US 9270720

I'll search for authoritative information on this specific patent number, including any 2026 Federal Circuit activity.

Let me dig into the specific IPR and any 2026 Federal Circuit docket activity on this patent.

Let me verify the claim set structure and check for any 2026 activity specific to the '720 patent.

US Patent 9,270,720 (the '720 patent)

Caveat on sources: I could not query USPTO PatentCenter or the CAFC docket system directly with these tools. The summary below is grounded in the patent text supplied, the Google Patents record for US 9,270,720, PTAB filings/decisions for IPR2020-00647, and the Federal Circuit opinion in Netflix, Inc. v. DivX, LLC, Nos. 2022-1203/2022-1204 (Fed. Cir. Oct. 25, 2023). Items I could not verify authoritatively are flagged.

Bibliographic data

Note: the Google Patents "anticipated expiration" of 2031-12-30 appears to run from the Dec. 30, 2011 parent filing and does not obviously fold in the recorded 36-month PTA; the effective term could therefore be later. Treat the expiration date as uncertain.

Abstract (verbatim, as published)

"Systems and methods for automatically generating top level index files for use in adaptive bitrate streaming in accordance with embodiments of the invention are disclosed. One embodiment of the method of the invention includes receiving a request from a playback device at a playback server, where the request identifies a piece of content, retrieving a list of assets associated with the identified piece of content using the playback server, filtering the list of assets using at least one predetermined criterion using the playback server, generating a top level index file describing each asset in the filtered list of assets using the playback server, and sending the top level index file to the playback device using the playback server."

Independent claims — plain-language overview

The issued claim 1 is materially narrower than the "summary" version quoted in the specification; it was amended during prosecution to add product-identifier/device-capability limitations (a Preliminary Amendment was filed Apr. 3, 2015) and terminal disclaimer (Sept. 28, 2015).

Claim 1 — Method of generating a top level index file (verbatim, as recited in the Federal Circuit opinion):

1. A method of generating a top level index file, comprising:
   • receiving a request from a playback device at a playback server system, where the request (i) identifies a piece of content and (ii) includes a product identifier;
   • retrieving, using the playback server system, (i) a list of assets associated with the identified piece of content and (ii) at least one device capability based upon the product identifier, wherein each asset is a different stream associated with the piece of content;
   • filtering the list of assets using the at least one device capability using the playback server system, wherein the playback server system maintains a database of product identifiers and associated device capabilities;
   • generating a top level index file describing each asset in the filtered list of assets using the playback server system; and
   • sending the top level index file to the playback device using the playback server system, wherein the top level index file is used by the playback device to determine which assets to request for playback on the device.

In plain language: a server, when a device asks for a title and identifies itself by product ID, looks up that device's capabilities, pulls the catalog of streams ("assets") for the title, removes the ones the device can't/shouldn't play, builds a custom playlist/index file from what's left, and returns it so the device can pick streams for adaptive streaming.

Claim 13 — Playback server system (verbatim, partial, as quoted in litigation record):
13. A playback server system, comprising: a database that stores descriptions of assets associated with specific pieces of content; a database that stores a plurality of product identifiers and associated device capabilities; a processor configured using a playback management application; wherein the playback management application configures the processor to: receive a request from a playback device, where the request (i) identifies a piece of content and (ii) includes a product identifier; retrieve (i) a list of assets associated with the identified piece of content and (ii) at least one device capability based upon the product identifier, wherein each asset is a different stream associated with the piece of content; filter the list of assets using the at least one device capability; generate a top level index file describing each asset in the filtered list of assets; …

In plain language: the apparatus counterpart of claim 1 — a server with two databases (asset descriptions per title; product IDs ↔ device capabilities) and a processor programmed by a "playback management application" to do the same receive → retrieve → filter → generate sequence. My quoted text is truncated at the "generate" step; the claim almost certainly also recites sending the index to the device, but I do not have the full verbatim remainder.

Possible third independent claim (machine-readable medium): The specification's summary recites "a machine readable medium containing processor instructions" that perform the same receive/retrieve/filter/generate/send process. I could not authoritatively confirm a corresponding issued independent claim in the '720 patent or verify the total claim count. Related litigation materials state that claims 2–12 depend from claim 1, and claim 13 is an independent system claim; the IPR challenged claims 1–5 and 12. Treat the exact claim set/total count as unverified.

Litigation and PTAB history (specific to the '720 patent)

• District court: DivX asserted the '720 patent against Netflix in the Central District of California (cases referenced as 2:19-cv-01602 and 2:19-cv-01606).
• IPR2020-00647 — Hulu, LLC and Netflix, Inc. v. DivX, LLC: petition filed Mar. 11, 2020 (Patent 9,270,720, app. 14/336,392); institution granted Oct. 1, 2020; Final Written Decision Sept. 27, 2021 determining no challenged claims unpatentable (Netflix's Pyle+Marusi and Lewis+Marusi obviousness theories rejected).
• Federal Circuit: Netflix, Inc. v. DivX, LLC, Nos. 2022-1203, 2022-1204 (Fed. Cir. Oct. 25, 2023) (precedential; reported at 84 F.4th 1371) — affirmed the Board, holding Netflix forfeited arguments not clearly presented in its petitions ("the Board should [not] have to decode a petition to locate additional arguments beyond the ones clearly made").
• Google Patents also lists CAFC case 22-1203 in the '720 family litigation table, consistent with the above.

2026 CAFC docket check

I found no 2026 Federal Circuit docket or decision specific to US 9,270,720. The 2026 appellate activity in the DivX–Netflix dispute concerns a different patent:

• Netflix, Inc. v. DivX, LLC, No. 24-1541 (Fed. Cir. Feb. 13, 2026) (precedential; argued Jan. 12, 2026) — a § 10,225,588 ('588 patent) IPR appeal (IPR2020-00558). The Court reversed the PTAB's construction, holding that "within the requested portions of the selected stream" modifies "encrypted portions of frames of video," not "encryption information," vacated, and remanded. Secondary coverage reports that, after this reversal, DivX gave Netflix a covenant rather than continue asserting infringement of the '588 patent.

Because the '588 patent shares a technical family with the '720 patent (both DivX adaptive-bitrate-streaming patents; the '720 specification incorporates by reference the Kiefer application on protected alternative streams), the Feb. 2026 decision is worth noting — but it is not a ruling on the '720 patent, and I did not find any 2026 CAFC filing for the '720 patent itself.

Explicit uncertainties

1. Exact total claim count of the '720 patent and the precise text of any machine-readable-medium independent claim — not verified.
2. Full verbatim text of claim 13's final "send" limitation — not verified (quote is truncated in the source I found).
3. Effective expiration date (interaction of the 36-month PTA with the 2031-12-30 figure) — not verified.
4. Current ownership nuances (Sonic IP vs. DivX, LLC as "current assignee") — Google Patents lists DivX, LLC, but Google expressly disclaims accuracy of its assignee list.

I'll search for litigation involving US patent 9270720.

Let me dig deeper into the specific cases and their details.

Litigation involving U.S. Patent No. 9,270,720

Scope note: I searched litigation-facing sources (Unified Patents PTAB/litigation portal links, Docket Alarm, CourtListener, Justia, the CAFC opinions site, and PTAB filings). Below I list every proceeding tied to the literal patent number 9,270,720 that I could confirm. I have excluded results for similar-but-different numbers (e.g., Parabit’s U.S. 8,523,072).

Date caveat: You stated the current date is April 26, 2026. Several of the authoritative sources returned describe events dated after that (a jury verdict of March 25, 2026, a final judgment dated May 7, 2026, and a Federal Circuit opinion of February 13, 2026 in a different DivX patent case). I report them as written in the sources rather than reconciling them, and I flag anything I could not independently confirm.

1. DivX, LLC v. Netflix, Inc. — C.D. Cal.

• DivX’s infringement contentions (filed as an exhibit in the related IPR) assert claims 1–5 and 12 of the ’720 patent against Netflix’s streaming system (e.g., Open Connect Appliances, “MSL” manifest requests).
• The case was stayed for several years pending the IPRs (WilmerHale states the stay was lifted May 2024).
• Jury trial commenced March 13, 2026. On March 25, 2026, the jury returned a unanimous verdict that Netflix did not infringe the asserted claims of U.S. Patent Nos. 7,295,673, 8,139,651, 8,472,792 and 10,212,486 (the four patents actually tried). The ’720 patent was not among the four tried to verdict because its asserted claims (1–2, 4, 5) had already been held invalid under § 101 on summary judgment; claim 1 of U.S. 9,998,515 was likewise held invalid under § 101.
• Final judgment document in the docket (No. 2:19-cv-1602-GW-AGRx) references the verdict and the § 101 rulings, dated in the document as May 7, 2026.

2. DivX, LLC v. Hulu, LLC (“Hulu I”) — C.D. Cal.

Note: I did not find a separate verdict or judgment on the ’720 patent in the Hulu action; the case was dismissed before trial.

3. Netflix, Inc. and Hulu, LLC v. DivX, LLC — PTAB IPR (post-grant, not district-court litigation)

4. Netflix, Inc. v. DivX, LLC — Federal Circuit

Google Patents also lists a Federal Circuit case 22-1203 and the PTAB case IPR2020-00647 under the patent’s litigation tags, consistent with the above.

Summary of outcomes / current status

• District court: The ’720 patent’s asserted claims (1–2, 4, 5) were held invalid for lack of patentable subject matter under § 101 by summary judgment in DivX v. Netflix (C.D. Cal. 2:19-cv-01602); a final judgment reflects that and the March 25, 2026 defense verdict on the other asserted patents. Claim 12 was asserted but is not listed among the claims invalidated in that § 101 ruling as reported.
• Hulu action: Dismissed Aug. 25, 2022; no ’720 merits judgment found.
• PTAB: All challenged claims survived IPR2020-00647 (FWD Sept. 27, 2021).
• Federal Circuit: Affirmed the PTAB on Oct. 25, 2023 (84 F.4th 1371).
• Patent status as listed by Google Patents: “Active,” with anticipated expiration 2031-12-30. (The PTO has not canceled the claims; the § 101 invalidity is a district-court ruling and, if appealed, would be reviewed de novo.)

Caveats

• I could not verify the exact complaint filing dates for 2:19-cv-01602 or 2:19-cv-01606 from the sources retrieved (both are 2019 filings).
• I found no other litigation (no ITC action, no other district-court suits, and no other IPR tied to the literal number 9,270,720) beyond the Netflix and Hulu matters above.
• Because your stated “today” (April 26, 2026) precedes several events appearing in the sources (jury verdict, final judgment, and the Feb. 13, 2026 CAFC opinion in Netflix v. DivX concerning the ’588 patent — a different patent), please treat those later-dated items as the sources’ own representations. I have not verified whether a post-judgment appeal concerning the ’720 patent is pending; that would require checking the C.D. Cal. docket and the Federal Circuit docket directly.

Primary sources: CAFC opinion 22-1203 (10-25-2023); PTAB IPR2020-00647 (FWD, Paper 28, Sept. 27, 2021); Docket Alarm PTAB/IPR record for IPR2020-00647; Justia docket and joint status report for DivX, LLC v. Netflix, Inc., No. 2:19-cv-01602 (C.D. Cal.); Hulu’s motion-to-stay memo in No. 2:19-cv-01606; WilmerHale press note (April 8, 2026).

I'll research the PTAB activity on this patent, starting with the IPR referenced in the patent page itself.

Strong signals. Let me pin down the FWD details, the panel, and any companion IPRs on the '720 patent.

Let me verify whether any other IPR targeted the '720 itself and confirm the '515 companion case mapping.

Important sourcing note

The structured "PTAB proceedings on file" block (USPTO ODP API) reports no AIA trial proceedings for US 9,270,720. That is incorrect/incomplete. The patent's own Google Patents "Family has litigation" field discloses a PTAB case — "PTAB case IPR2020-00647 filed (Final Written Decision)" — and web research confirms a full IPR that went to a Final Written Decision and a precedential Federal Circuit affirmance. I flag this below as the one confirmed AIA trial on this patent, and mark clearly where I could not independently verify something.

Proceedings overview

One AIA trial has been filed against US 9,270,720 — IPR2020-00647 (Hulu, LLC & Netflix, Inc. v. DivX, LLC), claims sustained: the Board challenged-and-instated claims 1–5 and 12, then issued an FWD holding no challenged claim unpatentable, and the Federal Circuit affirmed on 2023-10-25; the remaining claims were never tested. Bottom line for a defendant: this patent is hardened by a completed, affirmance-backed IPR — the obvious §103 attacks (Pyle+Marusi and Lewis+Marusi) failed on the merits, so do not expect a quick IPR kill; carve around the surviving claims or attack on grounds the PTAB never reached.

(Note: the companion IPR2020-00648 covers U.S. Pat. 9,998,515 — a continuation of the '720 — not the '720 itself; it is referenced only as context and should not be counted as a proceeding "on" 9270720.)

IPR2020-00647 — Hulu, LLC and Netflix, Inc. v. DivX, LLC

• Type: Inter Partes Review
• Filed: 2020-03-11
• Status: Final Written Decision — no challenged claims unpatentable; Fed. Cir. affirmed (2023-10-25). Plain-English gloss: Netflix/Hulu lost outright at the Board and lost again on appeal — every claim they attacked (1–5, 12) survived. (The structured ODP block shows this proceeding as "none," which is a data gap, not an absence.)
• Judge panel: Bart A. Gerstenblith (writing), Monica S. Ullagaddi, Iftikhar Ahmed — Administrative Patent Judges.
• Petition grounds: Claims 1–5 and 12 challenged under 35 U.S.C. § 103(a) on two grounds:
  • Ground I — Pyle (U.S. Pat. 8,782,268) in view of Marusi (EP 2 180 664)
  • Ground II — Lewis (U.S. Pub. 2012/0047542) in view of Marusi
  • Supported by the Declaration of Dr. Clifford Reader (Ex. 1003); Patent Owner relied on Dr. Kenneth A. Zeger (Ex. 2010).
• Institution decision: Instituted on all challenged claims and both grounds, 2020-10-01 (Paper 9). The Board credited Petitioner's contention that a POSITA would understand Pyle to teach generating new manifest files by "filtering the available list of assets to include only a subset of assets based upon the product identifier information."
• Final Written Decision: 2021-09-27, Paper 28 (JUDGMENT, "Final Written Decision Determining No Challenged Claims Unpatentable," 35 U.S.C. § 318(a)). Verdict at claim level: claims 1, 2, 3, 4, 5, and 12 — all of the challenged claims — held NOT unpatentable; none canceled. The Board found Petitioner failed to prove unpatentability by a preponderance. Key reasoning:
  • The Board identified a "disconnect" in Petitioner's use of Pyle. Because the "generating" limitation creates a top level index file describing "each asset in the filtered list of assets," the index content is necessarily the byproduct of the "filtering" step. Petitioner used Pyle's selection of a pre-existing manifest for the "filtering" limitation but Pyle's creation of a new, separate manifest for the "generating" limitation — so the new manifest could not describe each asset in the filtered list. (Fed. Cir. op. at 6.)
  • On the "retrieving . . . a list of assets" limitation, the Board treated Petitioner's argument as an obviousness challenge and found no adequate articulation of a modification; the Federal Circuit held the Board did not err because the petition used strong "obvious" language, invoking § 103 rather than § 102.
  • The Board also rejected Petitioner's attempt (in Reply) to recast the petition as arguing that Pyle's creation of a new manifest met the "generating" limitation, finding that argument was never presented.
• Settlement / termination: None. The case was decided on the merits by FWD; no termination or settlement.
• Appeal: Yes — affirmed. Netflix filed the notice of appeal; the appeal was docketed as No. 22-1203 (the '720 FWD, IPR2020-00647) and consolidated with No. 22-1204 (the '515 FWD, IPR2020-00648). Panel: Linn and Chen (majority), Dyk (dissenting). Decision 2023-10-25, reported at Netflix, Inc. v. DivX, LLC, 84 F.4th 1371 (Fed. Cir. 2023), designated precedential. Netflix did not challenge the Board's substantive analysis; it argued only that the Board failed to address arguments purportedly raised in the petitions. The court affirmed, holding the Board did not abuse its discretion in its interpretation of the petitions and that "all of Netflix's arguments on appeal were not presented below and are thus forfeited." Judge Dyk dissented, concluding the Board improperly ignored arguments as to the "filtering" and "retrieving" limitations and would have remanded. A Federal Circuit mandate was docketed at the Board in early 2024 (Paper 33; a related "other court decision" entry at Paper 32 is dated 2024-02-02) — the date the affirmance became final.
  • Opinion: https://cafc.uscourts.gov/opinions-orders/22-1203.OPINION.10-25-2023_2211191.pdf
  • CourtListener: https://www.courtlistener.com/opinion/[9435057](/patent/9435057)/netflix-inc-v-divx-llc/
• Defensive value: Claim 1 is alive and judicially confirmed — an IPR-based defense on the Pyle+Marusi or Lewis+Marusi theories is a dead end. The FWD is a roadmap of what not to argue: do not use one reference's pre-existing-manifest selection for "filtering" and a different reference/embodiment's new-manifest creation for "generating," and do not vaguely invoke "obvious" without articulating a modification. Note also that the surviving asserted claims in the co-pending district court case were 1, 2, 4, and 5 — all upheld.

Strategic summary

Which claims are canceled vs. sustained vs. untested. Nothing was canceled. The IPR reached only claims 1–5 and 12, and the Board held all of them not unpatentable; the Federal Circuit affirmed, so claims 1, 2, 3, 4, 5, and 12 are SUSTAINED through completed IPR and appeal. Untested: whatever claims of the '720 exist beyond 1–5 and 12 were never challenged at the PTAB (the FWD addressed only the challenged set). In the parallel litigation, DivX asserted claims 1, 2, 4, and 5, and a joint status report in DivX v. Netflix lists them as "All found patentable." The practical consequence: the disclosure in the joint status filing states that only five of nineteen asserted claims across the five DivX patents-at-issue were found unpatentable — the '720 was not among the losers.

Estoppel landscape. IPR2020-00647 ran to an FWD and a Federal Circuit affirmance, so 35 U.S.C. § 315(e)(2) estoppel has attached as to Hulu and Netflix (and their privies/real parties in interest — DivX identified Netflix, Inc. and Hulu, LLC as RPI; DivX identified itself and DivX CF Investors LLC): they are barred from raising in any civil action or ITC proceeding any ground they raised or reasonably could have raised in the IPR. Because § 315(e)(2) is petitioner-specific, a new defendant is not estopped — but the Board's merits findings on Pyle's and Lewis's disclosures are now part of the public record, cited in a precedential Federal Circuit opinion, and would be a severe headwind to any new petitioner recycling the same two combinations. The realistic remaining prior-art space is art outside the Pyle/Marusi/Lewis family and grounds (e.g., different primary references, § 112 or § 101 theories) that the panel never decided. Netflix did try § 101 in the district court (DivX v. Netflix, C.D. Cal.), and that attack does not carry § 315(e)(2) estoppel.

Pattern signals. This was not a defensive-aggregator case — the petitioner was a pair of commercial streaming competitors (Netflix and Hulu) concurrently being sued by DivX (2:19-cv-01602 against Netflix; 2:19-cv-01606 against Hulu). The same petitioners ran a coordinated multi-patent IPR campaign across the DivX portfolio — companion IPRs include IPR2020-00646 ('792 patent) and IPR2020-00648 ('515 patent, the '720's continuation), with other DivX-family IPRs (e.g., IPR2020-00558 on the '588 patent, IPR2020-00614, IPR2020-00511) also on file. Netflix appealed multiple FWDs; in this appeal it lost, while it obtained a claim-construction reversal in a different DivX patent (the '588 patent, Fed. Cir. 2026 — not the '720), showing the campaign yielded mixed results. Unified Patents does appear in the broader DivX family (Unified Patents, Inc. v. DivX, LLC, IPR2019-01379, Paper 52, 2021-02-08, which Unified did not appeal) — I could not confirm from the sources reviewed that Unified's petition targeted the '720, so I do not count it as a proceeding on this patent. DivX (patent owner) did not need to appeal here because it won; the identity of the aggressive appellant is Netflix.

Recommended next steps

• If you are a defendant facing assertion of the '720's claim 1 (or claims 2, 4, or 5): treat this patent as confirmed, not weakened. The dispositive FWD is Hulu/Netflix v. DivX, IPR2020-00647, Paper 28 (PTAB 2021-09-27), "Final Written Decision Determining No Challenged Claims Unpatentable"; the disposition was affirmed at https://cafc.uscourts.gov/opinions-orders/22-1203.OPINION.10-25-2023_2211191.pdf. The FWD is available via the PTAB record: https://www.docketalarm.com/cases/PTAB/IPR2020-00647/Hulu_LLC_v._DivX_LLC/docs/09-27-2021-Board/Final_Written_Decision__original-28-JUDGMENT_Final_Written_Decision_Determining_No_Challenged_Claims_Unpatentable.pdf.
• Do not fund an IPR recycling Pyle+Marusi or Lewis+Marusi. For Hulu/Netflix privies that is statutory estoppel-bait under § 315(e)(2); for everyone else it is a losing merits theory the Board and Federal Circuit have already reasoned through. If you file, lead with a different primary reference and a cleanly presented, separately expounded alternative theory — the precedential Netflix v. DivX opinion makes clear the Board will not "decode a petition" to find arguments you hinted at.
• Attack what is untested. Claims outside 1–5 and 12 were never challenged, and § 101/§ 112 theories were not squarely resolved at the PTAB — those remain open avenues. Verify the '720's exact claim count against the issued patent before relying on any "untested claim" theory (I could not independently confirm the total claim count from the sources reviewed).
• No live trial milestones to track here. There is no active PTAB proceeding on the '720 — the appeal mandate has issued and estoppel has attached. Any new IPR would have to be a fresh petition by a non-estopped party, subject to the § 315(b) one-year bar running from service of a new complaint.
• Watch the parallel litigation for pressure points. DivX, LLC v. Netflix, Inc., No. 2:19-cv-01602 (C.D. Cal.) and ITC Inv. No. 337-TA-1222 remain the enforcement venues; the Hulu district case (No. 2:19-cv-01606) was dismissed 2022-08-25. Netflix's live § 101 eligibility challenge (see DivX's Response to Motion to Dismiss, Case No. 2:19-cv-1602) is the most promising public non-IPR defense thread.

Confidence/caveats: The existence, docket number, panel, grounds, institution date, FWD date, and claim-level outcome of IPR2020-00647 are confirmed by the FWD text and the Federal Circuit opinion. I could not verify (a) the total claim count of the '720, (b) whether any additional, later-filed IPR targeted the '720 (my searches surfaced only IPR2020-00647 for this patent number), or (c) the precise target patent of Unified Patents v. DivX, IPR2019-01379. Do not treat those three items as established.

I'll research the assignment record for US 9270720 across the sources available.

Let me dig into the specific assignment reel/frame records and correspondents.

Assignment Reconstruction — US 9,270,720 B2 ("Systems and methods for automatically generating top level index files")

Scope note / data caveat. I do not have a live connection to the USPTO Assignment Center, and the authoritative text provided for this patent (the Google Patents record) summarizes the post-issuance "reassignment" events without the reel/frame numbers or correspondent fields. I therefore reconstruct the chain from (a) the Google Patents legal-events entries supplied in the authoritative text, (b) cross-referenced family legal events, and (c) litigation/PTAB record. Where a reel/frame is not confirmed for this patent, I say so explicitly rather than inventing one. Verify all reel/frame and correspondent data at assignmentcenter.uspto.gov (search by patent number) or assignment.uspto.gov/patent/index.html.

Inventors

• The patent is a continuation of Ser. No. 13/341,789 (filed 2011-12-30), which claimed priority to provisional 61/529,403 (filed 2011-08-31). The named inventors were employees of the DivX business (by 2011 the DivX unit sat inside Sonic Solutions/Rovi after the 2010 DivX→Sonic Solutions and 2011 Sonic Solutions→Rovi transactions).
• Inventor-exodus pattern: not present. Braness is a prolific, long-tenured DivX inventor who continues to appear as a named inventor on DivX applications through at least 2020–2023 (e.g., US 2020/0137434, US 2020/0106820). This is not a "all inventors walked out the door before a fire-sale" fact pattern; the inventors remained attached to the DivX entity across the ownership reorganizations.
• Cross-check from PTAB filings: "Eric W. Grab, Adam H. Li, Brian T. Fudge, Jason Braness, Evan Wallin, and Ederson Ferreira are named inventors on patents asserted in both Hulu I and Hulu II" — i.e., these are core DivX portfolio inventors, not itinerant inventors attached to a one-off shell.

Original assignee

• Entity named on the issued patent (2016-02-23): Sonic IP, Inc. (the Google Patents "Original Assignee" field reads "Sonic IP LLC"; the family listing shows Sonic Ip, Inc. for US 9,270,720; Sonic IP's address of record is 4790 Eastgate Mall, Suite 200, San Diego, CA 92121). The filing event on 2014-07-21 is recorded as "Application filed by Sonic IP LLC."
• Nature: Sonic IP, Inc. is an IP-holding entity, not an operating company. The operating company in the corporate family was DivX, Inc. (a publicly traded San Diego codec/software company that did ship a product embodying adaptive-streaming technology — the DivX codec/player software and DivX Plus/Matroska tooling).
• Primary line of business (operating predecessor): DivX, Inc. — digital video compression (DivX codec), consumer software, and video technology licensing.
• Current status of the corporate family: DivX, Inc. was acquired by Sonic Solutions (2010); Sonic Solutions was acquired by Rovi Corporation (2011). In 2014 the DivX business/portfolio was moved out into DivX, LLC in a transaction involving private investors (the 2019–2021 Hulu litigation disclosures list Fortress Investment Group and Parallax Capital Partners among potentially relevant third parties). Unified Patents publicly characterizes current DivX, LLC as "a subsidiary of well-known NPE Fortress Investment Group." The current assignee, DivX, LLC, is thus best described as an IP licensing/enforcement entity that retains a residual software business, not the original operating company.

Assignment timeline

Chronological, from the Google Patents legal-events record for US 9,270,720 (execution dates not exposed in the sources I could reach; recording dates shown below). Reel/frame and correspondent could not be verified for this patent — the two reel/frame entries below are cross-referenced from a sibling patent in the same family and are flagged as unconfirmed for '720.

• (execution n/d) / recorded 2014-07-28 — Reel —/— (not confirmed)

  • Conveyance: Assignment of assignors' interest
  • Assignor: Jason Braness; Evan Wallin; Ederson Ferreira
  • Assignee: DIVX, LLC
  • Correspondent: not retrievable from sources reached (likely the DivX prosecution agent, Christie, Parker & Hale LLP — see note below; unverified as assignment correspondent)
  • Context: Original inventor→company assignment; standard employee-inventor capture into the DivX entity.

• (execution n/d) / recorded 2014-07-30 — Reel —/— (not confirmed; a sibling-family record shows Reel 032293/0557 for a DivX, LLC→Sonic IP assignment, but that reel is dated 2014-02-24 and does not match this 2014-07-30 record, so it is not attributed to '720)

  • Conveyance: Assignment
  • Assignor: DIVX, LLC
  • Assignee: SONIC IP, INC.
  • Correspondent: not retrievable
  • Context: Internal reorg — the patent is parked in the family's dedicated IP-holding subsidiary (Sonic IP, Inc.) two days after the inventor assignment.

• (execution n/d) / recorded 2018-02-13 — Reel —/— (a sibling-family legal-events entry shows Reel 045310/0020, assignors SONIC IP, INC.; DIVX, LLC, recorded 2018-02-12, assignee DIVX CF HOLDINGS LLC, New York; likely the same bulk record but unconfirmed for '720)

  • Conveyance: Assignment
  • Assignor: DIVX, LLC and SONIC IP, INC.
  • Assignee: DIVX CF HOLDINGS LLC
  • Correspondent: not retrievable
  • Context: Transfer to a newly formed acquisition/holding vehicle — per an RPX litigation record, "In late 2017, a new and unrelated entity was formed called 'DivX CF Holdings LLC' for purposes of acquiring certain Old DivX" assets. This is the pivot from operating-company ownership to a financing-backed holdco.

• (execution n/d) / recorded 2018-03-05 — Reel —/— (not confirmed)

  • Conveyance: Change of Name
  • Assignor: DIVX CF HOLDINGS LLC
  • Assignee: DIVX, LLC
  • Correspondent: not retrievable
  • Context: Change of name only — DivX CF Holdings LLC renamed itself DivX, LLC, which is why "DivX, LLC" appears as current assignee on the 2018-restructured portfolio.

• (execution n/d) / recorded 2019-02-28 — Reel —/— (not confirmed)

  • Conveyance: Change of Principal Place of Business
  • Assignor: DIVX, LLC
  • Assignee: DIVX, LLC
  • Correspondent: not retrievable
  • Context: Change of address only — no change in ownership.

Litigation/PTAB context (not assignments, but load-bearing for the analysis):

• 2019 (early) — DivX, LLC v. Netflix, Inc., No. 2:19-cv-01602 (C.D. Cal.); DivX, LLC v. Hulu, LLC, No. 2:19-cv-01606 (C.D. Cal.). The '720 is among the asserted patents.
• 2020-03-11 — Hulu, LLC v. DivX, LLC, IPR2020-00647 (challenging the '720). Final Written Decision 2021-09-27: no challenged claims unpatentable (DivX prevailed).
• 2022–2023 — Fed. Cir. Appeal No. 22-1203 (Netflix, Inc. v. DivX, LLC); related DivX appeals and ITC Inv. No. 337-3578 (TCL smart TVs).
• DivX's own complaints state it "owns by assignment the 'entire rights'" to the asserted patents, "originally assigned to Sonic IP."

Timeline diagram

timeline
    title Ownership of US 9270720
    2011 : Priority provisional filed
    2014 : Filed as continuation
         : Inventors assign to DivX LLC
         : DivX LLC assigns to Sonic IP Inc
    2016 : Patent issues to Sonic IP Inc
    2018 : Assign to DivX CF Holdings LLC
         : Renamed DivX LLC
    2019 : Suits filed v Netflix and Hulu
    2020 : Hulu IPR filed on 720 patent
    2021 : PTAB upholds all challenged claims

NPE / troll-pattern signals

1. Shell-entity transfer — PRESENT (moderate-to-strong).
The patent moves out of the operating-side entities (DivX, LLC / Sonic IP, Inc.) into DivX CF Holdings LLC, which the RPX litigation record states was "a new and unrelated entity… formed [in late 2017]… for purposes of acquiring certain Old DivX" assets (recorded 2018-02-13, likely Reel 045310/0020, unconfirmed for '720). The "-LLC / Holdings" suffix and the stated single-purpose acquisition rationale support the signal. It then renamed to DivX, LLC (change of name, recorded 2018-03-05). Caveat: DivX, LLC is not purely anonymous — it still trades on the DivX brand — so this is a transfer-to-holdco finding, not a pure registered-agent mail-drop shell.

2. Known asserter in the chain — PRESENT.
Fortress Investment Group (a well-known NPE financier) is the backer of current assignee DivX, LLC; Unified Patents states in a public post that DivX, LLC is "a subsidiary of well-known NPE Fortress Investment Group" (unifiedpatents.com, Feb. 2020). Parallax Capital Partners also appears in DivX's litigation disclosures. The named DivX/Sonic entities are not on the classic Acacia/Marathon/IV lists, but the Fortress backing is the controlling fact.

3. Repeat correspondent across the chain — UNCLEAR / NOT VERIFIABLE.
The assignment correspondent of record is not exposed in the sources I could reach for this patent, so I cannot confirm recurrence. Note (for follow-up, not a finding): the DivX/Sonic prosecution agent of record on family PCT filings is John W. Peck, Christie, Parker & Hale LLP, P.O. Box 7068, Pasadena, CA 91109-7068 — but that is a prosecution agent, and I have no evidence it filed the assignment recordings. Do not treat a single prosecution-firm appearance as the recurrence signal; that signal requires verified repeat recording-firm entries.

4. Cascading transfers through chained LLCs — PRESENT (but internal-reorg in character).
Two tight clusters: 2014-07-28 → 2014-07-30 (inventors→DivX, LLC→Sonic IP, Inc., two days) and 2018-02-13 → 2018-03-05 (DivX, LLC + Sonic IP, Inc.→DivX CF Holdings LLC→DivX, LLC, three weeks). Four recorded transfers within the chain, with the 2018 pair moving through a chained/incoming LLC. The transfers look like internal restructurings rather than arm's-length sales to unrelated buyers, so I weight this as a structural cascade, not a classic multi-buyer flip.

5. Pre-litigation transfer (within 6 months) — NOT PRESENT.
The last ownership-changing record is 2018-02-13 (to DivX CF Holdings LLC); the first suits naming the '720 (DivX v. Netflix, DivX v. Hulu) were filed in early 2019 — roughly 12–13 months later, outside the 6-month window. The 2018-03-05 and 2019-02-28 records are name/address changes only.

6. Bankruptcy fire-sale — NOT PRESENT.
No Chapter 7/11 proceeding involving the assignors (DivX, Inc., Sonic Solutions, Rovi/TiVo, Sonic IP, Inc.) appears in the reachable record for this chain. The 2014 DivX divestiture and 2017–2018 "CF" restructuring are corporate/financing transactions, not insolvency sales.

7. Privateering — PRESENT (with nuance).
The operating-side portfolio was separated from the operating business and placed with a financing-backed entity (DivX CF Holdings LLC / DivX, LLC, Fortress-backed) that then ran a broad assertion campaign against streaming and consumer-electronics companies (Netflix, Hulu, TCL, Realtek; ITC Inv. 337-3578). The nuance: the "operating company" here (Rovi/TiVo) sold the DivX business rather than retaining it and directing suit, so this sits between classic privateering and a straight portfolio sale to a funder. The Fortress characterization by Unified Patents tips it toward privateering.

8. Defensive aggregator (anti-NPE) — NOT PRESENT.
The chain terminates at DivX, LLC, an active litigant. The relevant defensive-aggregator activity runs the other way: Unified Patents (a defensive aggregator) has repeatedly petitioned against DivX patents (e.g., IPR2019-01379 on US 8,139,651; IPR2020-00647 on the '720; IPR2021-01476 on US 10,326,987). The patent is being attacked by a defensive aggregator, not acquired by one. RPX/AST/LOT ownership: not present.

Verdict

NPE — high confidence.

The chain shows the '720 first parked in an IP-holding subsidiary (Sonic IP, Inc., inventors assigned to DivX, LLC on 2014-07-28 and DivX, LLC reassigned to Sonic IP, Inc. on 2014-07-30), then moved in February 2018 (recorded 2018-02-13, likely Reel 045310/0020) into a newly formed acquisition vehicle, DivX CF Holdings LLC (renamed DivX, LLC, 2018-03-05). Unified Patents documents that current assignee DivX, LLC is a Fortress Investment Group subsidiary and that DivX is a frequent video-codec plaintiff, satisfying the "known asserter / NPE financier" and "shell/holdco transfer" signals simultaneously; the subsequent assertion campaigns against Netflix, Hulu, TCL and Realtek confirm active monetization. The residual "DivX ships software" fact keeps this just short of a pure paper shell, but the Fortress-backed holding structure plus the enforcement campaign meet the 2+ strong-signal threshold for high confidence.

Verification links:

• USPTO Assignment Center (search "9270720"): https://assignmentcenter.uspto.gov/ and https://assignment.uspto.gov/patent/index.html
• Google Patents legal events (authoritative for this session): https://patents.google.com/patent/US9270720/en
• Unified Patents NPE characterization of DivX: https://www.unifiedpatents.com/insights/2020/2/11/divx-llc-patent-determined-to-be-likely-invalid
• PTAB IPR2020-00647 (Hulu v. DivX, '720): https://portal.unifiedpatents.com/ptab/case/IPR2020-00647

Two fields remain genuinely unresolved and should be pulled directly from the Assignment Center before any filing/publication: (1) the exact reel/frame of each recorded assignment, and (2) the correspondent of record on each. I did not find a reliable source for either for US 9,270,720 specifically, and I have not asserted one.

I'll research the prior art cited in US 9,270,720 using web searches and the patent text already supplied.

Let me pull the detailed cited-reference lists and the IPR prior art (the Pyle, Marusi, and Lewis references from IPR2020-00647).

Let me retrieve the full front-page citation list (including non-patent literature) for the '720 patent.

Prior-Art Analysis — US Patent No. 9,270,720 (the '720 patent)

0. Identification (search hit confirmed)

The USPTO record matching the literal query "9270720" is:

• US 9,270,720 B2, "Systems and methods for automatically generating top level index files"
• Application No. 14/336,392, filed July 21, 2014; granted Feb. 23, 2016
• Inventors: Jason Braness, Evan Wallin, Ederson Ferreira (OCR variants "Branes"/"Braness" appear in litigation exhibits — these are OCR errors in the source documents, not different patents)
• Original assignee: Sonic IP, Inc., San Diego, CA; current assignee listed as DivX, LLC
• Continuation of App. 13/341,789 (issued as US 8,787,570, "Automatic Generation of SMIL Files"), which claimed priority to Provisional 61/529,403, filed Aug. 31, 2011
• Effective filing date / priority: Aug. 31, 2011. Because all claims have an effective filing date before March 16, 2013, the pre-AIA versions of 35 U.S.C. §§ 102/103 govern (AIA § 3(n)(1)).

No results for similar-but-different numbers (e.g., 9,270,721, 9,270,720-family relatives such as 9,270,731 or 9,270,735) are reported here; the citations below are those tied specifically to 9,270,720.

Methodology / caveat. I was not able to query USPTO PatentCenter directly with the available tools. The "references cited" list below is reconstructed from (a) the '720 patent's own front page as reproduced in IPR2020-00647 Exhibit 1001 (the patent itself), and (b) the Justia/Google Patents "Referenced Cited" listing. Where the OCR of those sources produced garbled or contradictory entries, I flag it rather than silently correct.

Critical framing: The references printed on the '720 patent's face are not the references that mattered. None of the front-page references was used by the examiner to reject the claims (claim 1 issued over them), and the references that were actually litigated as prior art — Pyle, Marusi, and Lewis — do not appear on the patent's face. They were developed by IPR petitioners. I therefore split the analysis into "face-of-patent" references (Tables 1–3) and "litigated/asserted prior art" (Table 4), which is where the § 102/§ 103 action is.

1. U.S. patent documents cited on the face of the '720 patent

For each entry: citation — date — brief description — § 102 posture. Descriptions marked (low confidence) are inferences; I could not independently verify the exact subject matter of every one of these references with the tools available.

Note on entries 6–8 and 22/25: the publication/grant dates given are after the Aug. 31, 2011 priority date, so these can only be prior art under pre-AIA § 102(e) (if by "another" and filed before the invention date) or not at all. A definitive § 102 mapping requires each reference's filing date (and inventorship), which I could not pull from PatentCenter with these tools. Entries 6–8 in particular may be § 102(e) candidates only.

2. Foreign patent documents cited on the face of the '720 patent

Foreign-language references on the face are ordinarily accompanied by an English-language IDS summary (the docket shows "Foreign Reference Documents including citations" filed 9/28/2015). I could not retrieve those summaries; the descriptions above are therefore generic.

3. Non-patent literature cited on the face of the '720 patent

These are prosecution documents, not technical prior art. They carry no § 102 weight against the '720 claims.

4. The prior art that actually matters — IPR2020-00647 (most relevant prior art)

This is the operative prior-art record. In IPR2020-00647 (Hulu, LLC and Netflix, Inc. v. DivX, LLC), petitions filed Mar. 11, 2020, institution granted Oct. 1, 2020, Final Written Decision Sept. 27, 2021 holding no challenged claim (claims 1–5, 12) unpatentable, affirmed by the Federal Circuit in Netflix, Inc. v. DivX, LLC, Nos. 2022-1203/2022-1204 (Fed. Cir. Oct. 25, 2023), 84 F.4th 1371.

I'll research the prior art relied on in the PTAB proceedings for this patent, since the "Prior Art" section isn't included in my current context.

Let me gather specifics on the Marusi and Lewis references and the Board's motivation-to-combine analysis.

Obviousness Analysis — U.S. Patent No. 9,270,720 ("the '720 patent") under 35 U.S.C. § 103

Scope note and cross-reference flag

Missing input. Your instruction says to "use the results from the Prior Art section of this page," but no Prior Art section is present in the context I was given — I have only the previously generated Patent summary and Litigation summary. Rather than fabricate a prior-art list, I reconstructed the operative art from the PTAB record for IPR2020-00647 (the only proceeding that has actually tested these claims on § 103 grounds) plus the Google Patents record. If a prior-art section was generated elsewhere in this workflow, this analysis should be reconciled against it, and I flag any divergence.

Consistency with the earlier sections. This builds on — and does not repeat — the Patent summary (claim text, bibliographic data, IPR2020-00647 outcome) and the Litigation summary (Netflix/Hulu suits, the § 101 ruling). Two points of reconciliation:

• The earlier Patent summary identified the IPR theories as "Pyle+Marusi" and "Lewis+Marusi." That is confirmed and now expanded below with the actual reference identities and disclosures.
• Date inconsistency (carried forward, not resolved). The system header for this session states today is 2026-09-22; the task prompt and the earlier sections use April 26, 2026. I note it, as the earlier sections did, and do not attempt to reconcile it.

Identifiers (interpreted literally): 9,270,720 / US 9,270,720 B2; US 14/336,392; US 8,782,268; EP 2180664; US 2012/0047542; US 7,873,706; US 9,998,515; IPR2020-00647; 84 F.4th 1371; nos. 2022-1203 / 2022-1204.

1. The prior art actually asserted against these claims

All four references below come directly from the IPR2020-00647 exhibit list (Petitioners' Updated Exhibit List, June 8, 2021):

Both instituted grounds map to claim 1 of the '720 patent:

• Ground I — claims 1–5, 12 obvious over Pyle in view of Marusi.
• Ground II — claims 1–5, 12 obvious over Lewis in view of Marusi.

(Source: Petitioners' Request for Oral Argument, IPR2020-00647; Petitioners' Corrected Demonstratives, June 8, 2021; ipverse case record for IPR2020-00647. The Google Patents "prior art keywords" for the '720 — playback, assets, content, top level, level index — are consistent with this art being directed to the same subject matter.)

2. Claim 1, element by element

Claim 1 as recited in the Federal Circuit opinion (Netflix, Inc. v. DivX, LLC, 84 F.4th 1371 (Fed. Cir. 2023); nos. 2022-1203/2022-1204) contains these limitations:

Where the art is strong. Elements 1[a], 1[b], 1[d] (in isolation) and 1[f] are well disclosed:

• Marusi is unusually on-point for 1[c]-as-to-capability and 1[d]. It literally discloses a database of product identifiers and associated device capabilities used to filter a content list: the mobile terminal characterizing unit "can retrieve information from database 160 about the capabilities of the mobile terminal 110 based on the identifier or identification code received from the mobile terminal," and matching unit 150 "identif[ies] a subset of the multi-media content files having at least one specific digital format … supported by the at least one capability." Marusi's ¶[0143] worked example (Nokia 8800 → only MP3 "Take a Bow" granted; no access to WMV content) is a per-asset grant/deny based on device capability. (https://patents.google.com/patent/EP2180664A1/en)
• Pyle is strong for 1[e]/1[f] and supplies the "generate a device-tailored manifest in response to a request" concept (FIG. 4; "manifest files that can be created based upon particular device(s) or capabilities," quoted in the Board's Institution Decision).

Where the art is weak — and this is the crux. The two limitations that defeated both grounds are the intermediate data-handling steps:

• 1[c] "retrieving … a list of assets." Pyle "maintains multiple manifests" — the Board and the Federal Circuit treated "maintain"/"store" as materially different from "retrieve," and the art never discloses assembling a superset list and then passing it to a filter. (Fed. Cir. 2023 slip op.; 00647 POR at 37, 41–42.)
• 1[d] "filtering the list." Petitioners used Pyle in two mutually inconsistent ways: selection of a pre-existing manifest for "filtering," and creation of a new manifest for "generating." The Board found this a fatal "disconnect," because 1[e] requires the index to describe each asset in the filtered list — so the filtered list and the generated index must be the same lineage. Pyle's newly created manifest is undisputedly "unrelated to Pyle's pre-existing manifests." (Netflix v. DivX, 84 F.4th at the Board-discussion section; 2021 WL 4859625, at *15.)
• For Lewis, the Board held that relying on the result (a manifest listing URLs) to prove the antecedent step ("retrieving a list of assets") is "insufficient without also providing a reason with rational underpinning" (KSR / In re Kahn). (IPR2020-00648 FWD, at 22–24; https://lowensteinweatherwax.com/wp-content/uploads/2021/09/2021-09-27-IPR2020-00648-Final-Written-Decision.pdf)

3. Motivation to combine — the KSR rationales a petitioner would (and did) invoke

A § 103 case here rests on Pyle as the base with Marusi as the secondary reference. The articulated motivations are:

1. Same field, same problem, same solution type. Pyle and Marusi are both network media-delivery systems that adapt what is delivered to what the requesting device can handle. Pyle expressly frames its system as enabling "random devices to contact a Web server, find and play a composition matched to the given devices and users." Marusi is a system that "match[es] digital formats supported by the capabilities of the mobile terminal with the digital formats of the multi-media content files." That is the "familiar element … arranged as the prior art suggests" rationale under KSR 550 U.S. 398, 417 (2007).

2. Pyle supplies a need that Marusi fills. Pyle's own disclosure requires the server to know which device is asking in order to select/compose the right manifest; Marusi supplies exactly the missing mechanism — a database keyed by device identification code returning capabilities, then a match to available content. The Petitioner's expert (Dr. Reader) opined the two were "complementary" and that a POSITA "would have been motivated to combine Pyle with Marusi to enhance the efficiency of generating, storing, and delivering multimedia content tailored to the capabilities of particular playback devices." (IPR2020-00648 FWD, quoting Ex.)

3. Known technique, predictable result. "Using a lookup table keyed by device ID to determine capabilities, and filtering a list of encodings against those capabilities" was, by 2011, a well-understood, routine database operation. Marusi itself claims it generically (matching a 256-bit "bit key" of capabilities against 256-bit "multi-media format information"). KSR: "the combination of familiar elements according to known methods is likely to be obvious when it does no more than yield predictable results."

4. For Ground II specifically: Lewis already teaches applying "additional rules" to further customize its final dynamic manifest. Substituting/enhancing those rules with Marusi's device-capability database is the substitution of one known rule-evaluation mechanism for another, with a predictable outcome (a manifest containing only URLs the device can play).

Reasonable expectation of success is high on its face: both references operate on manifest/index files for HTTP-streamed media, and the combination requires no new hardware — only a database lookup and a list comparison.

4. Dependent and separate claims

• Claim 2 — filtering based on "at least one of a geographic location of the playback device, a language associated with the playback device, one or more user preferences, and one or more requirements of a content owner." (DivX's own characterization of '720 claim 2, quoted in Dkt. 67 in DivX v. Netflix, 2:19-cv-01602.) These are business-rule filters; Marusi-mediated capability filtering plus Pyle's language/version "representations" gives a rational basis, and such content-owner rules were notoriously conventional. This claim is the most vulnerable to a well-articulated § 103 theory, and note it was not among the claims invalidated on § 101 in the district court (claim 2 was listed among the § 101-invalidated claims 1–2, 4, 5 per the earlier Litigation summary).

• Claim 4 — "the playback server system maintains a database of assets associated with specific pieces of content." Directly met by Pyle's manifest/representation catalog and by Marusi's storage means 130 with per-file format information.

• Claim 5 — "the top level index file describes at least a bitrate of each asset … and identifies locations of the assets." Met by Pyle (manifests "describe at least one location of one or more content segments," plus representation attributes such as bitrate) and by Lewis (manifest = list of URLs + bitrate).

• Claim 12 — challenged in both grounds but I could not independently confirm its verbatim text; treat any specific mapping as unverified. (Flagging per the operating rules.)

• System/CRM claims — if the '720 includes a playback-server-system independent claim and/or a machine-readable-medium claim mirroring § 1 (the earlier Patent summary could not confirm the total count), the same Pyle+Marusi mapping applies with the added "database … processor configured using a playback management application" elements.

5. Why these grounds nonetheless FAILED — and what that means for a § 103 theory

This is the most important part of the analysis, and it must be stated honestly:

• PTAB (IPR2020-00647, FWD Sept. 27, 2021): "Final Written Decision Determining No Challenged Claims Unpatentable." The Board rejected both grounds. (ipverse case record.)
• Federal Circuit (Netflix, Inc. v. DivX, LLC, nos. 2022-1203/2022-1204, Oct. 25, 2023; 84 F.4th 1371): affirmed, but on a forfeiture/waiver rationale — the Board "need not decode a petition to locate additional arguments beyond the ones clearly made." (https://cafc.uscourts.gov/opinions-orders/22-1203.OPINION.10-25-2023_2211191.pdf)

Consequences for a fresh § 103 analysis:

1. The Federal Circuit did not hold that the art fails to teach these limitations as a matter of law; it held Netflix failed to present the arguments. The merits findings that do exist are the Board's (Pyle's "maintain" ≠ "retrieve"; Pyle's selected vs. newly-created manifest "disconnect"; Lewis's "result ≠ antecedent step"). Those are evidence-based findings that a differently framed petition could attack — but they are also persuasive, on-record fact findings that a later § 103 challenger must squarely overcome.
2. The strongest available § 103 theory is therefore not Pyle-alone-for-both-filtering-and-generating (the disconnect), but a theory that keeps a single lineage for the list: e.g., a base reference that retrieves a superset list and then filters it in place, combined with Marusi for the product-ID→capability database. Lewis's "additional rules further customize the final manifest" is the closest teaching of in-place filtering, but the Board found Lewis "generat[es] only one manifest file, not a manifest and then a filtered manifest," and that retrieving-then-filtering would be "nonfunctional overhead … for no intended benefit." (IPR2020-00647 Reply/POR excerpts; 00648 FWD at 22–24.) A petitioner would need an explicit teaching or a strong articulated reason for the two-step retrieve-then-filter architecture.
3. Estoppel/§ 315(e) and § 325(d) considerations: the same Netflix/Hulu petitioners are largely estopped as to grounds raised or reasonably raisable; a new petitioner faces discretionary-denial risk. This is a practical, not a technical, obstacle.

6. Alternative / broader combinations on the face of the art

Beyond the two asserted grounds, the reference set suggests additional § 103 combinations worth evaluating (all flagged as candidate theories, not record-proven):

• Pyle + Marusi + Lewis, using Lewis (not Pyle) for the "retrieve a list … then filter" architecture and Pyle for the device-tailored manifest generation. This addresses the Board's "disconnect" by giving the two-step data flow an express source.
• Marusi as primary + Pyle: Marusi alone discloses the receive-ID → retrieve-capability → match-subset → return-subset pipeline (claim 1's logical core); Pyle supplies the "top level index file describing each asset" and the adaptive-streaming context. The gap is that Marusi returns access grants rather than an index file, so a manifest-format reference is needed — which is exactly Pyle/Lewis.
• Thoen (US 7,873,706) and the other background exhibits (Verizon/HDS-HLS-HSS article, Ex. 2004; the Wowza and "Transcoding" exhibits, Ex. 2002–2003) are useful only as evidence of the state of the art / what a POSITA knew (e.g., that manifests listing bitrates and alternative streams were conventional), not as anticipation.

7. Bottom line

• On the technical merits, claim 1 (and claims 4 and 5) are, on the face of Pyle + Marusi, plausibly obvious: every element has a home — Pyle for the device-tailored manifest generation/delivery, Marusi for the product-identifier→device-capability database and the capability-based filtering of a content list. The motivation to combine is strong (same field, complementary teachings, predictable database operation, KSR rationales).
• But the two-step "retrieve a list of assets → filter that list → generate an index describing each asset in the same filtered list" architecture is where the art is genuinely thin, because Pyle's filtering and generating steps are disconnected and Lewis arguably skips the "retrieve-then-filter" intermediate step entirely. That is precisely why the claims survived IPR2020-00647, and the Federal Circuit's affirmance — although grounded in forfeiture — leaves those Board fact findings standing.
• Practical posture: the '720 patent's asserted claims (1–2, 4, 5) have already been held invalid under § 101 by summary judgment in DivX v. Netflix, No. 2:19-cv-01602 (C.D. Cal.), so § 103 is not the operative validity battleground for those claims today. § 103 remains live for claims not covered by that ruling (e.g., claim 12 and any unadjudicated dependents).

Explicit uncertainties

1. No Prior Art section was supplied in my context; the prior-art set above is reconstructed from the IPR2020-00647 exhibit list and reference documents.
2. Verbatim text of claims 3 and 12 (and the total claim count) — not verified.
3. Whether the § 101 summary-judgment ruling has been appealed and its current status — not verified; would require the C.D. Cal. and Federal Circuit dockets directly.
4. Whether any post-2023 § 103 challenge (new petitioner) has been filed against the '720 — none found; my tools reached their search limit before a complete 2024–2026 docket sweep.

Primary sources: Fed. Cir. opinion, nos. 2022-1203/2022-1204 (Oct. 25, 2023), https://cafc.uscourts.gov/opinions-orders/22-1203.OPINION.10-25-2023_2211191.pdf; IPR2020-00647 Petition, POR, and Petitioners' Corrected Demonstratives (Docket Alarm); IPR2020-00648 Final Written Decision (parallel '515 patent), https://lowensteinweatherwax.com/wp-content/uploads/2021/09/2021-09-27-IPR2020-00648-Final-Written-Decision.pdf; EP 2180664 A1 (Marusi), https://patents.google.com/patent/EP2180664A1/en; ipverse case record for IPR2020-00647; Netflix, Inc. v. DivX, LLC, No. 24-1541 (Fed. Cir. Feb. 13, 2026) (separate '588 patent), https://www.cafc.uscourts.gov/opinions-orders/24-1541.OPINION.2-13-2026_2647816.pdf.