJ names were not surfaced. Unknown — verify on PTAB E2E.
• Petition grounds: Challenged claims 1–3, 6–14, 16–18, 20, 21, 23–25, 27, and 28 of the '164 patent. The ground was § 102/§ 103 over patents and printed publications — the exhibit list (Ex. 1007–1018, 1023–1025) is composed of U.S. patent publications, patents, and litigation papers (e.g., U.S.P.N. 7,843,923 (Baum), U.S.P.N. 8,131,271 (Ramer), US 2007/0157247 (Cordray), US 2008/0108308 (Ullah)), consistent with an obviousness challenge across the challenged set. I could not retrieve the petition itself, so the precise art-to-claim mapping and the split between § 102 and § 103 is unverified.
• Institution decision: Denied on 2022-11-29 (Paper 8, "Institution Decision: Deny"). The Board's reasoning, as recited in a later expert declaration: "the petitioner in that case did not establish a reasonable likelihood of prevailing on" at least one challenged claim. Paper 9 (2023-01-04) is Petitioner's Request for Refund of Post-Institution Fees, followed by a Board notice on 2023-01-06 — consistent with a clean denial and case closure.
• Final Written Decision: None. A denial of institution ends the case; no FWD issued.
• Settlement / termination: No settlement. The case terminated by denial of institution.
• Appeal: None. A denial of institution is not appealable as of right under § 314(d); there is no FWD to appeal, and no CAFC docket number exists for this proceeding in any source I found.
• Defensive value: Roku — the defendant in the parallel W.D. Tex. suit (AlmondNet v. Roku, No. 6:21-cv-00731) — tried the full statutory claim set of the '164 patent and lost at the threshold. Because no institution issued, § 315(e)(2) estoppel never attached, so Roku remains free to run the same art in district court. For you, that cuts both ways: the Roku petition tells you what the Board has already rejected as not reasonably likely — re-packaging it is the fastest way to a second denial.

IPR2022-01506 — Samsung Electronics Co., Ltd. v. Intent IQ, LLC

• Type: Inter Partes Review
• Filed: Not confirmed in the sources I retrieved. The denial issued 2023-04-13; proceedings in the IPR2022-015xx series were docketed in the fall of 2022. Treat the filing date as unverified — pull Paper 1 (Petition) and the Notice of Filing Date Accorded on PTAB E2E. I decline to state a specific filing date.
• Status: Not Instituted – Merits (Google Patents front matter, verbatim).
• Judge panel: Not identified in the sources I retrieved. Unknown — verify on PTAB E2E.
• Petition grounds: Challenged claims of the '164 patent, expressly described by later petitioners as "different grounds from those presented" in the other '164 petitions. This is a § 102/§ 103 patents-and-printed-publications challenge (only such art is available in IPR). The specific claims and references are not in any source I retrieved.
• Institution decision: Denied on 2023-04-13, "because Samsung did not establish a reasonable likelihood of prevailing on at least one challenged claim." (Recited verbatim in later petition exhibits, e.g. https://ptacts.uspto.gov/ptacts/public-informations/petitions/[1555166](/patent/1555166) and .../1555167.)
• Final Written Decision: None.
• Settlement / termination: None; terminated by denial.
• Appeal: None. Same § 314(d) reasoning as above.
• Defensive value: Samsung — the defendant in AlmondNet v. Samsung, No. 6:21-cv-00891 (W.D. Tex.), where the asserted '164 claims were 1, 3–4, 16, and 22 per Samsung/AdGear's amended invalidity contentions — filed its own, different-ground petition and was likewise turned away. Two independent, differently-armed, well-funded industry defendants failed the reasonable-likelihood threshold. A defendant today should assume that any obviousness theory visible from the '164 patent's face has already been tried and screened out; the winning path runs through new art, a new motivation-to-combine theory, or a § 112/§ 101 attack (which IPR cannot reach).

Strategic summary

Claim status — nothing is canceled, nothing is adjudicated. Roku's petition put claims 1–3, 6–14, 16–18, 20, 21, 23–25, 27, 28 at issue; Samsung challenged claims of the '164 that later petitioners describe only as overlapping-in-part and on different grounds. Both petitions died at institution, so no claim of US 8,566,164 has been canceled, confirmed, or construed in a Final Written Decision. Every claim — including independent claim 1 and the all-important asserted set (1, 3–4, 16, 22 in the Samsung case) — remains valid and completely untested at the PTAB. Any statement that claims of this patent "survived an IPR on the merits" is wrong; they survived a threshold screen, which is a materially weaker signal than a post-trial affirmance. Do not let opposing counsel conflate the two.

Estoppel landscape — wide open. Because neither petition was instituted, 35 U.S.C. § 315(e)(2) estoppel never triggered for Roku, Samsung, or their privies — estoppel attaches only "after the petitioner requests or the Director institutes an inter partes review … and that review results in a final written decision." With no institution and no FWD, there is zero IPR estoppel on this patent. Practically, every prior-art ground that is a patent or printed publication is available to a current defendant in district court, including art Roku and Samsung raised, and system/on-sale/public-use art, which IPR could never have reached under § 311(b) in the first place. There is no "ground I reasonably could have raised" trap waiting for you. Conversely, if you file an IPR and lose at institution, you too walk away with no estoppel — but also with a denial on your record that the district court can be invited to weigh.

Pattern signals. Two different large-industry defendants, each facing parallel infringement suits in W.D. Tex., ran IPRs on the same patent in the same year (2022) — Roku first, Samsung second on different grounds — and both were denied in identical "no reasonable likelihood" terms. That is the classic one-two punch of a consolidated defense group, and its failure is notable. The "filed" vs. instituted distinction also matters: the Google Patents front matter labels both "Not Instituted - Merits," meaning the Board reached the merits of the petition rather than killing them on § 315(b) time bar or § 325(d) discretion — i.e., Roku and Samsung lost the substantive reasonable-likelihood fight, not a procedural one. No defensive aggregator (e.g., Unified Patents) appears as petitioner in either proceeding, notwithstanding that Google Patents sources the docket data from the Unified Patents portal — the portal is the data vendor, not the filer. I found no Federal Circuit appeal of either proceeding, consistent with the non-appealability of institution denials. For completeness: later IPRs against related but different Intent IQ patents (e.g., Viant's IPR2024-00421, LiveIntent's IPR2025-01317) cite these two '164 denials as background but do not attack the '164 itself.

Recommended next steps

1. Verify the two proceedings on PTAB E2E before relying on any of the above. Pull the Paper 1 (Petition), Paper 7/8 (POPR and Institution Decision) for IPR2022-00953, and the equivalent papers plus the Notice of Filing Date for IPR2022-01506, to confirm (a) Samsung's exact filing date, (b) both judge panels, and (c) the precise claims and references Samsung challenged. Case pages: https://ptacts.uspto.gov/ptacts/ (search by proceeding number). My panel and ground data for IPR2022-01506 is incomplete, and I have flagged rather than filled those gaps.

2. Get the two denials themselves. The IPR2022-00953 Institution Decision (Paper 8, 2022-11-29) is the single most valuable defensive document on this patent — it tells you exactly which claim limitations the Board thought Roku's art failed to reach, which is the roadmap of what not to file. Secondary corroboration of this docket is at https://services.patexia.com/lawsuits/Roku-Inc-v-Intent-IQ-LLC-id-[163930](/patent/163930)/documents/.

3. Do not build a defense on "the patent was invalidated." It was not. Nothing here gives you a Klein-style "the claims you're being sued on are canceled" argument. The disposition in both proceedings is institution denied / patent intact: https://patents.google.com/patent/US8566164/en (front matter, "Family has litigation").

4. Exploit the estoppel vacuum. Because no IPR was instituted, all § 102/§ 103 art on this patent is available to you in district court, including the Roku and Samsung references (no § 315(e)(2) bar on them or their privies). Focus on the claim limitations flagged in the Roku denial, on non-printed-publication art (§ 311(b) carve-out: on-sale bar, public use, system art), and on § 101/§ 112 theories, which the Board cannot entertain in an IPR and which therefore remain fully open.

5. If your demand-letter analysis turns on this patent's strength, weigh the remaining litigation risk. The '164 patent was asserted in W.D. Tex. (AlmondNet/Intent IQ v. Samsung et al., No. 6:21-cv-00891; v. Roku, No. 6:21-cv-00731) and later in D. Del. (No. 1:23-cv-00220, FreeWheel; No. 1:25-cv-00406, Mediaocean) and N.D. Cal. (No. 4:22-cv-07515). Two IPR denials + a live multi-front assertion campaign = a patent owner that has had its patent tested at the threshold and is still asserting it. Budget accordingly.

Confidence and gaps: High confidence on the existence, petitioners, challenged-claim lists, and denial dates/dispositions of both proceedings (multiple independent corroborations). Low confidence / unverified: Samsung's exact filing date, both judge panels, Samsung's specific references, and the precise § 102-vs.-§ 103 breakdown in either petition. I have not fabricated any of these; where I don't know, I have said so.

I'll reconstruct this from the authoritative patent text plus live searches for the assignment record, litigation, and asserter-directory cross-references.

Let me pull the specific assignment reel/frame data and the litigation/asserter-directory cross-references.

Now let me hunt for the specific reel/frame numbers and the recording correspondent.

Assignment & Ownership Reconstruction — US 8,566,164 B2

Patent: US 8,566,164 B2 — Targeted online advertisements based on viewing or interacting with television advertisements
App. No.: 11/968,117 · Filed: 2007-12-31 · Granted: 2013-10-22
Family: PCT/US2008/081037 → WO 2009/088554 A2; national counterparts EP 2238755, CA 2710774, JP 5707588, CN 101911689, KR 101562685, AU 2008347029, HK 1150918, IL 206660
Adjusted expiration (recorded): 2032-01-28 · Status: Active

Data-availability caveat, stated up front: the primary sources reachable in this session (Google Patents Legal Events, the Unified Patents patent record, PTAB filings, and district-court dockets) disclose conveyance type, assignor, assignee, and event dates for this patent, but do not expose the USPTO reel/frame numbers or the correspondent of record. I was unable to reach assignmentcenter.uspto.gov directly. I have therefore left every reel/frame field explicitly marked UNVERIFIED rather than reconstructing one from pattern. Each entry below is grounded in a cited, retrieved record; nothing is inferred from entity naming.

Inventors

Unusual-pattern check — not present. The frequent red flag ("all inventors depart the original assignee within 12 months of filing, preceding a fire-sale") does not apply here. Both inventors executed an assignment to AlmondNet, Inc. within two days of the 2007-12-31 filing, and the lead inventor remains the controlling principal of the patent-owning entity ~19 years later. This is the inverse pattern: founder-inventor retains control through every link in the chain.

Original assignee

Entity named on the issued patent: Intent IQ, LLC (Google Patents Current Assignee and Original Assignee both list Intent IQ, LLC; the W.D. Tex. § 290 notice of suit identifies the holder of the '164 patent as INTENT IQ, LLC, distinguishing it from the sibling patents held by ALMONDNET, INC.).

Note the nuance: Intent IQ, LLC was not the first assignee of record. The first recorded assignment (2008-01-02) runs to AlmondNet, Inc. Intent IQ acquired the application in the 2011-09-22 intra-group transfer described below and is the assignee named at issuance.

• Primary line of business: AlmondNet Group is an ad-tech / targeted-advertising R&D and IP-licensing house. Intent IQ, LLC is its identity-resolution / cross-device device-graph operating arm (cookieless ID clustering, first-party ID resolution, CTV/MAIDless activation), headquartered in Long Island City / New York City with R&D in Herzliya, Israel. AlmondNet, Inc. describes itself as "focused on R&D and the licensing of its extensive portfolio."
• Does it ship a product embodying the claims? Partially — yes. Intent IQ markets commercial identity-resolution products and states its portfolio includes "over 170 granted patents" (GlobeNewswire, 2026-04-07). However, its own litigation counsel, Reza Mirzaie, told AdExchanger that whether the company earns more as an "ad tech and data vendor" or as a "patent litigant" "depends on the year." This is the central ambiguity for the verdict below.
• Current status: Operating. Parent AlmondNet Group established 1998; survived the dot-com bust; reported near-insolvency in 2013 resolved by an asset-light license to Google (not a sale) per IAM's 2024 interview with Shkedi. No bankruptcy, no dissolution, no acquisition. Resolution/license agreement with Amazon announced 2026-04-07.

Assignment timeline

Three conveyance records are visible for this patent. Execution dates and reel/frame numbers are unavailable from the sources reachable here.

• Executed on or before 2008-01-02 / recorded 2008-01-02 — Reel/Frame UNVERIFIED

  • Conveyance: Assignment — "ASSIGNMENT OF ASSIGNORS INTEREST"
  • Assignor: Shkedi, Roy, Mr.; Shlomo, Ronen, Mr. (individual inventors)
  • Assignee: AlmondNet, Inc.
  • Correspondent: not retrievable in this session — no correspondent name recovered. No recurrence can be asserted.
  • Context: Inventor-to-company assignment of the invention at the moment of filing (application filed 2007-12-31, recorded 2008-01-02) — standard founder/employee obligation-to-assign, not a fire-sale. Source: Google Patents Legal Events, US8566164.

• Executed 2011-09-22 / recorded 2011-09-22 — Reel/Frame UNVERIFIED

  • Conveyance: Assignment of assignors' interest
  • Assignor: AlmondNet, Inc.
  • Assignee: Intent IQ, LLC
  • Correspondent: not retrievable in this session. Adjacent (not the same thing, flagged for follow-up): Intent IQ's post-issuance prosecution correspondence in this family is signed Louis J. Hoffman, 7689 East Paradise Lane, Suite 2, Scottsdale, AZ 85260, under customer number "Louis J. Hoffman/Donald Hertz" (Certificate of Correction filed 2020-03-30 in the '198 application). That is a prosecution-of-record correspondent, not a confirmed assignment-record correspondent — do not conflate them.
  • Context: Internal reorg / portfolio consolidation. The '164 application moved from the operating parent into the affiliate LLC that appears as assignee at issuance.

• Executed 2011-09-22 / recorded 2011-09-22 — Reel/Frame UNVERIFIED

  • Conveyance: License
  • Assignor: Intent IQ, LLC
  • Assignee: AlmondNet, Inc.
  • Correspondent: not retrievable in this session.
  • Context: Immediate license-back on the same day as the assignment above. Intent IQ takes record title while AlmondNet, Inc. retains a license. This dual-entity structure is corroborated for the sibling '398 patent, where Intent IQ's PTAB mandatory notice states verbatim: "The Patent Owner and real party-in-interest … is Intent IQ, LLC. AlmondNet, Inc. is a licensee of the '398 patent." (IPR2022-01420, Intent IQ's Updated Mandatory Notices, 2023-01-09.)

No further recorded assignments appear after 2011-09-22 — notably, no assignment accompanies any of the 15+ infringement suits, which means title was settled ~10 years before assertion began.

Timeline diagram

timeline
    title Ownership of US 8566164
    2007 : Application filed by Shkedi and Shlomo
    2008 : Inventors assign rights to AlmondNet Inc
    2011 : AlmondNet assigns to Intent IQ LLC
         : Intent IQ grants license back to AlmondNet
    2013 : Patent issued to Intent IQ LLC
    2019 : First AlmondNet portfolio suit on sibling patents
    2021 : First suits naming this patent filed
    2022 : Roku and Samsung IPR petitions filed
    2023 : Institution denied for both IPR petitions
    2026 : Amazon dispute resolved by license

NPE / troll-pattern signals

1. Shell-entity transfer — NOT PRESENT (as to a third-party shell); one weak structural tell.
The only transfer of '164 moved it within a single corporate family: AlmondNet, Inc. → Intent IQ, LLC on 2011-09-22, with both entities controlled by the named inventor Roy Shkedi (GlobeNewswire: "its subsidiaries Intent IQ and Datonics"; Shkedi is "CEO of Intent IQ and its parent company, AlmondNet Group"). The task's shell-entity tells are largely absent: Intent IQ is not at a registered-agent service address (37-18 Northern Boulevard, Suite 404, Long Island City, NY), it has employees and an R&D center, and it sells commercial products. The weak tell that survives: the same-day license-back (2011-09-22) plus a deliberate split of sibling-patent ownership — the W.D. Tex. § 290 filing lists '307, '822, '582, and '139 as held by ALMONDNET, INC. while '164 alone is held by INTENT IQ, LLC, and foreign counterparts split the same way (CA 2710774 names AlmondNet; EP 3448045 names Intent IQ). That is assertion-structuring, not shell concealment.

2. Known asserter in the chain — PRESENT (strong).
Current and record owner Intent IQ, LLC is a high-frequency plaintiff, surfaced by the Unified Patents directory. The Unified Patents patent page for US-8566164 lists, under Related Case IPR2022-01506 (Samsung, filed 2022-09-07; institution denied 2023-04-13): Industry: High-Tech; P.O. Entity Type: NPE (Small Company). Intent IQ / AlmondNet have sued Amazon, AWS, Microsoft (LinkedIn and Xandr), Meta, Roku, Samsung, Lotame, Viant, FreeWheel (Comcast), LiveIntent, Samba TV, AppLovin, Mediaocean, Flashtalking, Innovid, Beeswax, Oracle, and Yahoo/Oath. Neither entity appears on the classic enumerated lists (Acacia, Marathon, Intellectual Ventures, Vringo, Wi-LAN/Conversant, Pendrell, Round Rock, MPHJ, etc.) — the match is via the Unified Patents classification and the Stanford NPE Litigation Database, which is exactly the category the rubric asks for.

3. Repeat correspondent across the chain — UNCLEAR (could not be tested).
No assignment-record correspondent was recoverable for any of the three conveyances, so no recurrence finding is possible — and I will not manufacture one. The only family-wide correspondent data point I can substantiate is Louis J. Hoffman (Scottsdale, AZ; customer no. "Louis J. Hoffman / Donald Hertz") on Intent IQ's post-issuance prosecution papers in the sibling '198 patent (Certificate of Correction, 2020-03-30). That is a single appearance on a different filing stream; one appearance is not a recurrence finding under the rubric. Separately, litigation counsel of record is Russ, August & Kabat LLP (Reza Mirzaie, Benjamin T. Wang, James A. Milkey, Amy E. Hayden, James S. Tsuei, Daniel B. Kolko, Jason M. Wietholter — Los Angeles) with Farnan LLP as Delaware local counsel — but litigation counsel is not assignment correspondent and should not be counted.

4. Cascading transfers — NOT PRESENT.
Only three recorded events, two of them dated the same day (2011-09-22); the third is the 2008-01-02 inventor assignment. There is no chain of consecutive LLC-to-LLC transfers inside 24 months, no sharing of assignee address across links, and no established common correspondent. The ownership record is short and static, not cascading.

5. Pre-litigation transfer — NOT PRESENT.
Last ownership change: 2011-09-22. First suit naming the '164 patent: 2021-07-15 (AlmondNet, Inc. v. Roku, Inc., No. 6:21-cv-00731, W.D. Tex.; and Roku's mirror DJ action, 1:21-cv-01035, D. Del., same day). Gap ≈ 9 years 10 months. The chain was not arranged to set venue or clean standing immediately pre-suit. (That said, the split between AlmondNet-held and Intent-IQ-held patents did give the group a two-plaintiff structure — AlmondNet, Inc. and Intent IQ, LLC appear jointly as plaintiffs in the later filings, e.g. 4:22-cv-07515 and 1:25-cv-00406.)

6. Bankruptcy fire-sale — NOT PRESENT.
No Chapter 7/11 proceeding appears. The closest analogue is distress licensing, not distress selling: Shkedi testified he faced "near financial ruin" in 2013 and took a take-it-or-leave-it license deal with Google — an inbound license, not an outbound patent sale, and title never left the group (IAM, 2024-08-15).

7. Privateering — NOT PRESENT / inapplicable.
There is no operating-company → NPE handoff. AlmondNet/Intent IQ are the original inventor-side owner asserting its own portfolio; the '164 has been owned by the same corporate family since the day it was filed.

8. Defensive aggregator (anti-NPE) — NOT PRESENT.
The chain terminates at Intent IQ, LLC, an asserting entity. No RPX, AST, LOT Network, Unified Patents, or OIN ownership appears at any point. Unified Patents appears here on the adversarial side (its data underlies the Roku/Samsung/Delaware dockets and its portal classifies the owner as NPE), not as an owner.

Validity posture (value context, not an NPE signal): the '164 has survived repeated validity attack — IPR2022-00953 (Roku, filed 2022-05-24, institution denied 2022-11-29) and IPR2022-01506 (Samsung, filed 2022-09-07, institution denied 2023-04-13) — and the '164 was among the patents carried in the Roku and Samsung litigations that ended in licenses.

Verdict

NPE — moderate confidence.

Title to the '164 sits with a non-practicing affiliate — Intent IQ, LLC — which took the patent from AlmondNet, Inc. on 2011-09-22 and immediately granted AlmondNet a license back the same day, a two-entity structure under one principal, the named inventor Roy Shkedi, with the sibling patents deliberately split between the two LLCs (per the W.D. Tex. § 290 notice: '307/'822/'582/'139 held by AlmondNet, '164 held by Intent IQ). Unified Patents classifies this patent's owner as "NPE (Small Company)" on IPR2022-01506, and the family has been asserted against 15+ ad-tech defendants. I stop at moderate rather than high because two of the rubric's strong signals are affirmatively absent on the record: there is no third-party shell transfer (assignor and assignee are affiliates of the same founder), no pre-litigation transfer (2011 vs. a 2021 first suit), and Intent IQ genuinely ships identity-resolution products, which its own counsel says sometimes out-earn its litigation revenue.

Verify at:

• USPTO Assignment Center — https://assignmentcenter.uspto.gov/ (search patent 8566164 for reel/frame and correspondent of record)
• USPTO Patent Assignment Search — https://assignment.uspto.gov/patent/index.html
• Google Patents Legal Events — https://patents.google.com/patent/US8566164/en
• Unified Patents record (NPE classification, IPR2022-01506) — https://portal.unifiedpatents.com/patents/patent/8566164
• Stanford NPE Litigation Database — https://npe.law.stanford.edu/

Open items I could not close: (a) all three reel/frame numbers and execution dates; (b) the correspondent of record on the 2008 and 2011 recordings — the single highest-value remaining datum for the "repeat correspondent" test; (c) whether the 2011 license-back is exclusive or non-exclusive and its field/term, which affects standing to assert. Item (b) is a one-query lookup at Assignment Center and should be run before any of the three-signal conclusions are finalized.

I'll research the prior-art citations for US 8,566,164. Let me search authoritative sources.

Let me pull the actual front-page citations and cited-by data from patent databases.

I found the front-page reference list. Let me retrieve the complete citation sets from FreePatentsOnline and Google Patents.

Prior Art for US 8,566,164 B2 — Citation Analysis

1. Patent identification (confirmed literal match)

The number searched, 8,566,164, resolves to exactly one document:

This is distinct from similarly numbered patents (e.g., US 8,566,163; US 8,566,165) — no other number was substituted.

2. Important methodology and confidence caveats (read first)

Before the reference-by-reference analysis, three limitations must be stated plainly, because they materially affect how much weight to give the list below:

1. I could not retrieve the complete, authoritative front-page "(56) References Cited" list. I reached the patent via Google Patents, FreePatentsOnline (FPO), Justia, Unified Patents and Docket Alarm (all of which mirror USPTO data) — I did not have direct access to USPTO PatentCenter / Patent Full‑Text in this session. The FPO rendering returned the top (most‑recent‑first) ~24 U.S. patent references before truncating. A patent of this vintage typically cites a much longer list, and the older references (the ones most likely to be true §102 art) sit at the bottom of that list and were not retrieved. Treat the list in §4 as partial.

2. A bibliographic discrepancy to flag, not auto-correct. The Google Patents front matter gives the filing/priority date as 2007‑12‑31. One Unified Patents listing (for the later continuation US‑12063395‑B2) showed US‑8566164‑B2 with a date of 2007‑12‑30. I am flagging the one-day discrepancy rather than silently harmonizing it. The controlling date for the §102 analysis is the application filing date, 2007‑12‑31.

3. Critical-date filter (pre‑AIA, because the application was filed in 2007). A reference can only anticipate (or otherwise qualify as §102 art) if it predates 2007‑12‑31, or — for U.S. patents/published applications under pre‑AIA §102(e) — if its effective U.S. filing date predates 2007‑12‑31. This matters enormously below, because many of the retrieved references issued in 2010–2012, i.e., after the '164 filing date. They can be §102(e) art only if their underlying filings predate 2007‑12‑31; otherwise they are not prior art at all. I could not verify each reference's underlying filing date in this session.

4. "Potentially anticipates" is a screen, not a legal conclusion. Because the '164 claims issued over the examiner's cited art, none of these references actually defeated the claims. The analysis below identifies which references are closest in subject matter and which claim elements they appear to reach — a §103-style relevance screen — not a finding of anticipation.

3. The claims to be analyzed

Per the previously generated summary, only claim 1 (method) was authoritatively verified here; the full independent-claim set was not retrieved, and the abstract additionally recites a "tracking and recording online activity" alternative that may be claimed separately. Claim 1 requires, in combination:

• (a) responsive to online activity, automatically causing a selected online ad to be directed to the online device (identified by a first online access identifier);
• (b) selection based on a user profile that references/includes a first set‑top‑box identifier that is electronically associated with the online access identifier;
• (c) the profile info derived from data about user behavior toward a television ad delivered to the STB;
• (d) the STB identifier and online access identifier associated without using personally identifiable information (PII).

Anticipation requires one reference disclosing (a)+(b)+(c)+(d).

4. Retrieved citations (partial U.S. patent references on the '164 face)

Publication/issue dates are as printed on the face (FPO rendering). Descriptions are the patent titles/citations as listed, with brief technical characterization.

Additional art surfaced via Unified Patents' "Patent Art" table (173 items associated with the family), also relevant:

5. Anticipation assessment (claim‑by‑claim screen)

5.1 The single most relevant references — and why they still do not anticipate claim 1

US 7,861,260 B2 / US 2009/0049468 A1 (Shkedi), filed 2007‑04‑17 — the incorporated parent.
This is the only retrieved reference with a verified pre‑critical‑date filing (2007‑04‑17) and directly overlapping subject matter (linking TV‑ad delivery to a household's online identity). However, it discloses the reverse data flow: online behavior → targeted television advertisement. Claim 1 of the '164 patent requires the opposite — TV‑ad behavior → targeted online advertisement — plus the specific STB‑ID↔online‑access‑ID association of elements (b)/(d). On its face it therefore does not anticipate claim 1; it is at most combination/background art. (Note also the same‑inventor question: pre‑AIA §102(e)/(a) art must be "by another"; the '164 inventive entity is Shkedi + Shlomo, so this reference may still legally qualify as prior art, but this needs the full inventorship confirmed.)

The remainder of the Shkedi‑named references (items 2–7, 24). These are members of the same AlmondNet/Intent IQ family and, like US 7,861,260, are directed to online‑behavior‑driven TV ad targeting or to profile/privacy mechanics. For the same directional reason, none appears to anticipate claim 1.

5.2 References reaching individual claim elements but not the combination

• Element (a) – automatic targeted ad on online activity: addressed generally by US 8,132,202 (Swix), US 8,086,491 (Matz), US 2011/0258049 (Ramer/Jumptap), US 2012/0023522 (Anderson) — targeted‑content/impression systems. None retrieved shows the TV‑behavior trigger.
• Elements (b)/(c) – TV‑ad‑delivery/STB targeting: US 7,913,275 (Hendricks) ("switching targeted advertisements at a set top terminal"), US 8,024,765 (Ramanathan), US 7,900,229 (Dureau), US 2009/0172723 (Shkedi). These reach TV‑side targeting and STB identification, but not the subsequent targeted online ad delivered to an online interface device keyed to the STB identifier — i.e., they miss (a)/(b).
• Element (d) – non‑PII association: US 7,890,609 (Shkedi), US 7,966,647 (Igoe), US 7,937,383 (Hintze) — profile‑sharing/anonymization references relevant to the "without using PII" limitation.
• Cross‑device identity linking: US 7,971,260 (Spalink), US 7,856,372/7,856,373 (Ullah), US 8,015,286 (Jenkins).

5.3 Bottom line of the screen

On the retrieved (partial) list, no single reference discloses all of claim 1(a)–(d), and the subject‑matter‑closest references (the Shkedi family) disclose the opposite data flow. The realistic §102 exposure, if any, would be in the older references I could not retrieve (bottom of a newest‑first list) or in the art cited in the unsuccessful IPRs (below). The retrieved references map more naturally to a §103 combination theory than to §102 anticipation — and, as noted, the examiner allowed the claims over whatever he/she cited.

6. IPR prior art (the strongest identified challenge set)

The '164 patent survived two petitions that were denied institution on the merits — meaning the PTAB did not find a reasonable likelihood that the cited art anticipated/rendered obvious the claims:

• IPR2022‑00953 (filed 2022‑05‑24) — U.S. 8,566,164 — Not instituted.
• IPR2022‑01506 (filed 2022‑09‑07) — U.S. 8,566,164 — Not instituted.

A companion filing, IPR2022‑01505, appears in Docket Alarm as Samsung Electronics Co., Ltd. v. Intent IQ, LLC (petition dated 2022‑09‑07, i.e., the same day as ‑01506), with petitioner exhibits including an "EAST Search History (Prior Art)" and a PTO‑1449 IDS exhibit. I could not retrieve the specific prior‑art references or the grounds recited in the '164‑specific petitions; those petitions' exhibits (Petitioner Exhibit 1002 "EAST Search History" and the IDS exhibit 1064) are where the petitioners' asserted art would be listed, and retrieving them is the highest‑value next step. I also note the previously generated section flagged that the petitioner identities for ‑00953/‑01506 were not conclusively confirmed; the Samsung link above is suggestive for the ‑01505/‑01506 pair but is not confirmed for the '164‑specific petition.

7. Notes on non‑patent literature

I did not retrieve the '164 patent's own NPL/other‑references list. Caution: some search results returned NPL that belongs to different patents — e.g., a Gracenote/Invidi press‑cluster (Constine, TechCrunch 2012‑12‑26; Friedman, MediaDailyNews 2013‑01‑16; Roettgers, GigaOM 2012‑12‑14; Mandese, RTBlog 2013‑01‑17) and a Blinkx TV White Paper 1.0 (Jan 2005) appeared on the Unified Patents page for US‑12063395‑B2, not '164. Likewise, supplementary European search reports for EP3618784A4 and EP3308559A4 that surfaced in search contained medical‑device/ultrasound art (Ross, Raney, Nguyen, Dubois, Anis, Terwilliger; Bansal, Matsuoka) unrelated to the '164 subject matter — these are mismatched hits and should not be attributed to the '164 patent. I flag them so they are not mistaken for '164 citations.

8. Explicit uncertainty flags

• Incomplete citation list. The §4 list is the newest‑first top portion of the '164 "U.S. Patent References"; older references — the best §102 candidates — were not retrieved. Recommendation: pull the complete "(56) References Cited" from USPTO PatentCenter (application 11/968,117, file wrapper) or the full Google Patents "Patent Citations" table.
• Unverified filing dates. For every reference that issued after 2007‑12‑31, §102(e) status depends on a pre‑2007‑12‑31 U.S. filing date, which I did not verify for items 2–24.
• Inventorship/"by another." Whether the Shkedi‑named matches qualify as pre‑AIA §102 art turns on whether their inventive entity differs from the '164 entity (Shkedi + Shlomo) — not confirmed.
• Claim set. Only claim 1 was verified previously; there may be additional independent claims (e.g., a "tracking and recording activity" claim per the abstract) that this screen does not fully cover.
• IPR grounds unretrieved. The specific prior art in IPR2022‑00953/‑01506 (and the apparent Samsung companion) was not obtained.

Primary sources consulted: Google Patents US8566164B2 (https://patents.google.com/patent/US8566164/en); FreePatentsOnline (https://www.freepatentsonline.com/8566164.html); Unified Patents (https://portal.unifiedpatents.com/patents/patent/8566164); Docket Alarm IPR2022‑01505 record (Samsung v. Intent IQ) and W.D. Tex./D. Del. dockets; PTAB Portal pages for IPR2022‑00953 and IPR2022‑01506.

I'll research the prior art cited on the '164 patent and the closest known references in the cross-media targeting space before building the § 103 analysis.

Let me try to retrieve the actual prior-art references cited on the face of the '164 patent and in the two IPR petitions directed to it.

Good — I've located the actual prior art from the Roku IPR against this patent. Let me identify each reference and find the second (Samsung) petition's grounds.

Obviousness Analysis — U.S. Patent No. 8,566,164 B2

Under 35 U.S.C. § 103 (pre‑AIA § 103(a))

Tool budget note: I exhausted my search allotment mid‑investigation. Everything below is grounded in the sources retrieved in this session plus the patent text supplied in the prompt. Where I could not verify something, I say so rather than fill the gap. The Google Patents page as supplied to me did not contain a machine‑readable "References Cited"/"Prior Art" list; the prior art below therefore comes from (i) the prior art actually asserted against the '164 patent in the two IPRs on it, retrieved from PTAB petition documents, and (ii) the '164 specification's own statements and incorporated references. That is flagged repeatedly.

0. Corrections/updates to the previously generated sections

Source for the three rows: PTAB petition document at ptacts.uspto.gov/ptacts/public-informations/petitions/1555167 (declaration ¶¶ 62, 74), and …/petitions/1555168 ("'164 Patent, which is a continuation‑in‑part of the '260 Patent"). Note also a source artifact on Unified Patents' page: it recites "IPR2022‑00953, May 24, 2022, Re: Pat. No. 5,866,164 (institution denied)" — I interpret that literally as written but read it as a typo for 8,566,164, because the immediately adjacent entry says "IPR2022‑01506 … No. 8,566,164 (institution denied)" and the surrounding family data concerns the '164. https://portal.unifiedpatents.com/patents/patent/US-12063421-B1

1. Claim scope actually analyzed (and its limits)

Verified: independent claim 1 (method) — text as quoted in W.D. Tex./D. Del. pleadings, with clause (d) truncated after "…without using personally identifiable information pertaining to a user of the set‑top box that corresponds to the first set‑top box identifier" (the recovered sibling claim 18 language in the Roku petition shows the full clause structure; see ptacts.uspto.gov/…/petitions/1549592).

Inferred, not verified: the challenged claim set in Roku's IPR — claims 1‑3, 6‑14, 16‑18, 20, 21, 23‑25, 27, 28 (per petition doc 1555167, ¶74). This implies at least 28 claims and multiple independent claims, and the ground structure (one Madhavan+Whitehead mapping covering 1‑3, 6, 12, 14 and 16‑18 and 21, 23‑25, 28) is consistent with parallel method / system / article ("computer‑readable medium") independents. I did not open the '164 patent's official claim listing, so I do not assert the number or identity of the independent claims. All element mapping below is to claim 1 only.

Legal frame: Pre‑AIA § 103(a) applies (filed 2007‑12‑31, no post‑AIA priority). Graham v. John Deere factors; KSR Int'l v. Teleflex (2007) controls the motivation analysis. Note the CIP wrinkle: if any '164 claim is supported by the 2007‑04‑17 parent (Ser. No. 11/736,544), its effective date could be earlier than 2007‑12‑31, which would push the critical date for prior‑art qualification back to 2007‑04‑17. No claim‑by‑claim priority analysis exists in my sources; this materially affects the § 102(e) analysis in § 2.4 below.

2. The prior art

2.1 References the IPRs actually applied against the '164 patent

Per the declaration summary of Roku's petition (IPR2022‑00953), all grounds were § 103(a) only — no § 102 ground:

Source: ptacts.uspto.gov/ptacts/public-informations/petitions/1555167, ¶74. Caveat: this paragraph sits inside a declaration that also describes a parallel Roku petition against the '198 patent (US 10,321,198) that used the same reference families (Madhavan + Whitehead + Vallier/Varghese/Beyda — see …/petitions/1549592). References therefore risk conflation between the '164 and '198 petitions in that document. I could not open the IPR2022‑00953 petition itself or the Nov. 29, 2022 denial, so I treat the ground‑to‑claim mapping as reported, not confirmed. The same document also contains an apparent drafting artifact ("Samsung did not propose any constructions" appearing inside the description of the Roku petition).

2.2 What each reference teaches (as described in the petitions)

• Madhavan (primary reference; ROKU‑1008). Described as teaching content management across multiple devices in a local area network, a user profile holding "demographic data, geographic data, and behavioral data," with the user profile linked to events/activities in tables 230 and categories in table 220, and with targeted ads selected using usage from all devices in the LAN — "a PC, an internet‑connected set‑top box, a mobile handset" — and ads sent to a user at time t based on usage at earlier times. Madhavan expressly teaches association of device identifiers based on "IP access session" information with a common IP address and a timestamp. (I could not independently retrieve Madhavan's patent number/date in this session; the internal exhibit label is ROKU‑1008. Flagging as unverified — this matters for § 102(e) qualification.)
• Whitehead (US 2008/0281697 A1). Teaches association between a first online user interface device identifier and a set‑top box identifier based on IP access session information ("association of RI‑mytes including corresponding device identifiers based on the IP access session information association"), and targeted advertisement selection based on RI‑mytes from all devices in a common LAN (FIG. 5, data blocks 300‑360), linked to a user profile. Whitehead is the reference that supplies the claimed anchor — the non‑PII electronic association of an online access identifier to a set‑top‑box identifier.
• Vallier (US 9,495,682). "Converged marketing architecture and related research and targeting method utilizing such architecture"; filed August 31, 2006; teaches targeted advertising based on "TV viewing habits" of PVR users — "watching, skipping, rewinding, re‑watching."
• Beyda (US 9,928,522) and Varghese — cited for the dependent claims 9, 10, 13 and 11, 20, 27 respectively. I could not retrieve their subject matter; I will not speculate. Descriptive names/dates outside what is quoted should be treated as unverified.

2.3 The '164 specification's own admissions (usable as evidence of the art)

The specification supplied in the prompt contains inventor‑level admissions that are directly probative on the motivation‑to‑combine question:

• Association via a common IP address was a known, workable mechanism. "[T]he use of a common IP address to associate online access and set‑top box IP addresses is disclosed in application Ser. No. 11/736,544 … In some implementations, the association is such that the common IP address (IP_modem) can, at least initially, comprise the electronic association that enables targeted online advertising or tracking/recording based on observed user behavior related to a television advertisement." (Spec. at ≈ col. 19:52‑56 — the same passage the petitioner quoted as ROKU‑1001, 19:52‑56.)
• Association could be made "with or without using PII." The spec recites establishing the online‑IP ↔ STB‑ID association by "use of PII or non‑PII as needed, desired, or permitted; use of a common identifier or user pseudonym …; use of a common online access IP address used by both modem 32 and STB 36; cooperation between the ISP, TVP, and/or CAS … with or without using PII."
• Targeted online advertising and behavioral targeting keyed to cookies/IP were admitted to be conventional, and the spec names DoubleClick's DART for Publishers / DART for Advertisers as existing central ad servers. The spec also incorporates Ser. No. 11/736,544 (US 7,861,260) — which discloses the full architecture, with the only difference being the direction of targeting (online behavior → TV ad).

These admissions shrink the inventive contribution to (a) putting existing behavioral‑targeting machinery on the online side, (b) using a TV‑ad‑derived profile as the trigger, and (c) doing the linkage "without PII" — which the spec itself says was known.

2.4 Critical date problems that must be litigated

Two further doctrinal points a litigator must run down: (i) whether the '164 (a CIP naming Shkedi and Shlomo) is "by another" relative to Shkedi‑only references in its own family, and (ii) whether pre‑AIA § 103(c) / the CREATE Act disqualifies any commonly owned § 102(e) art. Neither is resolvable on my sources.

3. Ground 1 — Madhavan + Whitehead vs. claim 1

Mapping (limitations quoted from the verified claim‑1 text; reference support as described in the petition record):

Motivation to combine (KSR rationales (A), (C), (D), (F)):

1. Same field, same problem. Madhavan and Whitehead both address delivering targeted advertisements to multiple devices in a household/LAN. Whitehead itself supplies the express teaching/suggestion/motivation — it teaches associating device identifiers via an IP access session and selecting ads using activity from devices including the STB. Combining a content‑management/profile engine (Madhavan) with a device‑association/ad‑selection layer (Whitehead) is the "simple substitution of one known element for another" and "use of a known technique to improve similar devices in the same way," with predictable results (more complete profiles → more relevant ads).
2. Market forces. By 2007 the incentive to monetize TV‑viewing signal against higher‑margin online ad inventory was a well‑recognized industry objective; the '164 specification itself frames the goal as "increas[ing] the effectiveness of combined online and television advertising over that of either advertising medium alone."
3. No change in principle of operation. Both references key off the same IP‑session/common‑IP‑address infrastructure that the '164 specification itself identifies as "the electronic association that enables targeted online advertising."
4. KSR "obvious to try" with a finite, predictable set. Given a household's TV‑behavior profile and a known non‑PII IP/cookie linkage, the two candidate actions are: target the TV from online behavior, or target online from TV behavior. The '164 family claims the second while incorporating by reference the parent ('260) that discloses the first — strong intrinsic evidence that the reversal was a design choice, not an invention.

4. Dependent‑claim combinations

• Ground 2 (claims 7‑8): + Vallier (US 9,495,682). Motivation: Vallier monitors TV viewing habits (watch/skip/rewind/re‑watch) — supplying the ad‑level TV behavior data the profile needs, and knowledge of ad‑skipping in particular. Adding this to Madhavan+Whitehead yields the predictable benefit of "more effective targeted advertisements" (Vallier's own stated purpose), a classic "known technique to improve similar items in the same way with predictable results" rationale (KSR (C)). Vallier is also the reference whose 2006‑08‑31 filing date most cleanly supports § 102(e) status.
• Ground 3 (claims 9, 10, 13): + Varghese. Rationale asserted only as "same reasons" as the base mapping; I cannot state what Varghese teaches.
• Ground 4 (claims 11, 20, 27): + Beyda (US 9,928,522). Same limitation; subject matter unverified in this session.

Because I have not verified the dependent‑claim text, I cannot say which limitations Varghese/Beyda actually meet, and I expressly do not endorse those grounds beyond reporting them.

5. Additional § 103 theories available but not run in the IPRs

1. '260 patent (US 7,861,260) as the primary reference, with a direction‑reversal rationale. The '164 is a CIP of the '260, which teaches the entire architecture for associating an online access IP address to an STB identifier without PII and targeting the TV from online behavior. Reversing the targeting direction — using the STB‑side profile to select an online ad delivered to the already‑determined online access identifier — is the substitution of a known element (the same association + same profile) to obtain a predictable result. Add Whitehead for the express LAN/IP‑session association and ad selection at the online device. Hurdles: § 102(e)/"by another" analysis given the overlapping inventorship and the CIP relationship; whether '260 is disqualified under pre‑AIA § 103(c)/CREATE Act.
2. Specification‑as‑admitted‑prior‑art. The '164's own statements (common‑IP association "compris[ing] the electronic association that enables targeted online advertising"; association "with or without using PII"; conventional cookie/IP behavioral targeting; DART as an existing CAS) can be used to supply the "known technique" element and the motivation, minimizing the reference combination to Madhavan/Whitehead + admissions.
3. Non‑patent literature of record in the family. Archived DoubleClick web pages (e.g., "http://www.doubleclick.com (Aug. 10, 2005)" and "http://www.doubleclick.net (Aug. 15, 2000)") appear as prior art of record in this patent family on Unified Patents' page and are candidate printed publications for the "known ad‑serving infrastructure" element. I have not reviewed their content.
4. General addressable‑TV/behavioral‑targeting background art (e.g., Eldering US 6,216,129, Hendricks US 6,160,989, Merriman US 5,948,061, which appear on the face of the '260 patent) can support the level‑of‑skill/motivation narrative. I did not verify their content in this session and do not map them to any limitation.

6. What the record actually decided — and what it did not

These are institution denials, not merits holdings. There is no final written decision on the '164 patent, and no district‑court invalidity judgment identified in the earlier litigation section (Roku's § 101 motion attacked eligibility, not § 103; the Delaware consolidated case closed by dismissal in June 2025). So:

• The denial cuts against the strength of the as‑filed Ground 1 mapping — the Board evidently did not credit at least one limitation of at least one challenged claim. My best inference (flagged, not verified) is that the failure was on (c) ad‑level TV behavior in Ground 1 (which excluded Vallier) and/or on (d) the non‑PII negative limitation, since the petition reportedly offered a § 103(a)‑only case and the "without PII" clause is a specific negative limitation that must be affirmatively shown.
• The denial does not establish non‑obviousness and is not binding in district court (no estoppel; no FWD).

7. Counterarguments and secondary considerations

For the patent owner (non‑obviousness):

• The negative limitation. "Associated without using personally identifiable information" is a specific exclusion. Even if both references teach a device association, a combination that relies on account/subscriber linkage (PII) does not meet it. Whitehead's IP‑session association must be shown to be actually PII‑free in operation, not merely capable of being so.
• Direction of the record. The petitioner's own framing "the references applied during prosecution … did not disclose" the analogous association limitations in siblings (Ertugrul/Howcroft; Haeuser/Calamera/Bhatt) shows the Office repeatedly found household‑device association + cross‑device profile action missing from the art.
• Two‑year prosecution with an IDS listing the '260 family and ad‑serving art, moving to allowance; no § 103 rejection on this combination is reported.
• Secondary considerations: I found no evidence of commercial success, long‑felt need, failure of others, praise, or copying tied to the '164 claims in the sources retrieved. Absent that evidence, this factor is neutral on the present record.

Against the patent (obviousness):

• The specification's admissions and its incorporation of '260 remove most of the alleged novelty.
• The claim's preamble hardware is generic; the '164 claims no advance in computer/network technology (which is also the thrust of Roku's § 101 attack, D. Del. 1:21‑cv‑01035, Dkt. 25).
• The family breadth (continuations at 12/981,925 → 8,595,069; 16/435,051 → 11,095,952; 17/401,078 → 11,831,964) shows the inventors iterated on the same disclosure, consistent with a broad concept rather than a specific technical advance.

8. Bottom line and confidence

Can a prima facie § 103(a) case be made? Yes, articulable but not established. The strongest available combination is Madhavan + Whitehead (+ Vallier for the ad‑level TV‑behavior and ad‑skipping dependent claims, + Beyda/Varghese for the remaining dependents), with the motivations being: same field (LAN‑level targeted content/ad delivery), Whitehead's express teaching of the non‑PII IP‑session association of an online device identifier to an STB identifier and of ad selection from LAN‑wide activity, Vallier's TV‑viewing‑habit monitoring, the well‑recognized industry objective of cross‑media targeting, and the '164's own admissions (common‑IP association "compris[ing] the electronic association that enables targeted online advertising"; association "with or without using PII"; DART as a conventional CAS).

Why I cannot call the claims obvious:

1. Claim text is only partially verified — claim 1 provisionally, the independent/system/article claims not at all. Element mapping for claims other than 1 is therefore not performed and should not be relied on.
2. Reference identity/date gaps — I could not retrieve Madhavan's number or date, Varghese's or Beyda's disclosure, or Whitehead's filing date. Every one of Whitehead, Vallier, Beyda, Varghese (and probably Madhavan) must qualify as art only under pre‑AIA § 102(e) because they published after the 2007‑12‑31 critical date; that requires proof of pre‑critical‑date filing, "by another," and enablement — and the CIP‑to‑'260 relationship may move the critical date to 2007‑04‑17.
3. The reported Ground 1 is internally brittle — as reported it excludes Vallier, yet limitation (c) requires automatically collected data about user behavior with respect to a television advertisement, and limitation (d) requires the non‑PII association. Those two limitations are where the two institution denials most plausibly turned.
4. Both IPRs on this patent were denied institution (Roku 2022‑11‑29; Samsung 2023‑04‑13), and no adjudicator has held any '164 claim obvious.

Recommendation for follow‑up (each currently unverified): (a) pull the complete '164 claim listing from USPTO Patent Full‑Text/PatentCenter to fix the independent‑claim count; (b) retrieve the IPR2022‑00953 petition and the Nov. 29, 2022 decision, and the IPR2022‑01506 petition and the Apr. 13, 2023 decision, to get the exact ground‑to‑claim mapping and the Board's stated reason for denial; (c) confirm Madhavan's, Whitehead's, Beyda's, Varghese's, and Vallier's filing/publication dates and § 102(e) eligibility; (d) run the CIP/priority analysis for claim 1 and each independent claim.

Primary grounding sources for this section:

• https://ptacts.uspto.gov/ptacts/public-informations/petitions/1555167 (declaration ¶¶ 62, 74 — Roku/Samsung '164 IPRs, dates, grounds, denial dates)
• https://ptacts.uspto.gov/ptacts/public-informations/petitions/1555168 ('164 as CIP of the '260 patent; both '164 IPRs denied institution)
• https://ptacts.uspto.gov/ptacts/public-informations/petitions/1549592 (Madhavan + Whitehead + Vallier mapping; Whitehead device‑identifier/IP‑session association; Madhavan user tables 210 / database rationale)
• https://portal.unifiedpatents.com/patents/patent/US-12063421-B1 (IPR2022‑00953 / IPR2022‑01506 "institution denied"; archived DoubleClick pages)
• https://www.docketalarm.com/cases/Delaware_District_Court/1--21-cv-01035/…/25.pdf (the '164 as one of the "Set‑Top Box Patents")
• US 8,566,164 specification text (as supplied) — admissions at ≈ col. 19:52‑56 and the "with or without using PII" association list
• US 7,861,260 front matter ("References Cited" — Eldering, Hendricks, Merriman, et al.)