US 10715878

I'll search for the specific patent number and any appellate docket activity.

Let me search for the Federal Circuit appellate docket activity and the claim text.

Let me check the appellate outcome and full claim text.

Scope and verification note

I searched for the exact string 10715878 / 10,715,878. All results below pertain to US 10,715,878 B2 (the "’878 patent"), not to any similar number.

Two housekeeping items: (1) your prompt states the current date as April 26, 2026, while my working environment is dated September 28, 2026; docket facts below are stated with the retrieval dates shown. (2) Where the authoritative patent text you supplied conflicts with secondary sources, I follow the patent text and flag discrepancies.

1. Bibliographic data (grounded in USPTO/Google Patents bibliographic record)

Continuity (from the specification's Cross-Reference section): continuation of 15/803,631 (now 10,178,442) ← 15/179,356 (now 9,813,778) ← 14/246,525 (now 9,369,779) ← division of 13/097,885 (now 8,695,032) ← division of 12/860,666 (now 8,281,336) ← continuation of 11/736,544 (now 7,861,260). Later continuations claim priority to this application (e.g., US 11,303,973; 11,564,015; 11,589,136; 11,805,300; 11,974,025; 12,212,818 B1).

Note on the number itself (no auto-correction): the digits "10715878" correspond only to US 10,715,878 B2. Where the value "78" appears in claim numbering (claims 1–23) it is a different identifier and I do not conflate the two.

2. Abstract (verbatim summary)

"In a method for delivering targeted television advertisements based on online behavior, IP addresses indicating online access devices and IP addresses indicating television set-top boxes are electronically associated for a multitude of users. Using user profile information derived from online activity from one of the online access IP addresses, a television advertisement is selected, such as by using behavioral targeting or demographic information, and automatically directed to the set-top box indicated by the set-top IP address associated with that online access IP address. Preferably neither the user profile information nor the electronic association of online access and set-top box IP addresses includes personally identifiable information."

3. Independent claims — plain-language overview

The ’878 patent has 23 claims; claims 1, 13, and 22 are the independent claims (as recited by the Board in IPR2023-01281: "Claims 1, 13, and 22 are independent").

Claim 1 — method for identifying a device association on a common LAN and using it for a cross-device action (verbatim, Ex. 1001 at 24:25–62):

"1. A method for identifying an association among multiple electronic devices of a group of devices connected through a common local area network (LAN) and using the association for cross-device action, the method comprising:
(a) with a computer system, determining an association between device identifiers of two or more Internet-accessing electronic devices based on electronically stored association-detection information about a plurality of Internet-accessing electronic devices, wherein the stored association-detection information for each of the plurality of Internet-accessing electronic devices includes: (i) a device identifier; (ii) at least one dynamically assigned IP address assigned to a modem that connects a router of the LAN to the Internet; and (iii) at least one timestamp specifying a time when the Internet-accessing electronic device accessed the Internet via the router of the LAN;
(b) wherein determining the association in part (a) is done by using the stored association-detection information to determine that the two or more Internet-accessing electronic devices have separately accessed the Internet through a router of the LAN during a predetermined period of time via a common IP address;
(c) wherein the computer system determining the association in part (a) is outside the LAN and is connected to the LAN through the Internet; and
(d) with the computer system, using the association determined in part (a) to utilize user profile information associated with the device identifier of a first of the associated Internet-accessing electronic devices, and comprising behavioral or demographic data collected via the first of the associated Internet-accessing electronic devices, to take an automatic action with respect to a second of the associated Internet-accessing electronic devices using the device identifier of the second of the associated Internet-accessing electronic devices."

Plain language: A server outside the home LAN looks at stored records (device ID + dynamically assigned modem IP + timestamp) for many Internet devices and infers that two or more of them are on the same LAN because they separately went online through the same router at the same common IP address within a time window. The server then uses a behavioral/demographic profile gathered from device A to automatically act on device B (classically: serve a targeted TV ad to the household's set-top box). This is the "cross-device association by shared common IP address" concept.

Claim 13 (system): Independent claim drawn to a system (one or more computers programmed to carry out the association/cross-device-action steps paralleling claim 1). This characterization is consistent with the Board's treatment of claim 13 as an independent system claim; I do not have its verbatim text in the material available to me, so I am not quoting it.

Claim 22 (article / non-transitory medium): Independent claim drawn to an article comprising a non-transitory tangible medium encoding computer-readable instructions that, when applied to one or more computers, cause performance of the analogous association and cross-device-action method. Again, this is the standard Beauregard-style format used across this family; I do not have verbatim claim 22 text and will not fabricate it.

So, in one sentence: all three independent claims require (i) a common-LAN / common dynamically-assigned IP address basis for associating two Internet-accessing devices, (ii) the associating computer system being outside the LAN, and (iii) an automatic cross-device action on the second device based on profile data collected via the first device.

⚠️ Uncertainty statement: I verified claim 1's text directly from the IPR2023-01281 Final Written Decision excerpt. For claims 13 and 22 I have not personally verified verbatim language; treat the descriptors above as functional summaries, not quotes. Also, the ’878 specification you supplied includes a large glossary (CAS, STB, ISP/TVP, profile provider, "common IP address," etc.) that supports these claims, but the specification's glossary is not itself claim language.

4. Post-grant (PTAB) and litigation posture — this is where "10715878" is most active

PTAB (all on US 10,715,878):

Federal Circuit (the CAFC 2026 question):

• Meta Platforms, Inc. v. AlmondNet, Inc. and Intent IQ, LLC, No. 2025-1638 (lead appeal) — appeal from PTAB Nos. IPR2023-01281, IPR2024-00666, IPR2024-00701. Filed/docketed April 2025.
• No. 2025-1662 (Meta + Samsung) — consolidated as a member of 2025-1638 on 2025-05-09; Samsung was removed from 2025-1638 and 2025-1662 was dismissed per the court's nonprecedential (corrected) order of September 3, 2025; AlmondNet's extension to file its response brief (to Oct. 27, 2025) was granted. Mandate issued as to 2025-1662 only.
• Docket sources: https://dockets.justia.com/docket/circuit-courts/cafc/25-1662 ; https://www.cafc.uscourts.gov/09-03-2025-25-1638-meta-platforms-inc-v-almondnet-inc-order-25-1638-order-9-3-2025_2567215/ ; https://portal.unifiedpatents.com/litigation/Court%20of%20Appeals%20for%20the%20Federal%20Circuit/case/25-1638

⚠️ Uncertainty: I did not locate a merits opinion in No. 2025-1638 as of this writing; the latest authoritative item I found is the Sept. 3, 2025 nonprecedential order. If a 2026 merits decision exists, I could not verify it with the sources retrieved. I also saw a secondary news article (legalnewsfeed.com, ~Dec. 2025) describing a Federal Circuit ruling that "reinstated" a Meta infringement suit over an "implicitly construed" claim term. That description does not match the Sept. 3, 2025 order in 2025-1638 (a procedural consolidation/dismissal order), and may relate to a different appeal (e.g., an appeal of the N.D. Cal. dismissal). I therefore do not treat it as a finding about the ’878 patent without confirmation.

District court matters naming the ’878 patent (from the family-litigation listing and NPE database): W.D. Tex. 6:21-cv-00731 (Roku), 6:21-cv-00876, 6:21-cv-00891 (Samsung), 6:22-cv-01204/01205/01206 (Amazon/Meta/Microsoft); N.D. Cal. 4:22-cv-07515 and 4:23-cv-00438 (Meta); D. Del. 1:21-cv-01035 (Roku DJ) and 1:23-cv-01373 (LinkedIn). Asserted claims in at least one case were ’878 claims 13–18 and 20.

5. Bottom line

US 10,715,878 (Shkedi; Intent IQ LLC, licensed to AlmondNet) is a 2007-priority cross-device advertising patent that claims, in essence, inferring that two Internet devices share a home LAN via a common, dynamically assigned modem IP address, and then using one device's behavioral/demographic profile to trigger an automatic action (e.g., a targeted TV ad) on the other — with the analysis server outside the LAN. All three asserted PTAB challenges that reached a Final Written Decision (Meta/Roku/Samsung, IPR2023-01281 + joined IPR2024-00666 + IPR2024-00701) failed on 2025-02-19 ("no challenged claims unpatentable"), and the resulting appeal, Federal Circuit No. 2025-1638, remains the live appellate track after the companion appeal No. 2025-1662 was dismissed on 2025-09-03. I could not confirm a 2026 merits decision.

I'll research the litigation dockets for this specific patent number. Let me run targeted searches.

Let me verify each district court case and check for recent developments.

Litigation involving US 10,715,878 B2 ("the ’878 patent")

Scope note. I searched the ’878-specific identifiers and confirmed each case below by its docket record showing patent 10715878 in the case patent field, or by a filing in that case expressly naming the ’878 patent. I did not merge results for the many sibling patents in the same family (e.g., 7,822,639; 7,861,260; 8,244,582; 8,244,586; 8,281,336; 8,566,164; 8,595,069; 8,671,139; 8,677,398; 8,775,249; 8,959,146; 10,026,100; 10,321,198; 10,628,857). Where a case asserts the ’878 patent alongside those siblings, I say so; where a family case does not name ’878, I exclude it or flag it.

⚠️ Date discrepancy (carried forward): your task states April 26, 2026; my environment is dated 2026-09-28. All "current status" statements are as of the retrieval dates of the sources below.

1. District court cases naming the ’878 patent

Additional W.D. Tex. cases where ’878 appears in the patent field (verify before relying)

The ’878 patent also appears in the patent fields of the W.D. Tex. campaign against large platforms. Two of these are the Microsoft/Xandr matters, where the ’878 patent is expressly argued in the briefing:

• AlmondNet, Inc. et al. v. Microsoft Corporation and Xandr, Inc. — W.D. Tex., 6:21-cv-00897-ADA, related to 6:21-cv-01206-ADA. Microsoft/Xandr's summary-judgment motion of no direct infringement expressly addresses ’878 claim 1 (divided-infringement argument), confirming ’878 is in suit there. ⚠️ I did not obtain a final disposition.
• The 6:22-cv-01204 / 6:22-cv-01205 / 6:22-cv-01206 W.D. Tex. filings (associated with Amazon/Meta/Microsoft) appear in the ’878 family-litigation listing. ⚠️ I could not independently confirm that ’878 specifically (as opposed to only sibling patents) is asserted in each of these; treat those three case numbers as family-level, not individually verified for the ’878 patent.

Cases to EXCLUDE (family members, but NOT the ’878 patent)

• AlmondNet, Inc. v. Amazon.com, Inc. et al. — W.D. Tex. 6:21-cv-00898 ($122M June 2024 jury verdict). The verdict and JMOL briefing rest on U.S. Patents 7,822,639 and 8,671,139 — not the ’878 patent. Excluded.
• AlmondNet, Inc. et al. v. Oracle Corporation — W.D. Tex. 6:24-cv-00303 (filed 2024-06-03). Docket patent field lists 8,244,582 and 8,677,398 only. Excluded.

2. PTAB proceedings on the ’878 patent (post-grant, not Article III litigation but part of the enforcement picture)

3. Federal Circuit appeal (live track)

• Meta Platforms, Inc. v. AlmondNet, Inc. and Intent IQ, LLC, No. 2025-1638 (Fed. Cir.) — appeal from PTAB IPR2023-01281 + joined IPR2024-00666 + IPR2024-00701. Companion No. 2025-1662 (Meta + Samsung) was consolidated into 2025-1638 and then dismissed per the court's nonprecedential order of 2025-09-03 (Samsung removed; mandate issued as to 2025-1662 only). ⚠️ I found no merits decision in No. 2025-1638; the latest authoritative item I could verify remains the Sept. 3, 2025 order.

4. Bottom line

Yes — extensive litigation is known for US 10,715,878. The ’878 patent has been asserted or put at issue in at least six district-court matters: W.D. Tex. 6:21-cv-00731 (Roku), D. Del. 1:21-cv-01035 (Roku DJ), W.D. Tex. 6:21-cv-00891 (Samsung; transferred to N.D. Cal. 4:22-cv-07515), N.D. Cal. 4:23-cv-00438 (Meta), D. Del. 1:23-cv-01373 (LinkedIn), plus the Microsoft/Xandr matters (6:21-cv-00897 / 6:21-cv-01206). Most of these have resolved by dismissal/stipulation (Roku, LinkedIn) or stayed (Samsung), while Meta's validity challenge ran through PTAB IPR2023-01281 and joined IPRs (FWD 2025-02-19, patent upheld) and is now on appeal at the Federal Circuit in No. 2025-1638.

Caveats / unverified items: (1) The 6:22-cv-01204/01205/01206 W.D. Tex. filings are family-level and I could not confirm ’878 is individually asserted in each. (2) The Meta N.D. Cal. 4:23-cv-00438 disposition is unconfirmed. (3) The LiveIntent IPR2025-01317 patent linkage is unconfirmed. (4) No 2026 merits ruling in Fed. Cir. 2025-1638 was verifiable. I flag these rather than assert them.

I'll verify the PTAB proceedings against public sources, since the structured block and the patent text appear to conflict.

Good — the PTAB activity is real and contradicts the "no activity" block. Let me pull the remaining proceedings and the FWD panel/appeal status.

Let me pin down the FWD panel, institution date, and the appeal's current status.

I hit the search-step limit, but I gathered enough primary material to answer. One critical threshold issue first.

⚠️ Source conflict — the structured "PTAB proceedings on file" block is wrong

The task instructs me to treat the structured block as canonical and to default to "no PTAB activity on file." That block is contradicted by the authoritative patent text you supplied and by public PTAB records. The Google Patents record for US 10,715,878 itself carries a litigation section listing IPR2023-01281, IPR2024-00701, IPR2024-00666, IPR2024-00709, IPR2022-01505, and IPR2022-01315, and I retrieved actual PTAB papers (institution decisions, a rehearing decision, a joint motion to terminate, and a Final Written Decision excerpt) for each. So the ODP-derived block appears to be un-indexed/stale for this patent.

Per the operating rules ("prefer the search results when they contradict," and "don't fabricate"), I report the real proceedings below and do not adopt the "no activity" default. If you are relying on the ODP block for a clearance opinion, refresh it — it is materially incomplete.

Proceedings overview

Six AIA trials have been filed against US 10,715,878 across four petitioners: zero claims have ever been canceled — two petitions were denied institution (Roku, Samsung), one settled before institution (LinkedIn), and the one full trial to reach a Final Written Decision (Meta, with Roku and Samsung joined) ended 2025-02-19 with every challenged claim sustained. Defensive posture: this is a hardened patent on the § 103 prior-art front — claims 1–4 and 6–23 have survived a full merits test on the Baig/Laidlaw/Sitaraman art, so an ordinary obviousness-based IPR is a low-probability play; the realistic attack surface has shifted to § 101/§ 112 and to claim-construction/non-infringement arguments, plus whatever the pending Federal Circuit appeal (No. 2025-1638) yields.

Note on "claims-invalidated go to the top" ordering: there are none. I order by impact instead — the FWD proceeding first, then its joined members, then the settlement, then the two denials.

IPR2023-01281 — Meta Platforms, Inc. v. AlmondNet, Inc. & Intent IQ, LLC

(with joined IPR2024-00666 (Roku) and IPR2024-00701 (Samsung); caption on the joined papers reads "META PLATFORMS, INC., ROKU, INC. and SAMSUNG ELECTRONICS CO… v. ALMONDNET, INC. and INTENT IQ, LLC")

• Type: Inter Partes Review
• Filed: 2023-08-04 (Petitioner Meta Platforms, Inc.)
• Status: Final Written Decision issued 2025-02-19 (I use the date from the previously generated sections; the FWD text I retrieved is undated in the excerpt). Verdict: no challenged claims unpatentable.
• Judge panel: ⚠️ Not verified. The lead institution/FWD panel is not in the material I retrieved. The related 2022 denials (below) were decided by APJs Kristen L. Droesch, Stacey G. White, and Scott B. Howard; I do not assume the same panel sat on the FWD, and I will not guess.
• Petition grounds (from the Petition as filed and the exhibit list):
  • Ground 1 — claims 1–2, 8–10, 13–14, 19, 22 obvious over Baig (U.S. Pub. 2008/0113674) in view of Laidlaw (U.S. Pub. 2004/0128547), Sitaraman (U.S. Pat. 6,427,170), and Hahn.
  • Ground 2 — claims 3–4, 15–16 obvious over the Ground 1 art further in view of Xu (U.S. Pat. 7,730,030).
  • Ground 3 — claims 6–7, 17–18 obvious over the Ground 1 art further in view of Gerace (U.S. Pub. 2006/0282328).
  • Grounds covering claims 11, 12, 20, 21, 23 are not confirmed in what I retrieved (likely a further ground involving Bergs, Ex. 1012, but I will not state that as fact).
  • Statutory basis: § 103 (pre-AIA § 103(a)); the '878 patent claims 2007 priority.
• Institution decision: Instituted 2024-02-21 (per the declaration of record, "the Meta IPR … was instituted on February 21, 2024"). Challenged claims 1–4 and 6–23 (claim 5 was not challenged).
• Final Written Decision: "This Decision is a final written decision under 35 U.S.C. § 318(a) and 37 C.F.R. § 42.73 as to the patentability of claims 1-4 and 6-23 of the '878 patent. For the reasons that follow, we conclude that Petitioner has not shown by a preponderance of the evidence that claims 1-4 and 6-23 of the '878 patent are unpatentable." Claim-level result: claims 1–4 and 6–23 SUSTAINED (none canceled). Claim 5 was outside the challenge.
• Settlement / termination: none as to Meta.
• Appeal: Yes. Meta Platforms, Inc. v. AlmondNet, Inc. and Intent IQ, LLC, Fed. Cir. No. 2025-1638, appeal from PTAB Nos. IPR2023-01281, IPR2024-00666, and IPR2024-00701. On 2025-09-03 the Court issued a corrected nonprecedential order removing Samsung from 2025-1638, dismissing companion No. 2025-1662, and granting AlmondNet's extension to file its response brief to 2025-10-27. No merits opinion located as of retrieval — treat the appeal as pending. Order: https://www.cafc.uscourts.gov/opinions-orders/25-1638.ORDER.9-3-2025_2567215.pdf
• Defensive value: This is the controlling data point. Claims 1–4 and 6–23 have now survived a full IPR on the Baig/Laidlaw/Sitaraman combination — that art is now spent against these claims before the Board. A new defendant cannot expect an obviousness IPR on that record to fare better.

IPR2024-00666 — Roku, Inc. v. AlmondNet, Inc. & Intent IQ, LLC

• Type: Inter Partes Review ("copycat" joinder petition)
• Filed: 2024 (exact filing date not verified)
• Status: Instituted and joined/consolidated into IPR2023-01281; decided by the 2025-02-19 FWD — no challenged claims unpatentable.
• Judge panel: same as the lead proceeding (see caveat above).
• Petition grounds: Identical to IPR2023-01281 (Roku stipulated the challenges are the same; a joinder petition raising the Meta grounds). Roku stated it would "not pursue an invalidity defense in district court based on the same grounds as in this petition or any grounds that could have reasonably been raised" if joined. Also asserted: no § 325(d) issue because it relied on art not considered at prosecution or in the prior IPRs.
• Institution decision: instituted with joinder granted.
• Final Written Decision: none of the challenged claims unpatentable (by operation of the joined FWD).
• Settlement / termination: none.
• Appeal: covered by Fed. Cir. No. 2025-1638.
• Defensive value: Confirms Roku had to join Meta's winning-art petition (i.e., abandon its own 2022 theory) and still lost. Roku is now § 315(e)(2)-estopped on these grounds.

IPR2024-00701 — Samsung Electronics Co., Ltd. v. AlmondNet, Inc. & Intent IQ, LLC

• Type: Inter Partes Review ("copycat" joinder petition)
• Filed: 2024 (exact filing date not verified)
• Status: Instituted and joined/consolidated into IPR2023-01281; decided by the 2025-02-19 FWD — no challenged claims unpatentable.
• Judge panel: same as the lead proceeding (see caveat).
• Petition grounds: mirror of IPR2023-01281 (Samsung likewise concedes identical grounds to the Meta petition).
• Institution decision: instituted with joinder granted.
• Final Written Decision: none of the challenged claims unpatentable.
• Settlement / termination: none.
• Appeal: Samsung was a party to Fed. Cir. No. 2025-1662, then removed from 2025-1638 and 2025-1662 dismissed 2025-09-03 by the corrected order.
• Defensive value: Samsung's second bite (its first petition, IPR2022-01505, had been denied) also failed; Samsung is now estopped as to the joined grounds and has exited the appeal.

IPR2024-00709 — LinkedIn Corporation v. AlmondNet, Inc. & Intent IQ, LLC

• Type: Inter Partes Review
• Filed: 2024-03-20
• Status: Terminated – Settled (2024-07-16)
• Judge panel: not applicable (no institution decision; terminated pre-institution).
• Petition grounds: not retrieved; never reached the merits.
• Institution decision: none — the Board never instituted; it authorized the joint motion to terminate on 2024-06-12 and terminated the proceeding as to both parties.
• Final Written Decision: none.
• Settlement / termination: Joint motion to terminate under 35 U.S.C. § 317(a) / 37 C.F.R. § 42.71(a), granted 2024-07-16. The parties' Settlement Agreement is confidential (filed as Ex. 1024, business-confidential treatment requested). The co-pending D. Del. 1:23-cv-01373 was dismissed with prejudice (stipulation, Docket No. 20, 2024-06-06), each side bearing its own costs. Motion: https://www.docketalarm.com/cases/PTAB/IPR2024-00709/LinkedIn_Corporation_v._Intent_IQ_LLC/docs/06-24-2024-Petitioner/Motion__Motion_to_dismiss_due_to_settlement_pre_DI-7-Joint_Motion_to_Terminate_Inter_Partes_Review.pdf
• Appeal: none.
• Defensive value: A pure litigation-cost exit by a co-defendant; it produced no PTAB holding and gives a new defendant nothing. LinkedIn is not a source of estoppel or useful art.

IPR2022-01315 — Roku, Inc. v. Intent IQ, LLC

• Type: Inter Partes Review
• Filed: 2022-07-27 (challenging claims 1–23)
• Status: Institution denied; rehearing denied; POP review denied. ⚠️ The denial date is reported inconsistently in the sources I retrieved (2023-02-02 in one declaration, 2023-03-13 in another) — I flag the conflict rather than pick one.
• Judge panel: APJs Kristen L. Droesch, Stacey G. White, and Scott B. Howard; opinion authored by Judge White (rehearing decision, 2023-07-19).
• Petition grounds (pre-AIA § 103(a)):
  • Ground 1 — claims 1, 2, 5–7, 10, 13, 14, 17–19, 22 obvious over Madhavan in view of Baum, Harrington (U.S. Pat. 8,040,819), and Damick.
  • Ground 2 — claims 3, 4, 11, 12, 15, 16, 20, 21, 23 over Madhavan + Baum + Harrington + Damick + Rankin.
  • Ground 3 — claims 8, 9 over Madhavan + Baum + Harrington + Damick + Varghese.
• Institution decision: Denied. The Board "determined that Petitioner did not show a reasonable likelihood of prevailing" and, on rehearing, restated: "[W]e were not persuaded that the cited art taught or suggested the common LAN limitation found in each of the independent claims." Dec. 18–19. The panel reasoned that claim 1's preamble's "common LAN" provides antecedent basis for all later "the LAN" recitations and that Roku made "an insufficient factual showing to support its allegations regarding the recited common local area network." Rehearing decision: https://www.docketalarm.com/cases/PTAB/IPR2022-01315/Roku_Inc._v._Intent_IQ_LLC/docs/07-19-2023-Board/RehearingDecision_on_Request_for_Rehearing_-13-RehearingDecision_on_Request_for_Rehearing_.pdf
• Final Written Decision: none (never instituted).
• Settlement / termination: none.
• Appeal: Roku sought Precedential Opinion Panel review; POP denied and returned the matter to the original panel. No CAFC appeal (no appealable FWD).
• Defensive value: This is the "common LAN" gate. The Board has twice read the independent claims as requiring that the associating computer determine that multiple devices separately accessed the Internet through the router of a common LAN. Any new IPR must show that limitation in the art — that is where Roku failed.

IPR2022-01505 — Samsung Electronics Co., Ltd. v. Intent IQ, LLC

• Type: Inter Partes Review
• Filed: 2022-09-07 (challenging claims 1–23), with a concurrently filed motion for joinder to IPR2022-01315
• Status: Institution denied and joinder denied (2023-03-22)
• Judge panel: APJs Droesch, White, and Howard; authored by Judge White.
• Petition grounds: substantively identical to IPR2022-01315 — same claims, same § 103(a) combinations (Madhavan/Baum/Harrington/Damick, +Rankin, +Varghese), same expert declaration.
• Institution decision: Denied. The Board held the merits denial in the 1315 proceeding was dispositive, and because review was not instituted in the proceeding Samsung sought to join, it denied the joinder motion as well. Decision: https://www.docketalarm.com/cases/PTAB/IPR2022-01505/Samsung_Electronics_Co._Ltd._v._Intent_IQ_LLC/docs/03-22-2023-Board/Institution_Decision__Deny-10-Denying_Institution_of_Inter_Partes_Review_35_USC_%C2%A7_314_Denying_Motion_for_Joinder_35_USC_%C2%A7_315.pdf
• Final Written Decision: none.
• Settlement / termination: none.
• Appeal: none.
• Defensive value: A "me-too" petition that added no new art. It is a standing reminder that the Madhavan-based theory is dead and that the Board will not reward a second, identical filing.

Strategic summary

Claim status. No claim of US 10,715,878 has been canceled. Claims 1–4 and 6–23 were fully tested in IPR2023-01281 (and its joined members) and all were sustained on 2025-02-19; claims 1–23 were challenged but never instituted in IPR2022-01315/01505; claim 5 was never challenged at PTAB at all and is therefore untested. Put differently: every claim the Office has actually decided came back patentable, and the one claim (5) nobody touched is an unknown, not a weakness. "Canceled" set = ∅.

Estoppel landscape. Under § 315(e)(2), estoppel runs against parties who reached a FWD and their privies: Meta (grounds joined/reasonably could have been raised on Baig/Laidlaw/Sitaraman/Hahn/Xu/Gerace/Bergs), Roku, and Samsung, plus whatever Roku and Samsung stipulated away in their joinder petitions. A new, unaffiliated defendant is not estopped. But the practical runway is narrow: (i) the Meta/Roku/Samsung art has been adjudicated and largely rejected; (ii) reusing it invites § 325(d) and General Plastic discretionary denial; and (iii) § 101 and § 112 challenges are unavailable in IPR (only patents/printed publications under § 311(b)), so the strongest remaining attacks are district-court defenses, not AIA trials. The old Roku/Samsung Madhavan combination is the least attractive ground of all.

Pattern signals. Repeat-petitioner behavior is pronounced: Roku filed twice (IPR2022-01315 denied; IPR2024-00666 joined and lost) and Samsung filed twice (IPR2022-01505 denied; IPR2024-00701 joined and lost); Meta and LinkedIn each filed once. The patent owner litigated defensively and aggressively — it won institution denials against Roku and Samsung, defeated Meta on the merits, and obtained a POP review denial against Roku. On the appeal side, Meta and Samsung originally coordinated (consolidated appeals) but Samsung dropped out on 2025-09-03, leaving Meta as the lone appellant in No. 2025-1638. I found no confirmation that Unified Patents is a petitioner — the Unified Patents references in the Google Patents record are litigation-data links, not a party designation.

Recommended next steps

1. Anchor on the FWD. The operative disposition to cite is IPR2023-01281's FWD: "Petitioner has not shown by a preponderance of the evidence that claims 1-4 and 6-23 of the '878 patent are unpatentable." Retrieve the record from USPTO PTAB E2E / PTAB Center (proceeding IPR2023-01281) — the FWD excerpt I used is at https://ptacts.uspto.gov/ptacts/public-informations/petitions/[1555815](/patent/1555815)/download-documents?artifactId=ZhQDb9Hd0mRcqA4GK4Fx1dpecF0OI_A-4KgYzFFG8Vs_VLgoz7Y2Ae8. Confirm the panel composition and exact FWD date directly (I could not verify either).

2. Watch Fed. Cir. No. 2025-1638 as the single live risk. It is pending; AlmondNet's response brief was due 2025-10-27, so a 2026 argument/decision is plausible. Before relying on the PTAB outcome, check CourtListener / the CAFC docket for the merits opinion. If Meta prevails on an appeal issue (e.g., claim construction of "common LAN" or the "outside the LAN" limitation), the FWD's shelter disappears. Do not assume a 2026 decision exists — I could not locate one.

3. If you are a defendant now: don't build a defense on re-running the failed § 103 theories. Instead, (a) target the "common LAN" / "outside the LAN" limitations on the non-infringement side — note W.D. Tex./D. Del. construed terms including "automatically" = "done with little or no human control," which drove a Roku MSJ on non-infringement; (b) press § 101 (Roku litigated an Alice/§ 101 MSJ in D. Del. 1:21-cv-01035; the court/patent owner briefing points to material factual disputes that may defeat summary judgment, so scope the theory carefully); and (c) consider § 112 in district court, since it is unavailable in an IPR.

4. If you are contemplating a new IPR: weigh it heavily. You would need new art that teaches the common-LAN determination the Board has now twice found absent, and you would face § 325(d) and General Plastic headwinds given three prior petitions on this patent. Claim 5 is the only claim with no PTAB history — but challenging only claim 5 (a dependent claim) buys little.

5. Data hygiene: escalate the ODP "no PTAB proceedings" discrepancy. It is objectively wrong (the patent's own record lists six AIA trials), and any automated clearance report keyed to that block would understate both the patent's patentability track record and the estoppel already in place.

I'll research the assignment chain before writing. Let me run several targeted searches.

Let me try to pin down the assignment reel/frame details and correspondent of record.

I've reached my research limit. Note on sources: I could reach the Patent Assignment Center landing pages and Google Patents' legal-events feed (mirrored in the authoritative patent text), but I could not retrieve the Assignment Center record detail pages themselves — so the correspondent-of-record fields could not be verified, and I flag that rather than guess. Everything below that is stated as fact is grounded in the reel/frame numbers and dates that appear in the USPTO-sourced legal-events data embedded in the patent record.

Inventors

Pattern check — no unusual inventor-departure signal. The classic fire-sale precursor (inventors bailing within 12 months of filing) is absent: the lone inventor is the founder and current CEO of the assignee family. Shkedi's recorded chain (Shkedi → AlmondNet → Intent IQ, reel 047836/0899, 047836/0911) is an inventor-to-employer assignment, exactly as expected for a founder-employee, not an abandonment.

⚠️ Note: the sister patent '398 (app. 13/167,605) carries "Shkedi et al."; the '878 patent record itself names only Roy Shkedi. I did not locate an additional joint inventor on the '878.

Original assignee

• Entity named on the issued patent: Intent IQ, LLC (Google Patents bibliographic "Original Assignee: Intent IQ, LLC"; D. Del. AO-120 form lists the holder of the '878 as "INTENTIQ, LLC"). Intent IQ is a majority-owned subsidiary of AlmondNet, Inc. (per the AlmondNet/Yahoo E.D.N.Y. complaint: "Datonics and IIQ are majority-owned subsidiaries of AlmondNet").
• Primary line of business: Online advertising / ad-tech. Intent IQ is a privacy-first identity-resolution provider (probabilistic device graph across computer/mobile/CTV, cookieless and MAID-less environments); the parent AlmondNet "licenses intellectual property and enabling technology for the delivery of targeted ads"; sister company Datonics is audience data.
• Addresses of record: 37-18 Northern Boulevard, Suite 404, Long Island City, NY 11101 (AlmondNet, Datonics, and Intent IQ all share this address in the 2017–2021 filings); Intent IQ now stated as HQ New York City with an R&D center in Herzliya, Israel.
• Did they ship a product embodying the claims? Yes — Intent IQ markets an identity-resolution/device-graph product line; the group describes its portfolio as "licensed by some of the largest companies in the world, including Google and … Microsoft." AlmondNet's own description stresses R&D plus licensing of the portfolio.
• Current status: Operating. No bankruptcy, dissolution, or acquisition. The group publicly announced (2026-04-07) a licensing resolution with Amazon and continues as an operating plus licensing business. (Context: AlmondNet came close to insolvency in ~2013 and took a take-it-or-leave-it Google license; it survived and is still operating — no Chapter 7/11 was filed.)

Assignment timeline

Recorded against US 16/228,479 (the application that issued as US 10,715,878). All records below were retrieved from the USPTO-sourced legal-events data embedded in the patent record; execution dates are not exposed separately in that data and are shown as the recording date — treat execution as "on or about" that date.

Three records, all on 2018-12-20, all at reel 047836, then three corrections recorded 2019-09-06 at the same reel/frames. Note the assignee name was originally recorded with a typo, "ALMOLNDNET, INC." — I preserve it literally below and show the correction.

2018-12-20 / recorded 2018-12-20 — Reel 047836/0899

• Conveyance: Assignment (assignment of assignor's interest)
• Assignor: Shkedi, Roy (Mr.)
• Assignee: ALMOLNDNET, INC. (name as originally recorded; corrected to AlmondNet, Inc.)
• Correspondent: ⚠️ Not retrievable from the sources consulted; see note after the table.
• Context: Inventor-to-employer assignment — Shkedi assigns his rights in this continuation application to his own company, AlmondNet; i.e., the family's standing IP-holding restructure, not an arm's-length acquisition.

2018-12-20 / recorded 2018-12-20 — Reel 047836/0911

• Conveyance: Assignment
• Assignor: ALMOLNDNET, INC. (typo as recorded; corrected to AlmondNet, Inc.)
• Assignee: Intent IQ, LLC
• Correspondent: ⚠️ Not retrievable.
• Context: Internal reorganization — parent AlmondNet pushes title down to its IP-holding subsidiary Intent IQ (the entity named as patent owner thereafter). Consistent with Intent IQ's terminal disclaimer on the family, which lists "Owner: INTENT IQ, LLC — 100%."

2018-12-20 / recorded 2018-12-20 — Reel 047836/0917

• Conveyance: License (not an assignment of title)
• Assignor: Intent IQ, LLC
• Assignee: ALMOLNDNET, INC. (typo as recorded; corrected to AlmondNet, Inc.)
• Correspondent: ⚠️ Not retrievable.
• Context: License-back — Intent IQ retains ownership while granting parent AlmondNet a license. This exactly matches the PTAB representation that "The Patent Owner and real party-in-interest … is Intent IQ, LLC. AlmondNet, Inc. is a licensee of the '878 patent" (IPR2022-01505 Updated Mandatory Notice, 2023-01-09) and the litigation captions that sue as "AlmondNet, Inc. and Intent IQ, LLC."

2019-09-06 / recorded 2019-09-06 — Reel 047836/0899 (corrective entry on the above)

• Conveyance: Correction ("corrective assignment to correct the receiving party data")
• Assignor: Shkedi, Roy (Mr.)
• Assignee: AlmondNet, Inc. (name corrected)
• Context: Change of name only — fixes the "ALMOLNDNET" typo on the inventor-to-company assignment.

2019-09-06 / recorded 2019-09-06 — Reel 047836/0911 (corrective entry)

• Conveyance: Correction ("to correct the assignor's name")
• Assignor: AlmondNet, Inc. (corrected)
• Assignee: Intent IQ, LLC
• Context: Change of name only.

2019-09-06 / recorded 2019-09-06 — Reel 047836/0917 (corrective entry)

• Conveyance: Correction ("to correct the assignee's name … hereby confirms the license")
• Assignor: Intent IQ, LLC
• Assignee: AlmondNet, Inc. (corrected)
• Context: Change of name only — confirms the license record with the corrected parent name.

No post-issuance assignment to any third party appears in the record. There is no assignment after 2019-09-06. Net ownership today = Intent IQ, LLC (owner, 100%) with a license to AlmondNet, Inc. The three 2018 records simply replicate, for this continuation, the chain already recorded on the parent (app. 11/736,544 → now US 7,861,260).

⚠️ Correspondent-of-record gap (important for Signal 3): The Assignment Center exposes the recording attorney/agent for each reel/frame, but I could not retrieve those detail pages. I therefore cannot state who filed reel 047836/0899, /0911, or /0917, and I will not guess. What I can say is the prosecution correspondence for the family is an Intent IQ/AlmondNet in-house-style address (37-18 Northern Boulevard, Suite 404, Long Island City, NY 11101; roy@almondnet.com) with Louis J. Hoffman / Donald Hertz appearing as attorney of record on sibling app. 13/167,605, and Russ, August & Kabat (Reza Mirzaie et al.) as litigation counsel. Those are not the assignment correspondent and should not be conflated with it.

Timeline diagram

timeline
    title Ownership of US 10715878
    2007 : Priority application filed
    2018 : Continuation application filed
         : Shkedi assigns to AlmondNet
         : AlmondNet assigns to Intent IQ
         : Intent IQ licenses back to AlmondNet
    2019 : Name corrections recorded
    2020 : Patent issued to Intent IQ
    2021 : First assertion suits filed

NPE / troll-pattern signals

1. Shell-entity transfer — NOT PRESENT.
Title does move into a separate LLC (Intent IQ) via reel 047836/0911, but Intent IQ does not match the shell tells: no "IP/Holdings/Ventures/Licensing" suffix, it has a real product line (identity resolution/device graph), real offices (Long Island City NY; Herzliya IL R&D), a named chairman (Shkedi), and a stated corporate parent (AlmondNet). A transfer to an operating subsidiary is an internal IP-holding structure, not an anonymous shell. (The third record, 047836/0917, is only a license and transfers no title.)

2. Known asserter in the chain — NOT PRESENT.
Neither assignee (AlmondNet, Inc.; Intent IQ, LLC) appears on the named NPE lists (Acacia, Marathon, IV, IPNav, Wi-LAN, Mosaid/Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, Document Generation, Spangenberg entities), and no such entity appears anywhere in the reel 047836 chain. These are the original inventor's own companies. Caveat: they are a recurring plaintiff (RPX tracks the "AlmondNet/Intent IQ targeted advertising campaign" since 2016, and the Stanford NPE Litigation Database carries their cases — with at least one case tagged "Product company," i.e., category 8). Recurring litigant ≠ aggregator.

3. Repeat correspondent across the chain — UNCLEAR (not verifiable).
This is the one signal I cannot score, and I'm flagging that as a limitation, not a finding. The three 2018 records and three 2019 corrections sit at consecutive frames of a single reel (047836/0899, /0911, /0917), which is consistent with one filer/docket — but the Assignment Center correspondent fields were not retrievable, so I have no attorney name + firm to test for recurrence. A single appearance would not be a finding anyway; recurrence is, and I cannot establish it.

4. Cascading transfers — PRESENT IN FORM, BENIGN IN SUBSTANCE.
Three recordings on a single day (2018-12-20) at consecutive frames (0899 → 0911 → 0917), plus three corrections ~9 months later (2019-09-06). That is a cascade in form. But every party is the same corporate family (Shkedi / AlmondNet / Intent IQ), the middle step is a parent→subsidiary push-down, and the last step is a license-back, not a title transfer — so the "chained unrelated LLCs sharing a correspondent" tell is absent.

5. Pre-litigation transfer — NOT PRESENT.
The reel 047836 recordings (2018-12-20) predate the first suit naming the '878 (AlmondNet, Inc. et al. v. Roku, Inc., W.D. Tex. 6:21-cv-00731, filed 2021-07-15) by ~2.5 years — well outside the 6-month window. The recordings were made at application filing, tied to prosecution, not to an imminent complaint.

6. Bankruptcy fire-sale — NOT PRESENT.
AlmondNet neared insolvency around 2013 and took a Google license to survive, but there is no Chapter 7/11 and no sale of the '878 in any proceeding. The chain shows no estate/trustee assignment.

7. Privateering — NOT PRESENT.
No third-party operating company transferred patents to an NPE that asserts on its behalf. The enforcing entities are the original innovator family (Shkedi-controlled AlmondNet group), asserting its own R&D output; there is no separate operating-company beneficiary behind the reel 047836 chain.

8. Defensive aggregator — NOT PRESENT.
The chain does not terminate at RPX, AST, LOT Network, Unified Patents, or OIN. It terminates at Intent IQ, LLC (reel 047836/0911), which actively litigates.

Verdict

Operating-company assertion.

Justification: The record shows no shell transfer, no known NPE aggregator, and no fire-sale — title simply moved Shkedi → AlmondNet → Intent IQ (reel 047836/0899, /0911) with a license-back to AlmondNet (reel 047836/0917), all executed 2018-12-20 and merely corrected for spelling on 2019-09-06. The current 100% owner, Intent IQ, LLC, is an operating subsidiary of AlmondNet that actually ships identity-resolution/device-graph products and licenses its portfolio (Google, Microsoft, Amazon) while suing real ad-tech competitors (Roku 6:21-cv-00731; Samsung; Meta; Amazon; Microsoft; LinkedIn; Comcast; LiveIntent; Samba TV) — and Stanford's NPE database tags at least one AlmondNet case as "Product company." The only signal I could not clear is Signal 3 (repeat correspondent), which is unclear, not adverse, because the Assignment Center correspondent fields were not retrievable from my sources.

Verify at: USPTO Assignment Center search — https://assignmentcenter.uspto.gov/ (search US 10,715,878; legacy interface: https://assignment.uspto.gov/patent/index.html). The reel to inspect is 047836, frames 0899, 0911, 0917 (plus their 2019-09-06 corrective counterparts).

I'll search for the '878 patent's cited prior art references.

Let me get the full "References Cited" front page and the IPR grounds.

Let me pull the front-page reference list details, including foreign and non-patent citations.

Prior Art Analysis — US 10,715,878 B2 ("Targeted television advertisements based on online behavior")

Working date: April 26, 2026. Retrieval timestamp: 2026-09-28T20:44Z (environment clock). Building on the previously generated summary; I flag contradictions inline.

0. Scope, and an important honesty statement up front

You asked me to search USPTO for patent number 10715878 and to walk through each patent citation on the face of that patent, giving citation, date, description, and §102 claim mapping.

Three constraints I must disclose rather than paper over:

1. I could not query USPTO PatentCenter/Direct-Inspection or the "References Cited" (56) field directly. The Google Patents record supplied in the user message contains bibliographic data, classifications, description, abstract and litigation data — but it did not include the (56) References Cited list. I therefore reconstructed the citation list from the patent's own printed front page as filed in litigation/PTAB exhibits (URLs below). That is the same front-page text, but it is a secondary reproduction.
2. The front-page list I retrieved is OCR'd and truncated. Two independent reproductions disagree on several digits, and both cut off around the 2013–2014 entries. I mark every disputed entry rather than "fixing" it.
3. I will not fabricate descriptions or §102 mappings for references whose content I did not verify. For the same-inventor references and third-party patents where I could not confirm subject matter, I say so.

Contradiction flagged: the previously generated scope note recorded your prompt date as April 26, 2026 and the environment date as September 28, 2026. That discrepancy persists and is not resolvable from the sources I have; all docket facts are cited with retrieval context.

Contradiction flagged (relevant to this task): the previously generated section states the ’878 patent is a continuation of Ser. No. 11/736,544 (now U.S. Pat. No. 7,861,260). That matters enormously for §102: the ’878 patent's own priority document cannot be prior art to it. Some list entries below are the inventor's own family, not prior art. I explain in §3.

1. Sources used for the citation record

Not retrieved: the "Foreign Patent Documents" and "Other Publications" subsections of the ’878 front page. See §5 — this is a real gap you may want closed.

2. The ’878 patent's front-page (56) citations, as retrieved

2A. Same-inventor (Roy Shkedi) patents — listed, but generally not §102 prior art

These appear because the ’878 patent is one node in a very large Shkedi continuation family; the examiner/IDS lists them as related disclosures. They are marked below as "same-inventor."

Independent corroboration of one entry: the PTAB "State of the Art References for Asserted Patents" exhibit lists "Patent No. 7,861,260 (Shkedi) | 12/28/2010" alongside "10,715,878 (the '878 patent)." (https://ptacts.uspto.gov/ptacts/public-informations/petitions/[1555815](/patent/1555815)/download-documents).

2B. Third-party US patents on the face

I am not able to give you reliable per-reference content descriptions for §2B. I did not verify the disclosure of any of these third-party patents in this session, and I will not infer subject matter from assignee or title guesses. What I can say with confidence is that none of them was the basis of any adjudicated rejection of the ’878 claims, and none appears in the Board's Final Written Decision as anticipating art.

2C. Non-patent literature discussed in the ’878 specification (front-page listing not confirmed)

The ’878 specification itself discusses Battelle, The Search (Portfolio, New York, 2005) and a Google/BSkyB cross-media targeting arrangement. These read like IDS material, but I could not confirm that either appears under the front page's "Other Publications." Do not treat this as confirmed citation data.

3. The §102 framework that governs this list — and why most of it is not anticipation art

Critical dates. Earliest priority 4/17/2007 (pre-AIA regime). The ’878 patent's filing date (2018-12-20) is irrelevant to prior-art status because the chain reaches 11/736,544 (4/17/2007). So:

• §102(b) boundary: 4/17/2006 (one year before the earliest US filing in the chain).
• §102(e) boundary: a US patent/publication is §102(e) art only if "by another" and filed in the US before the applicant's invention date (≈4/17/2007). The issue date is not the operative date — so a 2010–2013-issued patent can still be §102(e) art if its underlying US filing is 2005–2006.

Why the §2A Shkedi list is not §102 anticipation art:

• US 7,861,260 is the ’878's own parent — same disclosure, same applicant. A priority document is not prior art to its own continuation. Its presence on the face is a continuity cross-reference, which is why it carries the asterisk and the G06Q 30/0241 classing.
• The other Shkedi patents are same inventive entity. Pre-AIA §102(a) and §102(e) both require the reference to be "by another." Roy Shkedi is the named inventor across this family. Where a family member names "Shkedi et al.," that would need case-by-case verification.
• Net effect: none of the ~20 Shkedi entries in §2A is a viable standalone §102 anticipation reference against claims 1, 13, or 22.

Why §2B is a weak §102 set:

• Pirani (1992) and Tsutsui (1993) clear the §102(b) date easily, but their content is decades pre-Internet-targeting and I did not verify any disclosure mapping to the "common LAN / dynamically assigned modem IP / cross-device action" limitations.
• The 2010–2013-issued third-party patents are only §102(e) candidates, and only as to subject matter the underlying pre-4/17/2007 application supports (MPEP 2136.02 / In re Wertheim). I did not verify filing dates or support.

Plain statement of the rule: anticipation under §102 requires one single reference disclosing every limitation arranged as in the claim. Every ’878 challenge that actually reached the merits was framed as §103, not §102.

4. The prior art that was actually litigated — this is the real "most relevant prior art" for the ’878 patent

Because the face-of-patent citations were largely the applicant's own family, the substantive prior art for this patent comes from the IPR petitions (Meta IPR2023-01281, with Roku IPR2024-00666 and Samsung IPR2024-00701 joined; LinkedIn IPR2024-00709 settled).

4A. Primary reference — Baig

⚠️ Flag on Baig's §102 status: Baig's publication (2008-05-15) postdates the 4/17/2007 priority date, and its utility filing (2007-10-31) also postdates it. It qualifies as prior art only through the 11/10/2006 provisional, and only if the specific invalidating disclosure is §112-supported by that provisional. That is a contested, two-pronged showing — worth noting if you intend to rely on it.

4B. Secondary references

4C. Grounds as pleaded, and result

Note the claim-set correlation to the previously generated summary: claims 1, 13 and 22 are the independent claims, and each of the three appears in Ground 1 — i.e., the only ground aimed at all three independents. Ground 1 failed, and with it the only route by which any single reference (Baig) came close to a §102-style full-limitation showing.

5. §102 mapping table — best available theory per reference, with confidence

6. Bottom line

• The ’878 patent's front-page (56) citations are dominated by the inventor's own Shkedi continuation family, which is why the printed list reads like a catalog of Shkedi patents (7,747,745 through 8,595,069 and beyond). As prior art, that block is largely inert: not "by another" under pre-AIA §102(a)/(e), and one of them (7,861,260) is the patent's own priority parent.
• The genuinely relevant §102-capable references are the IPR art: Baig (US 2008/0113674 A1, §102(e) via Provisional 60/858,204 filed 11/10/2006) is the closest thing to a single-reference anticipation theory against claim 1; Laidlaw and Sitaraman are clean §102(b) references but supply only elements. No §102 ground was actually pleaded or sustained.
• As adjudicated, no reference anticipates and no combination renders obvious claims 1–4 and 6–23: PTAB IPR2023-01281 + joined IPR2024-00666 + IPR2024-00701, Final Written Decision 2025-02-19 — "no challenged claims unpatentable." The appeal is CAFC No. 2025-1638 (companion 2025-1662 dismissed 2025-09-03). Consistent with the previously generated section, I found no 2026 merits decision.
• Claim-number hygiene (per the operating rule): "10715878" refers only to US 10,715,878 B2. The "78" inside the reference numbers above (e.g., 10,178,442; 2004/0128547) are distinct identifiers and I have not conflated them.

7. What I could not do, so you can decide whether to chase it

1. Full (56) list is truncated at ~2013–2014 in both reproductions; the continuation of the Shkedi block (2014–2020) and the Foreign Patent Documents / Other Publications subsections were not retrieved. Verify at https://patentcenter.uspto.gov/applications/16228479 ("References Cited") or the Google Patents "Citations" tab.
2. Filing/publication dates (as distinct from issue dates) are unverified for every §2B third-party patent except Laidlaw, Sitaraman, Baig, Xu, Gerace, Bergs. Without those, no honest §102(e) determination can be made.
3. OCR digit conflicts in §2A/§2B (e.g., 8,200,82x; 8,244,58x; 8,280,75x; Matz/Szalya; Ramanathan/Jenkins; Muldawer/Madhavan; Ige/Igoe) — I deliberately did not auto-correct them. Each needs a single authoritative lookup.
4. Content descriptions for ~20 third-party front-page patents were not verified. If you need them, the fastest path is Google Patents' "Cited By / Citations" view for US10715878B2, which renders each cited reference with title, filing date, and publication date in one pass.

I'll research the actual prior-art record before analyzing—the Google Patents page's "Prior Art" section lists keywords (internet, computer system, LAN, online, STB) rather than references, so I need to ground the § 103 analysis in the references actually cited in the IPRs.

Good — I now have the actual grounds. Let me get details on the key references and the Meta grounds.

⚠️ Two flags before the analysis

1. Date discrepancy (carried forward). This task states the current date as April 26, 2026; my retrieval environment is stamped September 28, 2026. I state docket facts with retrieval dates and do not silently reconcile the two.
2. What "the Prior Art section of this page" actually contains. The Google Patents page you supplied has a "Prior art keywords" field — internet, computer system, LAN, online, stb — and a "Prior art date" of 2007-04-17. It does not enumerate prior-art references. I therefore treat the keyword set as the examiner/PTO's conceptual field definition, and ground the § 103 analysis in the operative prior-art record for this patent: the IPR petitions and Board decisions, plus the references cited on the face of the ’878 patent. I flag verbatim-verified vs. functional descriptions throughout. I build on, and do not repeat, the prior sections' claim/claim-number and docket content.

Obviousness analysis — US 10,715,878 B2 under 35 U.S.C. § 103

1. Governing framework

Statute. Pre-AIA § 103(a) applies. The ’878 patent issued from app. 16/228,479 (filed 2018-12-20) but claims an unbroken priority chain to app. 11/736,544 (filed 2007-04-17), and no claim has an effective filing date on or after 2013-03-16. Petitioners in IPR2022-01315 expressly pleaded "pre-AIA U.S.C. § 103(a)" — consistent with this analysis. Consequences: (a) pre-AIA § 102(e) applies to published applications/patents, keyed to their filing dates; (b) § 102(b) art must predate 2006-04-17 by one year.

Standard. Graham v. John Deere factors as refined by KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007): the claims are invalid if the differences from the prior art would have been obvious to a POSITA, and a motivation to combine need not be found in the references themselves — it can come from the "design incentives and other market forces," from "common sense," or from the fact that the combination is "of familiar elements according to known methods… yielding predictable results." KSR, 550 U.S. at 416–21.

POSITA. Per the Meta/Samsung petition (Paper 2, IPR2023-01281/00701): a bachelor's degree in EE or CS plus two years' experience in network-based computer systems (LAN/WAN/Internet information transfer). The Roku petitions used Dr. Samrat Bhattacharjee; Meta/Samsung used Dr. Christopher M. Schmandt; Patent Owner used Dr. Erik de la Iglesia and Ms. Sylvia Hall-Ellis.

Claim construction. No independent construction issued; the Board treated claim 1 as representative of claims 1, 13, and 22 because they "recite 'substantially identical' subject matter" (Denial, IPR2022-01315, at 18). All § 103 groups below therefore rise or fall on claim 1, with dependents adding mostly conventional features.

Critical structural history. Every § 103 challenge tried against this patent to date has failed:

A § 103 analysis that ignored this record would be misleading. So I analyze two questions: (A) are the claims legally obvious on the art pleaded, and (B) why has that art nonetheless failed.

2. The two independent § 103 combinations on the record

Combination I — the "Meta" ground: Baig + Laidlaw + Sitaraman + Hahn

Baig — US 2008/0113674 A1 (utility filed 2007-10-31; prov. 60/858,204 filed 2006-11-10). The primary reference. As characterized by Petitioner and uncontested as to substance: Baig groups mobile devices that are connected to the same WiFi access point or hotspot (e.g., a coffee shop) into a group, then provides services (chat) to devices in the group. Baig supplies group membership determined by shared network attachment, device identifiers, and a server outside the local network.

Laidlaw — US 2004/0128547 A1 (published 2004-07-01, § 102(b)): session management, cookies, and a session table letting a web server recognize returning users. Used for claim-elements (a)(i) "device identifier" in cookie form, and claims 8–9.

Sitaraman — US 6,427,170 B1 (issued 2002-07-30, § 102(b)): DHCP-based dynamic IP address assignment. Used for (a)(ii) "dynamically assigned IP address." This is powerful because the ’878 specification itself admits DHCP dynamic assignment as known (’878 spec, DHCP/CMTS/DSLAM glossary; cited at 4:7–12 by Petitioner). Self-admitted prior art and a § 102(b) reference are mutually reinforcing.

Hahn — The Internet Complete Reference (2d ed. 1996): a printed publication (authenticated by the Hall-Ellis library-availability declaration) teaching automatic updating/delivery of webpages to visitors. Used for the "automatic action" element and claims 6–7.

Xu — US 7,730,030 and Gerace — US 2006/0282328 A1: dependents only (mobile-device behavioral data collected off-LAN and redirection; and ad selection from a plurality based on profile plus directing the ad to the second device).

Limitation mapping (claim 1):

Motivation to combine (KSR):

• Same field of endeavor: all five references address identifying networked client devices and delivering content/services to them. KSR permits combination of references "in the same field."
• Predictable result: the combination is the mechanical union of cookie identification (Laidlaw) + DHCP-recorded addresses (Sitaraman) + shared-attachment grouping (Baig) + automated delivery (Hahn). No reference's principle of operation is altered.
• Market force: monetizing household/device-level profiling for targeted advertising was a recognized commercial driver — the ’878 specification itself recites the Google/BSkyB alliance and Battelle's The Search (2005) as the state of the art, i.e., applicant-admitted prior art motivating exactly this combination.
• Design incentive: doing so without PII (claim 10) is recited as the goal, which cuts against nonobviousness of the goal itself.

Why this combination failed before the Board. The FWD of 2025-02-19 (IPR2023-01281/00666/00701) found no challenged claim unpatentable. PO's expert (de la Iglesia) mounted three defenses that map onto the claim text: (i) Baig associates users, not device identifiers — the claim requires association between device identifiers; (ii) Baig's hotspot is transient and public, so the modified system would not determine that devices "separately accessed the Internet… during [a] predetermined period of time" in the home-modem/DHCP-lease sense; (iii) the modified Baig "automatic action" (sending a user list) does not "comprise[] selecting an advertisement… based at least in part on… user profile information" as claims 6–7 require. These are genuine claim-differentiating gaps, not merely conclusory denials.

§ 102(e) vulnerability to monitor. Baig's utility filing (2007-10-31) post-dates the ’878 priority date; Baig is prior art only insofar as its relied-upon disclosures are supported by the 2006-11-10 provisional (Petitioner submitted a redline, EX1005, for exactly this reason). Dynamic Drinkware requires the petitioner to prove the provisional supports the challenged disclosure. If the provisional does not support the "grouping" disclosure, Combination I collapses on prior-art status alone.

Combination II — the "Roku/Samsung" ground: Madhavan + Baum + Harrington + Damick (+ Rankin, + Varghese)

Madhavan — US 8,180,674 (primary); Baum — US 7,843,923 (relied on for the common-LAN element, via Baum FIG. 5 — multiple IP-addressed devices on a single premises LAN — read with FIG. 8 — co-located IP addresses behind a common port/router); Harrington — US 8,040,819 and Damick — US 2006/0242294 A1 (secondary); Rankin — US 2003/0207685 and Varghese — US 7,908,645 (dependent claims only).

⚠️ Verification caveat: I have not independently verified Madhavan's, Harrington's, or Damick's verbatim disclosures. I rely on the petitions'/Board's characterizations: Roku asserted Madhavan teaches the IP-address-to-premises association and also argued that Madhavan alone renders the "common LAN" limitation obvious, with Baum as the primary common-LAN teaching.

Ground structure:

• Ground 1 — claims 1–2, 5–7, 10, 13–14, 17–19, 22 over Madhavan + Baum + Harrington + Damick.
• Ground 2 — claims 3–4, 11–12, 15–16, 20–21, 23 adding Rankin.
• Ground 3 — claims 8–9 adding Varghese.

Motivation to combine: Baum teaches a premises LAN with multiple IP-addressed devices interfacing through a same port and same router; Madhavan teaches using an IP address to associate a device with a location/premises for content targeting. A POSITA seeking to determine that two devices are co-located behind one modem/router would combine these as a predictable use of known network topology. Roku additionally invoked a related-family institution, IPR2022-00773 (Meta v. AlmondNet), Paper 34 (2022-10-21), which was instituted on a same-specification sibling patent.

Why this combination failed: the Board denied institution twice, holding the petitions presented "an insufficient factual showing… regarding the recited common local area network ('LAN')" and reading Baum FIG. 8 in isolation, finding no reference to multiple IP addresses "co-located at the same location or port number" (Denial at 18–24). Roku's rehearing request and POP request argued the Board misapprehended FIGS. 5/8 and its expert's unrebutted testimony — both refused. This is the single most important validity datum for the ’878 patent.

3. Additional § 103 avenues (less developed, some untried)

4. Would the claims be obvious? Assessment

Legally, yes — there is a defensible § 103 case. Cutting across both combinations:

• The ’878 claims are combinations of individually known elements: dynamic IP assignment (admitted/ Sitaraman), cookie device identification (Laidlaw), shared-network grouping (Baig/Baum/Cordray), timestamped records (Baig/Hahn), and delivering an ad to a second device based on a first device's profile (Gerace; and the ’878 spec's own recitation of the known cross-media art).
• KSR permits combination of familiar elements yielding predictable results, and does not require the motivation to appear in the references. The commercial pressure to monetize online profiling into TV-ad targeting is documented in the ’878 specification itself.
• The claim's most distinctive limitations are negative or architectural — "outside the LAN," "no personally identifiable information," "during a predetermined period of time via a common IP address" — and each has an obvious engineering motivation (privacy, accuracy, DHCP leases).

But the claims have so far survived on the merits, and the survival is substantive, not formal. Across six PTAB challenges and two institution denials, the failing limitation is consistently the same: proving that a single server associates two device identifiers — not two users — on the basis of their separately accessing the Internet through a common LAN's dynamically assigned modem IP address within a predetermined window, and then acts on one device using the other's profile. Petitioners' art either (a) groups users (Baig), (b) shows co-location without establishing separate access through a common router (Baum FIG. 8 alone), or (c) requires an unsupported leap from a chat service to ad selection from a plurality based on user profile (Baig→Gerace).

My bottom-line assessment:

Caveat on secondary considerations. I found no evidence in the record I retrieved of objective indicia (nexus-bearing commercial success, long-felt need, licensing due to validity, copying) either asserted by Patent Owner or rebutted by Petitioners. Under Graham/KSR those factors can be outcome-determinative and their absence from the retrieved record means this analysis cannot exclude them.

Sources (retrieval 2026-09-28 unless noted):

• Google Patents record for US 10,715,878 B2 (prior-art keywords; prior-art date 2007-04-17): https://patents.google.com/patent/US10715878/en
• Meta petition & exhibit list, IPR2023-01281: https://www.docketalarm.com/cases/PTAB/IPR2023-01281/Meta_Platforms_Inc._v._Intent_IQ_LLC/docs/03-27-2024-Petitioner/Notice__Exhibit_list-10-Petitioners_Updated_Exhibit_List.pdf
• Samsung/Meta joinder petition text (Ground 1–4 tables; Baig/Laidlaw/Sitaraman/Hahn/Xu/Gerace): https://ptacts.uspto.gov/ptacts/public-informations/petitions/[1555815](/patent/1555815)/ and .../1555823/
• Roku expert declaration (Madhavan/Baum/Harrington/Damick/Rankin/Varghese; Grounds 1–3): https://www.docketalarm.com/cases/PTAB/IPR2022-01315/Roku_Inc/docs/07-27-2022-Petitioner/Exhibit-1003-37-Declaration_of_Dr_Samrat_Bhattacharjee.pdf
• Roku request for rehearing & POP request (common-LAN; Baum FIG. 5/8): https://www.docketalarm.com/cases/PTAB/IPR2022-01315/Roku_Inc._v._Intent_IQ_LLC/docs/04-12-2023-Petitioner/Rehearing_request_Institution-10-Petitioners_Request_for_Rehearing.pdf
• IPR2024-00701 docket (institution 2024-07-24; FWD 2025-02-19): https://ipverse.greyb.com/ptab-web/cases/case-details/IPR2024-00701
• CAFC corrected order, No. 2025-1638 (2025-09-03): https://www.cafc.uscourts.gov/opinions-orders/25-1638.ORDER.9-3-2025_2567215.pdf