US 8204783

I'll search for authoritative information on US 8204783, including any 2026 CAFC activity.

Let me verify whether US 8204783 itself is tied to the 2026 CAFC activity, and check its litigation history.

Let me confirm whether the '783 patent itself appears in any 2026 Federal Circuit docket, and identify the other Texas case.

US 8,204,783 B2 — Summary

1. Bibliographic data

Source: https://patents.google.com/patent/US8204783/en

Note on identifiers: Do not confuse US 8,204,783 with NL 8204783 A (a 1982 Philips Dutch application for applying a metal layer to a substrate) — entirely unrelated. Also note that some litigation-derived docket text renders an AlmondNet patent as "U.S. Patent No. 8,224,582"; the PTAB record for IPR2022‑01436 identifies it as 8,244,582 B2. Per your instruction I am reporting both as they literally appear rather than silently correcting.

2. Abstract (verbatim)

"An automatic system facilitates selection of media properties on which to display an advertisement, responsive to a profile collected on a first media property, where a behavioral-targeting company calculates expected profit for an ad correlated with the profile and arranges for the visitor to be tagged with a tag readable by the selected media property. The profit can be calculated by deducting, from the revenues that are expected to be generated from an ad delivered based on the collected profile, at least the price of ad space at a media property where the BT company might like to deliver ads to the profiled visitor. When the calculated profit is positive (i.e., not a loss), the BT company arranges for the visitor to be tagged with a tag readable by the selected media property through which the BT company expects to profit."

3. Independent claims in plain language

Claim 1 (method). For each of many visitors to a first media property, when the system receives information about one of several possible profile attributes applicable to the visitor, it (a) automatically authorizes a third-party second media property to allow display of a profile-correlated ad to that visitor when the visitor later visits the second property, subject to a visitor‑specific condition (e.g., a time window — see dependent claims 10 and 15); and (b) automatically authorizes a third-party third media property to allow display of a profile-correlated ad to that same visitor when the visitor later visits the third property, where the price the third property charges is less than the profile‑attribute‑dependent price an advertiser is willing to pay for the display. In short: two authorizations, one gated by a visitor-specific condition, the other gated by a profitability/price-cap comparison. Note the claim language is "authorizing … to allow display," i.e., permission-granting rather than serving the ad.

Claim 18 (tangible computer-server device). The same two-step method, but recited as a programmed server ("tangible computer-server device programmed to implement" it). Substantively identical limitations to claim 1.

Claim 35 (system). The same two-step method, recited as "a system comprising one or more computers programmed to implement" it. Again substantively identical to claim 1.

Representative dependent-claim coverage (claims 2–17, 19–34, 36–51 largely mirror each other across the three independent claims): building a look-up data structure mapping profile attributes to second media properties (claims 2, 3); acting responsive to a redirect of a visitor computer carrying profile-attribute information (claims 4, 10); recognizing an action on the first property (claim 5); receiving a search request (claim 6); receiving a visitor "characteristic" (claim 7); categorizing the attribute into predefined categories (claim 8); placing a tag readable by the second media property (claim 9), including by redirecting the visitor so the second property places its own tag (claims 13–17); authorization limited to a specified time period (claims 10, 15); ad space in a different medium than the first property (claim 11); ad space controlled only temporarily (claim 12); and delivering the condition to the second property via a redirect URL (claims 17, 34, 51).

4. Litigation / PTAB / Federal Circuit status (as of April 26, 2026)

• 2026 Federal Circuit activity involving this patent number: none found. The April 13, 2026 Federal Circuit judgment that invalidated AlmondNet claims (Law360 headline: "Fed. Circ. Affirms Meta's Win Over AlmondNet Ad Tech Patent") arises from AlmondNet, Inc. v. Meta Platforms, Inc., No. 2024‑1834, which is an appeal from IPR2022‑01436 addressing U.S. Patent 8,244,582 B2 — not the '783 patent. It was a nonprecedential Rule 36 affirmance (Lourie, Dyk, Taranto). Sources: https://cafc.uscourts.gov/opinions-orders/24-1834.RULE_36_JUDGMENT.4-13-2026_2674964.pdf and https://law.justia.com/cases/federal/appellate-courts/cafc/24-1834/24-1834-2026-04-13.html.
• The other 2026 AlmondNet appellate matter, 26‑1326 (and cross‑appeal 25‑1967), was voluntarily dismissed under FRAP 42(b) (mandate noted April 6, 2026 in W.D. Tex. 6:21‑cv‑00898). The patents reported at issue there are US 8,671,139 B2 and US 7,822,639 B2 — again not the '783 patent. Sources: https://www.law360.co.uk/retail/articles/[2461759](/patent/2461759)/amazon-almondnet-drop-appeal-of-136m-patent-verdict and the 6:21‑cv‑00898 docket.
• District court: The '783 patent was asserted by AlmondNet, Inc. and Intent IQ, LLC in AlmondNet, Inc. et al. v. [Amazon.com, Inc.](/litigations/by-plaintiff/Amazon.com%2C%20Inc.) et al., No. 6:22‑cv‑01204 (W.D. Tex., Judge Albright), filed Nov. 18, 2022, alongside US 7,747,745, 7,861,260, 10,715,878, 8,959,146, 8,775,249 and 7,979,307. The case was dismissed with prejudice by joint stipulation dated April 26, 2023, with each side bearing its own fees/costs. Source: https://www.docketalarm.com/cases/Texas_Western_District_Court/6--22-cv-01204/AlmondNet_Inc._et_al_v._Amazon.com_Inc._et_al/docs/34.pdf
• The Google Patents page for this family also lists a second W.D. Tex. case, 6:22‑cv‑01205, as family litigation; a Meta complaint excerpt asserts the '783 patent in W.D. Tex. (6:21‑cv‑00896). I could not independently confirm in this session which defendant corresponds to 6:22‑cv‑01205, so I flag that as unverified.
• PTAB: '783 prosecution history was filed as an exhibit in Microsoft petitions IPR2022‑01319 (Ex. 1005) and IPR2022‑01324 (Ex. 1009). I could not confirm in this session whether the '783 patent itself was a challenged patent in those IPRs or merely a related-family exhibit — treat that as uncertain. Separately, IPR2022‑01505 (Samsung v. Intent IQ) lists US 8,204,783 in a patent-family exhibit table.

5. Explicit uncertainty statements

• I found no 2026 CAFC docket in which US 8,204,783 is the patent on appeal; if such a docket exists, it is not reflected in the sources retrieved above.
• Litigation-sourced OCR of docket/PTAB text for the Meta IPR rendered the patent as "8,224,582" while the PTAB Final Written Decision caption reads "Patent 8,244,582 B2"; I did not resolve which is a typographical error, and I have reported both literally.
• All bibliographic and claim information above is taken from the full patent text you supplied, which is authoritative for this task; the litigation/PTAB/Fed. Cir. items are from the cited secondary sources and dockets.

I'll research litigation involving US patent 8204783 using the sources you specified.

Let me dig deeper into the specific cases.

Let me search for additional cases and the litigation portal data.

Litigation Involving U.S. Patent No. 8,204,783

Summary of what the sources show

U.S. 8,204,783 ("Media properties selection method and system based on expected profit from profile-based ad delivery"), assigned to AlmondNet, Inc. (inventor Roy Shkedi), has been drawn into the AlmondNet/Intent IQ campaign against large ad-tech platforms. The Google Patents record for US8204783B2 itself carries a "Family has litigation" flag pointing to two W.D. Tex. cases (6:22‑cv‑01205 and 6:22‑cv‑01204) plus a Darts‑ip worldwide-family litigation entry (family 38832126).

Note on the Unified Patents portal: the caselist URL you gave (portal.unifiedpatents.com/litigation/caselist) is a JavaScript application and did not return case records in retrievable text form; I therefore corroborated via Stanford NPE/Lex Machina, UniCourt, Justia, PACER-derived dockets, PTAB filings (which reproduce the AO‑120 forms and docket excerpts), and the Google Patents litigation links.

---

1. AlmondNet, Inc. et al. v. Meta Platforms, Inc. — CONFIRMED to assert 8,204,783

Source: https://npe.law.stanford.edu/case/[269881](/patent/269881) ; PTAB exhibit lists referencing "Joint Stipulation to Stay and Order filed in AlmondNet, Inc. v. Meta Platforms, Inc., Case No. 4:23‑cv‑00438 (N.D. Cal.)"; UniCourt docket for 6:22‑cv‑01205 (https://unicourt.com/case/pc-db5-almondnet-inc-et-al-v-meta-platforms-inc-[1342872](/patent/1342872)).

---

2. AlmondNet, Inc. et al. v. [Amazon.com, Inc.](/litigations/by-plaintiff/Amazon.com%2C%20Inc.) et al. — family-level link; '783 assertion NOT verified

Source: https://docs.justia.com/cases/federal/district-courts/texas/txwdce/6:2022cv01204/[1197266](/patent/1197266)/36 ; https://patents.google.com/patent/[US8204783](/patent/US8204783)/en

---

3. Related PTAB proceedings (not district-court litigation, but they name/attach '783)

• [Microsoft Corp.](/litigations/by-plaintiff/Microsoft%20Corp.) v. AlmondNet, Inc., IPR2022‑01319. Petitioner's Exhibit 1005 (filed July 27, 2022) is captioned "Prosecution History of U.S. Patent No. 8,204,783." A prosecution-history exhibit is normally filed for the challenged patent, so this IPR very likely targeted 8,204,783 — but I am not stating that as confirmed; I could not retrieve the IPR's certificate/patent caption. Source: https://services.patexia.com/lawsuits/Microsoft-Corporation-v-AlmondNet-Inc-id-[179779](/patent/179779)/documents/sort_by/created/sort_dir/asc

• Meta Platforms, Inc. v. AlmondNet, Inc. IPR proceedings include IPR2022‑01436; reported as "challenging claims of the '582 patent." The resulting appeal is AlmondNet, Inc. v. Meta Platforms, Inc., Fed. Cir. No. 2024‑1834, decided by Rule 36 affirmance on April 13, 2026 (panel: Lourie, Dyk, Taranto); Law360 headline: "Fed. Circ. Affirms Meta's Win Over AlmondNet Ad Tech Patent" (PTAB invalidated all claims Meta challenged). Caution: the record ties IPR2022‑01436 to the '582 patent, not 8,204,783 — so do not conflate this affirmance with the '783 patent. Sources: https://cafc.uscourts.gov/opinions-orders/24-1834.RULE_36_JUDGMENT.4-13-2026_2674964.pdf ; https://www.law360.com/texas/articles/[2464582](/patent/2464582)/

• [[Samsung Electronics Co.](/litigations/by-defendant/Samsung%20Electronics%20Co.) Ltd.](/litigations/by-plaintiff/Samsung%20Electronics%20Co.%20Ltd.) v. Intent IQ, LLC, IPR2022‑01505 (patent 10,715,878) — Exhibit 1002 lists "US‑8204783" only as related/background material. Not a challenge to '783.

---

4. Other AlmondNet campaign matters (no verified '783 link)

• AlmondNet, Inc. and Intent IQ, LLC v. Samsung Electronics Co., Ltd.; Samsung Electronics America, Inc.; Adgear Technologies Inc. — W.D. Tex., filed 8/26/2021 (per AO‑120). Patents listed: 7,979,307; 8,200,822; 8,244,582; 8,566,164; 8,671,139, later supplemented with 8,677,398; 8,959,146; 10,321,198; 10,715,878. 8,204,783 is not on that list.
• Roku, Inc. v. AlmondNet, Inc. — declaratory-judgment action, D. Del. (referenced on the same AO‑120); patent(s) at issue not identified in what I retrieved.
• An IAM interview (Aug. 2024) reports AlmondNet had pending litigation against Samsung, Roku, Meta, LinkedIn and Oracle, and a June 2024 W.D. Tex. jury verdict of $122 million against Amazon on U.S. 7,822,639 and 8,671,139. Neither the verdict patents nor any of those pending suits were identified as involving 8,204,783.

---

Bottom line

• The only case I can affirmatively confirm as asserting U.S. 8,204,783 is AlmondNet, Inc. et al. v. Meta Platforms, Inc., filed as W.D. Tex. 6:22‑cv‑01205 (11/18/2022), transferred to N.D. Cal. as 4:23‑cv‑00438, and stayed in April 2023 pending the IPR decisions.
• The Amazon case (6:22‑cv‑01204, dismissed with prejudice 6/1/2023) is family-linked on Google Patents but I could not confirm '783 was pleaded there.
• Microsoft's IPR2022‑01319 is the one PTAB proceeding whose exhibit set squarely references the '783 patent's prosecution history; whether '783 is the challenged patent should be verified directly in the PTAB record.
• I found no final judgment holding 8,204,783 invalid or unenforceable, and no CAFC merits decision on the '783 patent itself (the April 13, 2026 Rule 36 affirmance appears to concern a different family member, the '582 patent).

If you need certainty for an opinion or disclosure, the next step is to pull the actual docket in N.D. Cal. 4:23‑cv‑00438 (patent-in-suit list and current stay status) and the PTAB docket/certificate for IPR2022‑01319 (to confirm the challenged patent). I'd want those primary documents before relying on the '783-specific assertions above beyond the Meta case.

I'll verify the structured block's "no PTAB activity" default with targeted searches before writing this up.

The structured block says no AIA proceedings, but search shows AlmondNet has faced many IPRs on sibling patents. I need to pin down whether any targeted the '783 patent specifically.

Good — I've confirmed the '783 patent's siblings were heavily attacked. Now I need to determine whether any proceeding targeted 8204783 itself, plus its parent '745.

Proceedings overview

Zero AIA trial proceedings (IPR / PGR / CBM) have ever been filed against U.S. Patent No. 8,204,783. The USPTO Open Data Portal structured block in my prompt returns no AIA trial proceedings, and my web searches — which surfaced a very large PTAB docket against sibling patents in this family — turned up no petition, institution decision, FWD, or appeal naming the '783 patent as the challenged patent; the 0/0/0/0/0 breakdown (active / invalidated / sustained / settled / denied institution) is therefore a genuine null result, not a data gap. Bottom-line posture for a defendant: you are facing an asserted-but-never-adjudicated patent — its claims have survived zero IPRs and zero Markman constructions — which cuts both ways: no adverse validity record exists to exploit, but the '783 patent was also asserted against Meta Platforms and then voluntarily dismissed by the patent owner in 2024, and its four sibling continuations have been on the receiving end of roughly fifteen PTAB petitions, one of which ended in a Final Written Decision cancelling all challenged claims.

Confidence and caveats. I am confident about the negative finding for '783 (two independent source lines agree). I am not able to rule out an older, unindexed proceeding filed in the 2012–2017 window with full certainty — I found no evidence of one, but I could not exhaustively review every PTAB filing year-by-year against this patent number before my search budget ran out.

---

Proceedings targeting US 8,204,783

None. There is no {PROCEEDING_NUMBER} — {Petitioner} v. {Patent Owner} section to write, and I will not manufacture proceeding numbers, panels, or dispositions to fill the template.

One thing worth flagging explicitly, because it is the kind of thing that looks like a '783 proceeding but is not: the '783 prosecution history appears as an exhibit in IPRs directed at its children — e.g., "Prosecution History of U.S. Patent No. 8,204,783" is Ex. 1005 in IPR2022-01319 ([Microsoft Corp.](/litigations/by-plaintiff/Microsoft%20Corp.) v. AlmondNet, Inc., patent 8,671,139) and Ex. 1009 in IPR2022-01324. Being cited as an exhibit is not being challenged. Do not let a docket-scraper or an adversary's spreadsheet conflate the two.

---

Family-adjacent proceedings (context — these are NOT proceedings against '783)

'783 is the middle link of a continuation chain: 7,747,745 → 8,204,783 → {8,200,822, 8,671,139} → {8,959,146, 8,244,582, 7,979,307, 8,244,574, 8,566,164} → … → 11,830,159. Practically every link except '783 has been petitioned.

Sources: AlmondNet v. Oracle, No. 6:24-cv-00303-ADA (W.D. Tex.), Case Readiness Status Report (https://www.docketalarm.com/cases/Texas_Western_District_Court/6--24-cv-00303/AlmondNet_v._Oracle/docs/25.pdf); PTAB decision denying institution in IPR2023-00227 (https://bannerwitcoff.com/wp-content/uploads/2023/06/IPR2023-00227.pdf); Samsung v. AlmondNet, IPR2022-01259 Patent Owner's Response, 2023-06-02 (https://www.docketalarm.com/cases/PTAB/IPR2022-01259/Samsung_Electronics_Co._Ltd._v._AlmondNet_Inc/06-02-2023-Patent_Owner/PO_Response_to_Pet-14-Patent_Owners_Response/); National Law Review, "PTO Litigation Report – June 5, 2017" (CBM2017-00056/-00058 filings).

One structural point that matters most to a '783 defendant. Samsung's IPR2022-01259 petition challenged claims 35–51 of the '822 patent, and the claim-35 language quoted in that proceeding ("A system comprising one or more computers programmed to implement a method of directing electronic advertisements, the method comprising: for each of a multitude of different electronic visitors to a first media property, responsive to receiving information about at least one of a plurality of profile attributes possibly applicable to the visitor, automatically electronically authorizing a third-party second media property to allow display …") reads the same as '783 claim 35 in the patent text I was given. The same is true of the '139 claim 37 language in IPR2022-01319. I could not verify whether the claim sets are verbatim identical (both patents carry 51 claims and the '822 is subject to a terminal disclaimer, which is inconsistent with literal identity), so treat this as strong family resemblance and a source of reusable prior art and claim-construction argument — not as issue preclusion.

---

Strategic summary

Claim status. Nothing about US 8,204,783 is CANCELED. Nothing is SUSTAINED. All 51 claims — independent claims 1 (method), 18 (tangible computer-server device), and 35 (system), plus every dependent claim — are UNTESTED at the PTAB. There is no IPR certificate, no FWD, and no disclaimer narrowing the '783 claim set that I could identify. The patent's Google Patents record shows it live with an anticipated expiration of 2027-06-14, which is the operative date to plan around. Contrast this with sibling '582, where a PTAB FWD of 2024-03-13 found all challenged claims unpatentable (on appeal at Fed. Cir. 2024-1834) and an ex parte reexamination (90/015,285) has rejected all claims.

Estoppel landscape. Because no IPR against '783 ran to a Final Written Decision, no § 315(e)(2) estoppel attaches to '783 at all — for anyone. Every prior-art ground a defendant can find is theoretically available in an IPR against '783, subject only to the § 315(b) one-year bar from service of an infringement complaint and to § 325(d) discretion. Note two traps: (i) § 315(e)(2) estoppel flowing from the '822, '139, '582, '398 or '615 IPRs does not bind the petitioner as to '783, because estoppel is patent- and claim-specific; (ii) conversely, there is no adverse estoppel running against AlmondNet on '783 either — but AlmondNet has made useful-ly-sweeping statements about the family's claims, most notably in CBM2017-00052, where it argued the claims "do not deal with human decision-making or activities" and are "performed automatically by a programmed computer," which Roku later used to argue preambles are limiting. Statements like that are portable ammunition against '783's "automatically electronically authorizing" limitations.

Pattern signals. (1) No petitioner has ever targeted '783 — the closest is Samsung's attack on '822 claims 35–51, which overlap in claim structure. (2) The patent owner, AlmondNet (with affiliate Intent IQ), is a serial, aggressive litigant — the family has been asserted against Amazon, Meta, Microsoft, Samsung, Roku, Oracle, Adobe, Criteo, LinkedIn, Viant and Bytedance/TikTok — and has litigated its PTAB losses up to the Federal Circuit (No. 2024-1834). (3) The '783 patent was asserted against Meta in AlmondNet II, No. 4:23-cv-00438-JST (N.D. Cal.), and then dismissed as one of six patents AlmondNet dropped in the March 2024 joint stipulation that produced a stay of the Meta actions pending the sister-patent IPRs. (4) A defensive aggregator is in the chain at the litigation-monitoring level — the structured data carries Unified Patents litigation-portal links to the W.D. Tex. cases 6:22-cv-01204 (AlmondNet v. Amazon, terminated) and 6:22-cv-01205 (AlmondNet v. Meta, transferred to N.D. Cal. 4:22-cv-08911). I found no evidence Unified Patents or any other aggregator filed an IPR on '783.

Recommended next steps

1. Do not build a defense narrative around an IPR that doesn't exist. If opposing counsel or a budgeting memo refers to "the IPR on 8,204,783," it is wrong. Confirm independently at PTAB E2E / the PTAB decisions search before relying on any proceeding number.
2. The '783 claims were never construed. AlmondNet II (4:23-cv-00438-JST) was stayed before any Markman ruling, and '783 was dismissed rather than construed — so there is no binding or persuasive claim construction on '783 claim 1's "visitor-specific condition," "authorizing … to allow display," or the price-cap limitation in claim 1(b) ("a price charged by the third media property that is less than a profile-attribute-dependent price that an advertiser is willing to pay"). That is your cleanest attack surface, and it is untouched.
3. Reuse the family's IPR record but respect its limits. The Meta FWD in IPR2022-01436 (2024-03-13) cancelling all challenged '582 claims, and the Beyda/Narin and Perry/Wilson/Bronson art sets from IPR2022-01319 and IPR2022-01259, are the obvious starting inventories. Because '783's claims are untested, that art is not estopped — but you must map it to '783's actual limitations, and you must clear § 315(b)'s one-year window from service of your complaint. Obtain the FWD and institution decisions through USPTO PTAB E2E (https://ptacts.uspto.gov) rather than relying on secondary summaries, including mine.
4. If you are already in suit on '783, verify the § 315(b) clock today. Given how aggressively AlmondNet litigates and how the Meta '783 assertion was dropped rather than resolved, a re-assertion against a new defendant is a live possibility, and the one-year bar is the single most common way a meritorious '783 petition dies procedurally — see the § 315(b) denial in IPR2023-00227 as the cautionary example.
5. Treat the "no IPR" fact as a two-sided signal, not a defense. Well-asserted patents usually attract IPRs; the absence here most plausibly reflects that AlmondNet dropped '783 from the Meta case before the family's IPR wave reached it — i.e., it is a timing and case-management artifact, not evidence that the claims are strong.

I'll research the assignment record and litigation posture before drafting.

Good — I have core signals. Let me dig for the recorded assignment details and litigation posture.

I could not reach the USPTO Assignment Center or its indexed mirror directly (both require the interactive search UI, and my web retrieval of that endpoint did not return reel/frame detail). I therefore reconstructed the chain from the patent's own prosecution record, the USPTO/PTAB file histories, district-court dockets, and AlmondNet's public statements, and I flag explicitly below where a reel/frame is unverified rather than inventing one.

Inventors

• Single-inventor portfolio. Every filing in this family (11/763,286; 12/823,709; 13/418,320; 13/491,454; and the later continuations) names Roy Shkedi as sole or first named inventor, and he is listed as the applicant on the filing receipts for 13/491,454 and 15/360,762.
• No inventor-departure pattern. Shkedi founded AlmondNet in 1998, is still CEO, and is Chairman of the subsidiaries Intent IQ and Datonics (IAM interview reproduced in PTAB filings; AlmondNet press release, 2026-04-07). He has therefore never left the assignee — the "all inventors gone within 12 months" tell that often precedes a portfolio sale is absent.
• The one unusual feature is the reverse of a fire-sale: the inventor is also the controlling principal of the assignee, so inventor and owner interests are aligned rather than separated.

Original assignee

AlmondNet, Inc. — named as both original and current assignee on the face of US 8,204,783 (Google Patents). The assignment instrument of record describes the assignee as "AlmondNet, Inc., a Delaware corporation." Address of record on the patent: New York, NY, later Long Island City, NY.

• Line of business: profile-based online advertising. AlmondNet was an operating ad-tech company (explicitly grouped with Tacoda and RevenueScience in the specification's own background section, lines describing "BT companies"), founded 1998 and relocated to the US in 2000. Its subsidiaries Intent IQ (identity resolution) and Datonics (data/audience segments) sell products; AlmondNet described itself in April 2026 as "focused on R&D and the licensing of its extensive portfolio."
• Did it ship a product embodying the claims? Unclear. AlmondNet ran behavioral-targeting ad delivery, which is the field of claim 1. I found no evidence of a currently marketed AlmondNet-branded product reading on the '783 claims specifically; the company's own 2026 description emphasizes "R&D and the licensing of its extensive portfolio," while its subsidiaries market separate identity/data products.
• Current status: operating, privately held, no bankruptcy filing surfaced. It is not publicly traded, so no 10-K/8-K exists to corroborate ownership — a gap for this exercise, not a finding.
• Financial distress, not insolvency: Shkedi testified that AlmondNet faced near financial ruin in 2013 and survived via a take-it-or-leave-it Google licensing deal (IAM, Aug. 2024). No Chapter 7/11, no recorded security interest surfaced.

Assignment timeline

The chain for this patent is a single link. There is no recorded post-issuance transfer, and no recorded security agreement, merger, change of name, or release that I could locate.

• 2012-03-13 (recorded) — Reel/Frame not retrieved; unverified (Assignment Center search UI not reachable; see note below)
  • Conveyance: Assignment of Assignors' Interest ("ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS)," per Google Patents legal events; the underlying instrument is the founder-to-company assignment recited in the 13/491,454 filing materials)
  • Assignor: Shkedi, Roy (individual, sole inventor)
  • Assignee: AlmondNet, Inc. (Delaware corporation)
  • Correspondent: not retrieved from the assignment record itself. All recording and prosecution filings for this family go through USPTO Customer Number 26362 — Louis J. Hoffman, P.C., 14301 North 87th Street, Suite 312, Scottsdale, AZ 85260 (attorneys Louis J. Hoffman and Donald Hertz; the 13/491,454 filing receipt, filed 2012-06-07, lists the filer as "Louis J. Hoffman/Donald Hertz" with the e-mail domain valuablepatents.com). Recurrence flag: this same correspondent appears on every AlmondNet family filing I sampled (13/491,454 filed 2012-06-07; 15/360,762 filed 2016-11-23; the 14/763,963 national-stage filings), so it is the presumptive correspondent on the 2012-03-13 recording.
  • Context: internal founder-to-company assignment, not an acquisition or fire-sale. Note the timing: the instrument was recorded 2012-03-13 — one day after the 13/418,320 continuation was filed (2012-03-12) and about three months before the '783 patent issued (2012-06-19), which suggests it was perfected as part of the March 2012 continuation-filing spree rather than as a standalone transaction.

Verification note: Google Patents legal events for US 8,204,783 list exactly one reassignment entry (2012-03-13, AlmondNet, Inc., assignor Shkedi, Roy) and no subsequent entry through its 2023 descendants. To confirm the reel/frame and correspondent on the face of the recorded document, pull the record at the USPTO Assignment Center (https://assignmentcenter.uspto.gov/) or the indexed mirror (https://assignment.uspto.gov/patent/index.html) using patent number 8204783 or application 12/823,709. I am not supplying a reel/frame number because I did not verify one and will not guess.

Timeline diagram

timeline
    title Ownership of US 8204783
    2006 : Provisional applications filed by Shkedi
    2007 : Parent application filed
         : Shkedi assigns invention to AlmondNet Inc
    2010 : Continuation filed as 12823709
    2012 : Assignment recorded to AlmondNet Inc
         : Patent issues 19 Jun 2012
    2017 : Yahoo CBM petitions denied institution
    2022 : AlmondNet sues Amazon and Meta
         : Samsung and Roku campaigns pending
    2023 : Meta case transferred to N D Cal
         : Amazon West Texas case dismissed
    2024 : Jury verdict against Amazon on related patents
    2026 : Amazon dispute settled with license

NPE / troll-pattern signals

1. Shell-entity transfer — not present. The chain never leaves AlmondNet, Inc. There is no assignment to any "IP / Holdings / Licensing / Ventures" entity, no registered-agent address of record, and no second LLC anywhere on the patent's assignment or legal-events record. The only recorded link (2012-03-13) is founder → operating company.

2. Known asserter in the chain — partially present. AlmondNet does not appear on any of the enumerated classic NPE rosters (Acacia, Marathon, IV, IPNav, Wi-LAN, Conversant/Mosaid, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, Document Generation, Spangenberg). However, it is indexed as a patent asserter in the Stanford NPE Litigation Database (case 4:23-cv-00438, AlmondNet, Inc. et al v. Meta Platforms, Inc., N.D. Cal., filed 2023-01-31, listing patent asserters AlmondNet, Inc. and Intent IQ, LLC, asserted patents including 7,747,745, 8,204,783, 8,775,249, 8,959,146 and 10,715,878), and it is one of the highest-frequency active patent plaintiffs against large operating companies: W.D. Tex. 6:22-cv-01204 (Amazon) and 6:22-cv-01205 (Meta) — the two litigation entries Google Patents flags for this patent — plus N.D. Cal. 4:22-cv-07515 (Samsung), 4:23-cv-00438 (Meta), D. Del. 1:21-cv-01035 / 1:22-cv-01540 (Roku), 1:23-cv-01373 (LinkedIn), 1:23-cv-00174 (Viant), 1:23-cv-00220 (FreeWheel Media), and 6:24-cv-00303 (Oracle). Classified as a hybrid asserter rather than a pure shell NPE.

3. Repeat correspondent across the chain — present, but weak as a troll tell. Louis J. Hoffman, P.C. / Customer No. 26362 recurs across the AlmondNet family (13/491,454 filed 2012-06-07; 15/360,762 filed 2016-11-23; 14/763,963), and is the presumptive correspondent on the sole recorded assignment. This is one attorney serving one stable owner, not the "shell LLCs change, the lawyer doesn't" syndicate pattern the signal is designed to catch. I found no evidence tying Hoffman to a Unified Patents / RPX / Patent Progress NPE assertion list. Separately, the litigation dockets list Adam Hoffman and Reza Mirzaie as plaintiffs' counsel in 6:22-cv-01205; do not conflate these with Louis J. Hoffman, and I have not verified their firm affiliation from a primary source here.

4. Cascading transfers — not present. One recorded link, in 2012, with no successor assignments in the following ~14 years and no chained assignees at all.

5. Pre-litigation transfer — not present. The only assignment predates the first suit naming this patent (W.D. Tex. 6:22-cv-01204, filed 2022-11-18, where the '783 patent was asserted at Compl. ¶46) by roughly 10 years and 8 months. Standing and venue were not arranged by a late-breaking assignment.

6. Bankruptcy fire-sale — not present. No Chapter 7/11 docket, no recorded security agreement, no release-by-secured-party entry. The 2013 near-insolvency described by Shkedi was resolved by a private Google license, not by proceedings, and produced no assignment record.

7. Privateering — not present in the literal sense. Privateering requires an operating company to hand patents to an NPE that asserts on its behalf. Here the patent has never left its operating originator; AlmondNet asserts its own portfolio. If anything the pattern is "operating company that wound down operations and monetizes internally," which is a different (and less clearly NPE) posture.

8. Defensive aggregator — not present. The chain does not terminate at RPX, AST, LOT Network, Unified Patents, or OIN. In fact the patent has been attacked by RPX/Unified-aligned defendants through IPR and CBM (e.g., IPR2022-01420 Microsoft; IPR2022-01436 Meta; IPR2022-01260 Samsung; IPR2022-01455 Amazon; CBM2017-00058 Yahoo! v. AlmondNet) — the inverse of a defensive-aggregation endpoint.

Verdict

NPE — moderate confidence.

The assignment record itself contains no classic NPE mechanics: there is exactly one recorded conveyance (2012-03-13, Shkedi → AlmondNet, Inc., Delaware), no shell-entity hand-off, no cascading LLCs, and no pre-suit transfer — which by itself would point to "insufficient data / original assignee retains." What tips the call is one strong external signal: AlmondNet is the plaintiff in a long-running assertion campaign against Amazon, Meta, Samsung, Roku, Microsoft, Oracle, LinkedIn, Viant and FreeWheel on this family, is indexed as a patent asserter in the Stanford NPE Litigation Database, and describes itself as a portfolio-licensing business (Stanford NPE database, 4:23-cv-00438; W.D. Tex. 6:22-cv-01204 and 6:22-cv-01205, both filed 2022-11-18 with assertion of the '783 patent). It stops short of high confidence because AlmondNet is a genuine 1998-vintage ad-tech developer with operating subsidiaries and a founder-CEO who never sold the patent, so the "no products" and "single-purpose LLC" prongs are not established on this record.

Verify the assignment entries at the USPTO Assignment Center: https://assignmentcenter.uspto.gov/ (search patent 8204783 or application 12/823,709), mirrored at https://assignment.uspto.gov/patent/index.html.

I'll search for the patent's citation records before drawing any conclusions.

Let me pull the actual citation records from the patent's front page and prosecution history.

Bottom line up front

I could not retrieve a verified "References Cited" (§56) listing for US 8,204,783 in this session, and the authoritative full text you supplied does not contain one — the fetched Google Patents record jumps from the bibliography/classifications straight to Definitions, Abstract, Description and the 51 claims, with no (56) section, no Citations table, and no Cited-By table. Rather than reconstruct that list from memory (which would risk fabricating patent numbers and dates), I'm reporting exactly what I verified, what I found fragments of, and what remains unverified. Details and sources below.

---

1. Patent identity (verified from the authoritative text)

Identifier-hygiene note (per your instruction not to auto-correct): the sources disagree on the priority date. The authoritative Google Patents record says 2006-06-16; Unified Patents' portal pages state 2006-06-15 for US-8204783-B2 and for family members. I am reporting both literally and not reconciling them. A separate, unrelated hit — NL 8204783 A (N.V. Philips, 1982, "Werkwijze voor het aanbrengen van een metaallaag op een substraat") and EP 0111957 B1 — is a Dutch application that merely shares the digits "8204783." It is not the US patent and should be excluded.

2. What the text tells us about the subject matter being searched

The record lists the examiner's/algorithmic search concepts as visitor, profile, media property, authorizing, media. Those are the operative claim terms. Claim 1 has two distinct limbs that prior-art mapping must treat separately:

• Limb (a): responsive to receiving profile-attribute information about each of a multitude of visitors to a first media property, automatically electronically authorizing a third-party second media property to allow display of a profile-correlated ad to that visitor when the visitor later visits it, subject to a visitor-specific condition.
• Limb (b): automatically electronically authorizing a third-party third media property to allow display of a profile-correlated ad, for a price charged by the third media property that is less than a profile-attribute-dependent price that an advertiser is willing to pay (i.e., the profitability / price-cap element).

The specification's profit framework (Pr = Rev(profile) − P(mp) − C − Mar; "price cap") is the substance of limb (b). Any §102 analysis must find both limbs in a single reference.

3. Citation evidence I could actually verify

1. Microsoft IPR2022-01319, Exhibit 1005 = "Prosecution History of U.S. Patent No. 8,204,783" (321 pages). This is the single most valuable artifact for your question, because the examiner's actual applied art lives in it. The retrievable snippets confirm:

• A PTO-892 "Notice of References Cited," Part of Paper No. 20110210 (i.e., the Feb. 2011 Office Action in 12/823,709).
• A Form PTO-1449 (Modified) List of Patents and Publications for Information Disclosure Statement, and a search-string record showing "US-PGPUB; USPAT … ADJ … '7747745'".
• Bibliographic header: Serial No. 12/823,709, Applicant Shkedi, Art Unit 2441, Examiner Kostas J. Katsikis, Conf. No. 8426.
• The Nov. 21, 2011 final Office Action and the Feb. 21, 2012 after-final response in which the claim set was rewritten into the issued 51 claims.
  Source: https://www.docketalarm.com/cases/PTAB/IPR2022-01319/Microsoft_Corporation/docs/07-27-2022-Petitioner/Exhibit-1005-5-Prosecution_History_of_US_Patent_No_8,204,783.pdf
  I could not open the pages of Exhibit 1005 that contain the PTO-892's actual reference table, so I cannot state its contents.

1. A PTO-892 (Date Considered 02/05/2017) from the related prosecution of 15/360,762 (same family, Art Unit 2441, Examiner Kostas Katsikis) cites a long run of Shkedi's own patents/publications — e.g., items 190–210 including US 8,600,815; 8,607,267; 8,677,398; 8,683,502; 8,695,032; 8,713,600; 8,775,249; 8,959,146; 9,071,886; 9,078,035; 9,083,853; 9,208,514; 9,226,019; 9,271,024; 9,351,053; 9,369,779; plus publications 2005/0086112 (Shkedi), 2009/0049468 (Shkedi), 2009/0172723 (Shkedi et al.). Source: https://www.docketalarm.com/cases/PTAB/IPR2023-00384/Amazon.com_Inc/docs/12-22-2022-Petitioner/Exhibit-1013-...Prosecution_history_of_US_Patent_9,830,615.pdf
   Caveat: these are AlmondNet's own later family members, cited in a later application. They share the 2006-06-16 priority date and therefore are not §102 prior art against US 8,204,783.

2. The sibling patent US 8,200,822 B1 (same specification, filed 2012-03-12, same family) has a published "Patent Art (131)" list on its Unified Patents page, including US 2009/0049468 A1, US 2003/0065558 A1, US 2011/0040607 A1, US 6,073,241 A, US 7,181,438 B1, US 5,710,884 A, US 2007/0294401 A1, US 6,324,519 B1, US 2002/0082923 A1, US 2003/0220837 A1, US 6,487,538 B1, US 2003/0149938 A1, US 6,366,298 B1, US 5,794,210 A, and others. Source: https://portal.unifiedpatents.com/patents/patent/US-8200822-B1
   Caveat: this is the sibling's record, not 8,204,783's, and the list mixes backward and forward citations (some entries have priority dates after 2006-06-16, e.g., 2007-04-16 and 2009-08-04, which can only be forward citations). I therefore cannot treat it as 8,204,783's §56 list.

3. Non-patent literature (105 items) appears in the AlmondNet-family Unified Patents record, including the DoubleClick "Boomerang" archive (2000), Commercialization of the World Wide Web: The Role of Cookies (1997), RFC 2109 HTTP State Management (Feb. 1997), NetGravity AdServer 3.0 (1998), Engage/ECHO remarketing (Sep. 25, 2000), and Yahoo's invalidity contentions (Jun. 30, 2017). Source: https://portal.unifiedpatents.com/patents/patent/US-10628857-B2. Again — family-level, not confirmed as 8,204,783's list.

4. Candidate references and their §102 exposure — stated with confidence levels

Below, bibliographic data I state comes from the retrieved records (marked ✔) or is flagged as unverified. Under pre-AIA §102 (which governs here, since the effective filing date derives from a June 2007 non-provisional), single-reference anticipation requires every element of the claim.

Do not treat the following as §102 art against 8,204,783, even though they appear in related records: US 7,747,745; 8,200,822; 8,671,139; 8,959,146; 8,341,247; 8,501,444; and publications 2007/0294401 (priority 2006-06-18 — two days after) and 2009/0049468 (2007-04-16). These are same-family or later-filed AlmondNet/Intent IQ matters.

5. Why I am not giving you a finished §102 chart

1. The claim set that issued was created during prosecution. The examiner's PTO-892 (Paper 20110210, Feb. 2011) applied art to the original claims; the applicant replaced them with the 51 issued claims in the Feb. 21, 2012 after-final response. Mapping the 2011-cited references onto the issued claim language — particularly the "visitor-specific condition" and "price … less than a profile-attribute-dependent price that an advertiser is willing to pay" limitations — is exactly the work the Microsoft IPR2022-01319 petition and the Yahoo invalidity contentions purport to do, and I could not open those substantive pages.
2. Anticipation requires a single reference disclosing every element. The Merriman-family and Rosenberg references cover the re-targeting/tagging architecture but nothing about a media-property-charged price benchmarked against advertiser willingness-to-pay; the Shkedi mercantile patents cover the pricing economics but not the third-party cross-property authorization-with-condition structure. On the record I can see, no single reference appears to disclose all of claim 1 — which is consistent with the claims having issued over the art of record, and would push the challenge toward §103 combinations rather than §102.

6. Where to get the authoritative (56) list

• USPTO PatentCenter / Global Dossier for application 12/823,709 — the PTO-892 and PTO/SB/08 IDS forms are in the file wrapper, unredacted. https://patentcenter.uspto.gov/
• Microsoft Exhibit 1005 in IPR2022-01319 — 321 pages of the 8,204,783 prosecution history, including PTO-892 forms. https://www.docketalarm.com/cases/PTAB/IPR2022-01319/Microsoft_Corporation/docs/07-27-2022-Petitioner/Exhibit-1005-5-Prosecution_History_of_US_Patent_No_8,204,783.pdf
• The IPR2022-01319 petition itself (Microsoft Corp. v. AlmondNet) — this will name the specific primary and secondary references with paragraph-level pin cites and, by construction, is the best single statement of "most relevant prior art" for this patent.
• Google Patents citations tab (https://patents.google.com/patent/US8204783B2/en) — the page did not render its citation tables to my fetcher; viewing it directly should expose "Citations" and "Cited By" separately. Note that "Cited By" lists forward citations, which are not prior art; only "Citations" (backward) matters for your §102 question.

Explicit uncertainty disclosure: I have not verified the contents of any PTO-892 or IDS actually naming references for 12/823,709, and I have not verified that the specific patents listed in §4 above appear on 8,204,783's face. The entries marked ✔ are corroborated by at least one retrieved source; the others reflect bibliographic knowledge I could not cross-check in this session and should be confirmed before use in any filing.

I'll start by locating the actual prior-art references cited in this family, since the supplied Google Patents page text does not include a "Prior Art" citation list.

The supplied page contains no prior-art citation list, so I'll retrieve the actual cited references and the invalidity contentions/PTAB petitions directed at this family.

Good leads. The family's own prosecution history cites O'Kelley in view of Eldering as §103 art. Let me pin down those references and the invalidity-contention art for the sibling '822 patent (same specification).

Preliminary note on the "Prior Art section"

The supplied Google Patents page contains no Prior Art / "References Cited" section. The capture includes bibliographic data, classifications, "Definitions," landscape, abstract, description, claims (1–51), and the family/priority tables — but no front-page U.S. patent-document list, no foreign references, and no non-patent-literature list. I therefore could not "use the results from the Prior Art section" literally, and instead substituted the two authoritative §103 records that exist for this exact specification:

1. The examiner's own §103 rejection in the direct parent, US 7,747,745 (same specification, same priority), preserved as Exhibit 1006, "Prosecution History of U.S. Patent No. 7,747,745," in Microsoft's IPR2022‑01319 — rejecting claims over O'Kelley in view of Eldering (and, for claims 37–56, further in view of Boubek and Bandura).
   • https://www.docketalarm.com/cases/PTAB/IPR2022-01319/Microsoft_Corporation/docs/07-27-2022-Petitioner/Exhibit-1006-6-Prosecution_History_of_US_Patent_No_7,747,745.pdf
   • https://www.docketalarm.com/cases/PTAB/CBM2017-00058/Yahoo%21_Inc._v._ALMONDNET_INC/docs/09-13-2017-Patent_Owner/Exhibit-2002-Exhibit_2002___Remainder_of_file_history___Serial_No_15360762.pdf
2. Defendant-side prior-art contentions against the sibling patent US 8,200,822 (identical title and specification to the '783) — "Exhibit G — State of the Art References for Asserted Patents," which lists, inter alia, U.S. Patent No. 5,948,061 (Merriman et al.), 9/7/1999.
   • https://ptacts.uspto.gov/ptacts/public-informations/petitions/[1555815](/patent/1555815)/download-documents

I flag explicitly: the '783's own front-page reference list is unknown to me in this session, and I have not reconstructed it. Where below I name specific reference numbers, they come from the searchable record; where the record gives only a surname (e.g., "O'Kelley," "Boubek," "Bandura"), I say so rather than inventing numbers.

---

1. Legal framework and the critical date (this drives everything)

• Pre‑AIA applies. The examiner's actions in this family state the applications are "being examined under the pre-AIA first to invent provisions" (15/360,762 Office Action). The '783 is a continuation of 11/763,286 filed June 14, 2007, now US 7,747,745, which claims benefit of provisionals 60/804,937 (June 16, 2006) and 60/805,119 (June 19, 2006).
• Effective filing date for the '783 claims = June 16, 2006, provided the granted claims are supported under §112 by the '745/parents. Its own filing date is June 25, 2010; a challenger who could show the two-"authorizing"-step claims lack written-description support in the June 2006 provisionals would move the date to 2010 — a materially different art landscape. I have not verified support; treat this as an open issue.
• Therefore:
  • §102(b) art must predate ≈ June 14–16, 2005.
  • Art published/patented June 2005 – June 2006 is available only as §102(a)/(e) art.
  • Art from June 2006 – June 2007 (e.g., the family IDS's US 2006/0271690, Banga et al., published Nov. 30, 2006) is not §102(b) art and is usable only on an earlier-filing §102(e) theory. I flag this because it is a common error in this art.
• Status: "Active," anticipated expiration listed as 2027‑06‑14; terminal disclaimers tie it to the earlier-issued family members (the 2017 terminal disclaimer lists 9508089, 9208514, 8959146, 8671139, 8204783, 8200822, 7747745).
• Enforceability posture: the '783 was asserted against Meta (N.D. Cal., No. 4:23‑cv‑00438, filed Jan. 31, 2023, per Stanford NPE database), alongside 10715878, 7747745, 7979307, 8775249, 8959146. Burden: clear and convincing in court (the W.D. Tex. jury charge in the Amazon case confirms this standard), preponderance at the PTAB.

(Minor housekeeping: this task says "current date April 26, 2026," while the patent-text capture is timestamped 2026‑09‑28. I use the April 2026 date for status statements, consistent with the earlier section.)

---

2. POSITA and claim construction

POSITA — I adopt the formulation already adopted by the Board in an AlmondNet-family IPR (IPR2022‑01259, involving US 8,671,139): "a bachelor's degree in electrical engineering or computer science, and two years of work experience in the field of network-based computer systems, such as systems for sending and receiving information over local networks and wide area networks (such as the Internet or World Wide Web)," with the caveat that more formal education and less experience, or vice versa, could suffice. This is the correct yardstick because the '783 is a networking/server-programming patent dressed in advertising economics.

Constructions that matter for §103:

---

3. What claim 1 actually requires (element map)

The heart of the case is C + E: cross-site profile-based ad display gated by a profitability comparison.

---

4. Prior-art references I can ground in the record

Not prior art, despite appearing in the family record: US 8,015,184 (Zito, 2011), US 8,880,677 (Julia, 2014), US 2006/0271690 (Banga, 11/30/2006, post-priority) — these appear as IDS/co-pending citations, not as §103 art against a June 2006 priority.

---

5. Grounds of obviousness

Ground 1 — Merriman ('061) in view of Gerace ('396) and Montulli ('670) [primary recommended ground]

Coverage. Merriman discloses the networked advertising server that selects and delivers ads to users of affiliated sites, targets on an observed user profile/behavior, and logs/tracks the user — elements A, B, C (in the "later visit" sense of an ad server recognizing a user on another affiliate site), and the "allowing display by another media property" reading of "authorizing." Gerace supplies the cookie-based behavioral profile built from a user's navigation across sites (B). Montulli supplies the persistent-cookie mechanism by which a first server's tag is written to and read from the client (claims 9, 13, 14, 17).

Motivation (KSR/MPEP 2143 rationales):

• Same field of endeavor, common problem. All three address delivering advertisements to networked users based on who they are; combination is not a field-crossing leap (KSR rationales (A), (C)).
• Predictable result / known technique. Using a persistent cookie to mark a user so a different server recognizes that user later was expressly admitted prior art by the applicant (DoubleClick Boomerang; Montulli). Applying it to Merriman's ad server yields nothing more than the expected operation of the known technique.
• Design incentive. The specification itself states the problem to be solved — the cost of ad space at a second property "might not be covered by revenues generated for a BT company buying the space" — a classic market-forces motivation (KSR rationale (F)) to add a profitability test to the known architecture.

Weak point: none of Merriman, Gerace or Montulli is shown to disclose element E (the price-vs-advertiser-willingness-to-pay comparison). This ground therefore needs a fourth reference — see Ground 2.

---

Ground 2 — Ground 1 further in view of a venue-level ad selection/pricing reference (the examiner's O'Kelley) and/or the disclosed profit equation

Coverage of E. The '783 specification itself reduces the limitation to arithmetic: "Pr = Rev(profile) − P(mp)," and "positive profit … equates to testing … whether P(mp) < Rev(profile)," which the specification literally calls "the same as testing whether the price charged by a media property is less than a price cap." O'Kelley (per the examiner's and applicant's own characterizations) selects an ad for a given electronic venue — i.e., it operates on the venue/price side of the equation.

Motivation. "[I]f a technique has been used to improve one device, and a person of ordinary skill in the art would recognize that it would improve similar devices in the same way, using the technique is obvious" (KSR). Comparing an asking price to an expected yield to decide whether to buy is the definition of routine commercial practice; implementing it as a server-side "authorize if price < expected revenue" test requires no more than ordinary programming. The claimed "profile-attribute-dependent price" is merely the ad price that the ad-serving art already determines per-profile.

Why this ground is the strongest on paper: it is the examiner's own rejection, applied to the same specification in the '745 (Exhibit 1006), and re-applied in the 15/360,762 family member.

Why it is contestable — see §6.

---

Ground 3 — Merriman/Boomerang admissions + an ad-network inventory/yield reference (e.g., Merriman '599-type art) + Montulli

For claim 1(b)'s "third media property … for a price" and dependent claims 12 ("ad space … only temporarily") and 11 (different medium), the record shows ad-space allocation/intermediation art existed: the '783 specification itself describes the ad-network model where a site gives unsold inventory to an ad network that "will have temporary control of the ad space." That is an admission and it supplies the "temporary control" and "third media property" concepts directly. Combining it with the profile/tag art of Ground 1 is motivated because the entire commercial purpose of the BT model is to buy someone else's inventory and resell profile-targeted impressions at a margin.

Note: I could not, in this session, confirm a specific, pre-June-2006 inventory-optimization patent number (the Microsoft "Exhibit G" list is the right place to look, and it confirms Merriman '061 but I could not read the full list). This ground should be treated as directionally sound but not fully cite-supported.

---

Ground 4 — O'Kelley + Eldering + DoubleClick-Boomerang admission (the "later visit" repair)

The strongest available expectation for AlmondNet's rebuttal is the one it actually used: that O'Kelley selects an ad "for immediate presentation," so it does not teach "authorizing … to allow display … when the visitor visits … at a time after the visitor visits the first media property." The Boomerang admission (and Merriman/Gerace/Montulli) directly repairs exactly that gap, and the motivation is supplied by the applicant's own stated problem. Under KSR, the fact that O'Kelley's reference is being modified to add a later-visit recognition step is not a "change in principle of operation" so long as the resulting system still does what O'Kelley does (select/price ads) — which it does. AlmondNet's In re Ratti argument is the single most vulnerable plank in its earlier win, and a challenger that supplies an explicit motivation statement (the spec's own efficiency/profitability problem) can defeat it.

---

Ground 5 — Mr. Shkedi's own later admissions and family consistency

Dependent claims 2–3 (look-up data structure mapping profile attributes → media properties) and claims 10/15 (time-limited authorization) are squarely met by the look-up-table embodiment the '783 specification itself describes as an implementation option ("A look-up table can be used that lists, based on the kind of profile, the media properties that should be selected"), and by standard profile-decay reasoning (the spec: "some kind of profiles … are time-sensitive"). A reference combination need only supply what the claims add beyond the disclosed-but-claimed alternatives.

---

6. Anticipation and the applicant's likely rebuttals

§102 anticipation. I do not see a single-reference anticipation of claim 1 on the record I retrieved. Merriman/Boomerang lacks E; O'Kelley reportedly lacks the later-visit architecture. Any §103 case must be a combination case.

Rebuttals AlmondNet has actually made (CBM2017‑00058 file history, Ex. 2002), and how they fare:

1. "Neither O'Kelley nor Eldering disclose arranging tagging by a computer controlling ad space on unrelated (third party) media properties … conditioned on the ad price being less than a price cap, during a later visit." — This is a genuine gap-filling argument. It is answerable by Grounds 1/4 (Merriman/Gerace/Montulli + Boomerang admission) but the challenger must affirmatively map the "third-party" requirement, because the Board has already construed "third-party" to require a distinct entity in this family (IPR2022‑01259). If the combination's "media property" and "authorizing server" collapse into one entity, claim 1 fails to read on it.
2. "Modification … would fundamentally alter O'Kelley's operating principles" (In re Ratti/MPEP 2143.01(VI)). — Rebuttable with an explicit rationale tied to the spec's stated problem.
3. Enablement/§112 and double-patenting history — the family repeatedly used terminal disclaimers to clear obviousness-type double patenting across the seven parents. That history is not §103 art, but it does show the claims were understood as obvious variants of one another, which a challenger can use rhetorically.

Structural features that make claim 1 harder to invalidate than it looks:

• The two distinct authorizations with two distinct gates (visitor-specific condition in (a); price comparison in (b)) — a challenger must show both, and must articulate why one reference's gate supplies the other.
• The claim requires "for each of a multitude of different electronic visitors," i.e., automated, visitor-by-visitor decisions. Art showing batch/aggregate deal-making is insufficient.
• §101 overlay: AlmondNet's own file history shows the examiner ultimately accepted that the claims are not "the basic abstract idea of using advertising as an exchange or currency" and that the "price cap" language is not "paramount." That does not defeat a §103 ground, but it means the challenger cannot shortcut with a §101-flavored obviousness argument ("mere automation of a fundamental economic practice") — the Board has previously treated that framing skeptically in this family (cf. the Fintiv-stage POPR discussion in the 1559018 petition: "the Petition fails to allege that any limitations of the independent claims are disclosed by the prior art; instead, the Petition relies on expert testimony to satisfy missing claim limitations").

---

7. Objective indicia (secondary considerations)

• Licensing/commercial success: the family's commercial significance is documented (AlmondNet's Google licensing deal described in the IAM interview; the 2024 Amazon verdict of $122M on different family members, 7,822,639 and 8,671,139). No nexus to the '783 claims is shown in anything I retrieved. A challenger should argue no nexus; AlmondNet must show the success is attributable to the claimed two-gate authorization architecture specifically.
• Long-felt need / failure of others: arguable, and the specification frames the need well ("cost of ad space at a second media property might not be covered by revenues … Alternatively, the revenues might cover the media cost but not deliver satisfactory margins"). But the need is a business/financial need, which cuts both ways — it is exactly the kind of market pressure that supplies a KSR motivation.
• Copying: none evidenced in the retrieved record.

---

8. Bottom line

Most likely to succeed: Grounds 2 and 4, framed as Merriman + Gerace + Montulli + O'Kelley, or O'Kelley + Eldering + the applicant's admitted Boomerang/cookie matching art, with an explicit KSR motivation statement rooted in (i) the specification's own statement of the profitability problem, (ii) the admitted known technique of domain-readable cookie tagging across properties (claims 9, 13–17), and (iii) the admitted ad-network practice of temporary ad-space control (claim 12).

Most likely to fail: any attempt to invalidate independent claims 1/18/35 without an affirmative showing on the "third-party … media property" distinctness requirement and on the second, price-based gate — because the family's own record shows a prior art rejection was overcome precisely on those points, and the Board has already construed "third-party" narrowly in a sibling patent.

---

9. Explicit uncertainty statements

1. The '783's own "Prior Art" list is not in the supplied text, and I could not retrieve it. My reference inventory is drawn from the family's prosecution/IDS/CBM/IPR record and from defendant contentions against the sibling '822, not from the '783 front page.
2. I could not recover patent numbers, dates, or titles for the "O'Kelley," "Eldering," "Boubek," and "Bandura" references, which the examiner used in the '745 parent prosecution. I have described them only as the applicant's own remarks characterize them. If "O'Kelley" is the Right Media/AppNexus-associated O'Kelley reference, its pre-June-2006 availability must be confirmed before use.
3. Baig, Laidlaw, Sitaraman, Hahn, and Xu (IPR2024‑00709 grounds against US 10,715,878) are not established here as prior art to the '783; their dates relative to June 16, 2006 are unverified, and the '878 claims differ materially from the '783 claims.
4. Priority-date vulnerability is unresolved. I have not verified §112 support in provisionals 60/804,937 / 60/805,119 for the granted two-authorization claims. If unsupported, the operative date becomes June 25, 2010, which would open a much larger art field and materially strengthen every ground above.
5. No 2026 Federal Circuit or PTAB outcome on the '783 itself appears in the record retrieved (consistent with the earlier section: the April 13, 2026 Rule 36 judgment involved US 8,244,582 in IPR2022‑01436, and the dismissed appeal 26‑1326 involved 8,671,139 and 7,822,639).
6. I am not offering any opinion on infringement or on §101, both of which are outside the §103 question asked.