US 11676600

I'll verify the patent's current status and check for any litigation/docket activity before summarizing.

Let me run one more targeted check for litigation/PTAB activity tied to this patent specifically.

Summary — U.S. Patent No. 11,676,600 (US 11,676,600 B2)

Verification / sourcing note

I checked the USPTO/Google Patents record and ran targeted searches for Federal Circuit and district-court dockets. I found no U.S. Court of Appeals for the Federal Circuit docket, appeal, or PTAB proceeding specifically identifying patent 11,676,600 as of this writing. The Google Patents page carries a generic "Family has litigation" flag pointing only to a Darts‑IP family-lookup link (family 51529448) — that is a pointer, not evidence of a suit on this patent. Do not treat the family litigation flag as a CAFC 2026 case tie. I am flagging this explicitly rather than inferring an appeal.

---

Bibliographic data

Assignment events: a security agreement to Wells Fargo Bank, N.A. as collateral agent (recorded 2024‑04‑15), released on 2025‑01‑02.

Date discrepancies to be aware of: Unified Patents' portal lists priority 2013‑03‑11, application date 2021‑08‑08, and grant 2023‑06‑12, and expiration 2033‑03‑11 — one-day offsets from the Google Patents record. These are consistent with UTC-vs-local timezone rendering artifacts, but they are literal differences in the records, so I am noting them rather than silently reconciling them.

Important uncertainty: the fetched full text is truncated. It contains the abstract, the "Definitions," the Summary of Invention, the description and the figure list, but the verbatim "What is claimed is" claim set is not present in the material I have. The claim overview below is therefore derived from the Summary of Invention paragraphs, which track the independent claims nearly verbatim in this document, and from the reference numerals used in the figures. Claim numbering, total claim count, and dependent-claim scope cannot be verified from what I have.

---

Abstract (as published)

"According to some aspects, a method of monitoring an acoustic environment of a mobile device, at least one computer readable medium encoded with instructions that, when executed, perform such a method and/or a mobile device configured to perform such a method is provided. The method comprises receiving acoustic input from the environment of the mobile device while the mobile device is operating in the low power mode, detecting whether the acoustic input includes a voice command based on performing a plurality of processing stages on the acoustic input, wherein at least one of the plurality of processing stages is performed while the mobile device is operating in the low power mode, and using at least one contextual cue to assist in detecting whether the acoustic input includes a voice command."

---

What the patent is about (plain language)

The patent is a Cerence (ex-Nuance) "always-on" voice activation disclosure. The core problem: prior art voice assistants required either a manual trigger (touching an icon, pressing a button) or a spoken trigger phrase, and required the device to already be awake. The patent describes instead continuously monitoring the microphone and deciding whether speech is an actionable voice command without any explicit trigger, and doing so in a power-stingy, staged/hierarchical manner so the phone can stay in a low-power mode as much as possible.

Three technical levers are emphasized:

1. Tiered processing stages — cheap first (voice activity detection: amplitude, duration, energy, spectral slope, periodicity, cepstral analysis, phoneme/phone-loop matching), then expensive (limited-vocabulary ASR, keyword spotting, statistical classification, full ASR, natural language processing). A cheap stage that rules input out as "spurious acoustic activity" terminates processing; a stage that is confident it sees a command escalates to response.
2. Contextual cues used as side information — time of day, location (GPS, Wi-Fi/BT inference), motion (accelerometer "picked-up" gesture), recent device activity/events, battery state, background acoustic profile, speaker verification via voiceprint, and microphone-array direction-of-arrival (TDOA).
3. Two-processor / dual-tier architecture — a low-power secondary processor (DSP) runs early stages so the main CPU need not wake; the primary CPU (or a network server) is engaged only when needed.

---

Independent claim overview (plain language)

Based on the Summary of Invention, the patent appears to present its independent claims in three matched families (method / computer-readable medium / mobile device), across three distinct inventive framings. Nine independent claims appear likely.

Group A — Triggerless detection (generally applicable)

1. Method: monitor the mobile device's acoustic environment; receive acoustic input; detect whether that input includes a voice command from a user without requiring receipt of an explicit trigger from the user; and initiate responding to the detected voice command. ("Explicit trigger" is defined to cover both manual triggers and spoken trigger phrases like "Hello, Dragon.")
2. Computer-readable medium counterpart encoding instructions to perform Group A's method.
3. Mobile device counterpart: at least one input (microphone) plus at least one processor configured to do the same.

Group B — Low-power mode plus contextual cue
4. Method: while the device is operating in a low power mode, receive acoustic input; detect whether it includes a voice command by performing a plurality of processing stages, with at least one of those stages performed while still in the low power mode; and use at least one contextual cue to assist the detection.
5. Computer-readable medium counterpart.
6. Mobile device counterpart (input + processor configured accordingly).

Group C — Two-processor staged processing
7. Method on a device having first and second processors: receive acoustic input in low power mode; perform a first processing stage on the first processor, prior to engaging the second processor, to evaluate whether the input includes a voice command; perform a second processing stage on the second processor only if further processing is needed; and initiate responding when either stage determines the input includes a voice command.
8. Computer-readable medium counterpart.
9. Mobile device counterpart: input + first processor + second processor, with the first stage performed before engaging the second, the second only if needed, and response initiated if either concludes a command is present.

Plain-language gist of all three groups: where the prior art demanded a button press or a magic wake word and an awake device, these claims cover deciding on-the-fly that ordinary speech is a command, doing it incrementally so the battery survives, feeding in non-acoustic context to sharpen the decision, and splitting the work across processors so the main chip stays asleep until it is actually needed.

Architectural support in the figures (useful for mapping claim elements): primary processor 115, secondary processor 125, audio codec 105, common bus 155, transducers 130, wireless component 160, motion detector and other sensors 150/170, voice response system 750, VAD stages 710, contextual cues 715, speech processing stages 720, NLP stages 730, response process 740 and parameters 785, mobile device 700, user 790, network(s) 795, servers 775.

---

Enforcement context (not patent-specific unless confirmed)

Cerence Operating Company has been active in voice-IP enforcement, but I could not confirm that 11,676,600 is among the asserted patents in any of the following:

• Cerence Operating Company v. [Amazon.com, Inc.](/litigations/by-plaintiff/Amazon.com%2C%20Inc.) et al, E.D. Tex. No. 2:26-cv-00372 (filed April 2026, consolidated; Judge Gilstrap), plus a parallel ITC action against Amazon.
• Cerence Operating Company v. Sony Group Corporation et al, E.D. Tex. No. 2:25-cv-00757 (filed 2025-08-01; stayed pending the ITC).
• Cerence Operating Co. v. Apple Inc., W.D. Tex. No. 7:25-cv-00400 (filed 2025-09-04).
• Cerence's October 2023 Samsung complaint (five Cerence patents, per Cerence's own SEC filing).

None of the complaint exhibits I surfaced list the '600 patent specifically. Treat any assertion that the '600 patent is in suit as unverified.

---

Confidence levels

• High confidence: bibliographic data, abstract, inventors, assignee, priority/filing/issue dates, and the technical content of the specification (all from the authoritative patent text).
• Moderate confidence: the identification and structure of the nine independent claims (Group A/B/C method–CRM–device triplets) — this tracks the Summary of Invention but is not a reading of the actual claim set.
• Low confidence / explicitly unknown: exact claim numbering and count, dependent-claim scope, any PTAB proceeding, and any CAFC 2026 appeal or district-court assertion involving this specific patent.

I'll search litigation sources for this specific patent number.

Let me dig deeper on the patent family's litigation and Cerence enforcement activity.

Key lead: a Cerence v. Apple case asserts a sibling patent ('885). Let me verify whether the '600 patent itself appears in any complaint.

Bottom line

I found no litigation that specifically asserts U.S. Patent No. 11,676,600. Despite targeted searching of Unified Patents' patent/litigation pages, CourtListener/Justia dockets, and patent-litigation news coverage, no complaint, IPR, ITC action, or appeal identifying 11,676,600 as an asserted patent was located.

Note one important limitation: my searches pointing at the ITC complaint and the asserted-patent lists for the Cerence v. Apple / Samsung / Amazon matters were cut off before returning results, so this answer is based on the sources actually retrieved and is not an exhaustive clearance search. See "How to verify" below.

What the record does show

Patent family litigation flag (not case-specific). The Google Patents record for US 11,676,600 carries a "Family has litigation / First worldwide family litigation filed" indicator (Darts-ip family ID 51529448). This tells you the family has been litigated somewhere, but it does not establish that the '600 patent itself is asserted. Notably, the source data are inconsistent about dates (Google Patents lists priority 2013‑03‑12 and grant 2023‑06‑13; Unified Patents lists priority 2013‑03‑11 and grant 2023‑06‑12), so treat those shifted values literally as reported.

The likely source of that family flag is Cerence's assertion of a sibling patent in the same "Methods and apparatus for detecting a voice command" family — U.S. Patent No. 9,361,885 — not the '600.

Related cases in the same family/portfolio (NOT confirmed to assert the '600)

These are provided because they are the plausible family-level litigation, and each is expressly flagged as not confirmed to involve 11,676,600:

Patents reportedly asserted in these matters (none listed as 11,676,600):

• Cerence v. Apple (7:25-cv-00400): U.S. 7,251,367; 7,453,439; 7,750,891; 8,712,755; 9,256,580; and 9,361,885 — "Methods And Apparatus For Detecting A Voice Command." The '885 is described as covering "methods for monitoring voice commands while operating in low power or idle mode," accused against iPhones/iPads (iOS5+), Apple Watches (watchOS 5+), and HomePods (Feb. 2018+).
• Cerence v. Samsung: reported patents include the '671 and '486 patents and U.S. 8,914,291 (Cerence's counts), with Samsung counterclaiming on U.S. 10,395,657; 10,720,162; 11,823,682; and 9,583,103.
• Cerence v. Amazon: press reporting describes five Cerence patents on speech recognition/related technology, without naming them in the retrieved sources.

Because the '600 and the '885 share title and lineage, if the '600 is eventually added to any of these pleadings (via amended contentions or a new complaint), Cerence v. Apple is the most likely candidate.

Deliberate false-positive warning (interpreted literally)

One search hit referenced an IPR2021-00598 petition that discusses a "'766 Patent." That proceeding concerns U.S. Patent No. 8,232,766 (litigated in Fundamental Innovation Systems International LLC v. Coolpad / Lenovo / TCT Mobile). That is not U.S. 11,676,600 — different patent, different number, different technology. Do not conflate the two.

Outcome / disposition

No outcome exists to report for 11,676,600, because no case asserting it was found. For the related matters above: all are pending; none has reached judgment.

How to verify (recommended, since my search was incomplete)

1. Unified Patents litigation portal — https://portal.unifiedpatents.com/litigation/caselist and the patent page https://portal.unifiedpatents.com/patents/patent/US-11676600-B2 (the retrieved snippet showed no litigation entries).
2. PACER / CourtListener RECAP — full-text search for "11,676,600" across D. Del., E.D. Tex., W.D. Tex., N.D. Cal., D. Mass., and the ITC.
3. Docket Navigator / Lex Machina — patent-level "asserted in" reports, which are the most reliable for a granted-2023 patent.
4. Darts-ip — family ID 51529448, via the Google Patents link, to see the specific family member(s) involved.
5. USPTO Patent Trial and Appeal Board — IPR/PGR petitions naming US 11,676,600 (none found).

If you need a litigation-clearance opinion, I'd treat the current answer as "no known litigation on the '600 as of this search," not as a definitive negative.

PTAB / AIA Trial Analysis — US 11,676,600 B2 ("Methods and Apparatus for Detecting a Voice Command")

Data-source note: The canonical list for this task is the structured "PTAB proceedings on file" block, which reports that the USPTO Open Data Portal returns no AIA trial proceedings for US 11,676,600 as of the most recent ingest. My web searches did not surface any petition, institution decision, FWD, or appeal naming US 11,676,600 (application 17/397,522). I did find a substantial cluster of PTAB activity against sibling patents in the same Cerence family (the '750 and '461 patents) — I report that below as context only, and I am explicit that those proceedings are not on the '600. I did not invent any proceeding numbers.

---

Proceedings overview

Total AIA trial proceedings on US 11,676,600: 0 — 0 active, 0 claims invalidated, 0 claims sustained, 0 settled, 0 institution denied.

Bottom line for a defendant: the '600 patent itself has never been challenged at the PTAB. That is not "hardened" and it is not "dead" — it is untested at the Board. Because no IPR/PGR petition against the '600 was ever instituted, no § 315(e)(2) estoppel attaches to the '600, and a defendant facing an assertion of this patent retains a full, unburdened invalidity playbook (district court and a first-ever IPR). The offsetting risk is procedural, not merits-based: the Board has recently denied institution against Cerence family patents on Fintiv discretionary grounds where parallel district court litigation was advanced, and the Federal Circuit has declined to disturb that (In re Google LLC, No. 25-144). So the practical question for a new petitioner is not "will the art work" but "will the Board even institute."

---

Proceedings on US 11,676,600

None. No IPR, PGR, or CBM proceeding has been filed against this patent, per the structured ODP block. There is no institution decision, no FWD, no settlement/termination, and no Federal Circuit appeal to report at the claim level. I will not manufacture claim-level outcomes that do not exist.

---

Related AIA proceedings in the same Cerence "voice command" family (context — these are NOT proceedings on the '600)

These matter to a defendant because the '600 shares a specification/family lineage and priority date (2013-03-12) with the patents below, and the same challengers (Samsung, Google) are already litigating the family. None of the proceeding numbers below names US 11,676,600.

• IPR2024-00978 — Samsung Electronics Co., Ltd. et al. v. Cerence Operating Company (family patent: US 11,087,750 B2, per the asserted-patent set)

  • Type: Inter Partes Review
  • Status: Terminated — Post-DI Settlement (Termination Decision dated 2025-08-11; joint motion to terminate filed 2025-08-04)
  • Institution: Instituted 2025-02-13 (Institution Decision: Grant; Scheduling Order same day)
  • Note: The docket shows a Panel Change Order (2025-04-10) and repeated joint stipulations to modify due dates — i.e., active settlement discussions well before the FWD date.
  • Defensive value: Ends in settlement, so there is no FWD and no claim-level estoppel on the '750. A defendant gets no free claim-cancellation benefit from this case.

• IPR2024-00979 (family patent, Samsung v. Cerence) — instituted 2025-01-31.

• IPR2024-01267 (family patent, Samsung v. Cerence) — instituted 2025-01-31.

  • Both institution decisions are cited for the proposition that Samsung's Sotera stipulations were "dispositive" in overcoming discretionary denial at that time. (Source: Google v. Samsung mandamus petition, https://www.reexamlink.com/wp-content/uploads/2025/09/Google-v-Samsung-Mandamus.pdf)

• IPR2024-01464 and IPR2024-01465 — Google LLC v. Cerence Operating Company (filed ~2025-04-23; IPR2024-01465 exhibit list identifies US 11,087,750 B2 "750 Patent"). (Source: https://ccianet.org/wp-content/uploads/2025/08/Joint-Amicus-In-re-Google-.pdf; PTAB E2E https://ptacts.uspto.gov)

• IPR2025-00457 (Samsung/E.D. Tex.-related, 2025-08-04) and IPR2025-00458, -00459, -00460 — [Samsung Electronics America, Inc.](/litigations/by-plaintiff/Samsung%20Electronics%20America%2C%20Inc.) v. Cerence Operating Company (filed 2025-06-25) — additional family challenges. (Source: CCIANet Joint Amicus, https://ccianet.org/wp-content/uploads/2025/08/Joint-Amicus-In-re-Google-.pdf)

• Institution denials (Fintiv) + Federal Circuit appeal — In re Google LLC, No. 25-144 (Fed. Cir.)

  • The Board denied institution of the petitions challenging the '461 and '750 patents purely on Fintiv grounds, applying the Acting Director's retroactive rescission of the Sotera stipulation standard.
  • Google and Samsung petitioned for writ of mandamus; the Federal Circuit denied it on 2025-11-06.
  • Verbatim: "Samsung ... and Google LLC jointly petitioned for inter partes review (IPR) of patents owned by Cerence Operating Company. The then-Acting Director ... denied the petitions, concluding such review would be an inefficient use of resources given the progress of ... parallel district court proceedings involving Cerence's assertion of the same patents. Google and Samsung now petition for a writ of mandamus. We deny the petition." (Source: PACER, Case 25-144, Doc. 29; https://www.patsnap.com/de/resources/blog/litigation/in-re-google-llc-25-144-ipr-mandamus-denied-patsnap/)

Underlying district court exposure (for context):

• Cerence Operating Co. v. Samsung Electronics Co., Ltd. et al., No. 2:23-cv-00482-JRG-RSP (E.D. Tex.) — asserts five Cerence patents including the '750 and US 9,361,885 ('885) (all "Methods and apparatus for detecting a voice command" family members).
• Cerence Operating Co. v. Apple Inc., No. 7:25-cv-00400 (W.D. Tex.) — asserts, among others, the '885 patent for low-power "Hey Siri" wake-word detection. (Source: https://ai-lab.exparte.com/case/dct/txwd/7:25-cv-00400/)

⚠️ Accuracy flag: I was able to confirm the patent number for IPR2024-01465 ('750) and the general family membership of the other proceedings, but I could not independently verify the exact patent number targeted by every one of IPR2024-01267 / -01464 / IPR2025-00457 / -00458 / -00459 / -00460. Treat the patent-number-to-proceeding mapping for those six numbers as unconfirmed, and confirm against PTAB E2E before relying on it.

---

Strategic summary

1. Which claims of the '600 are canceled / sustained / untested. All of them are UNTESTED. There is no PTAB record on the '600, so every claim stands as issued for PTAB purposes. The aggressive activity you may see in press coverage is aimed at sibling family members ('750, '461, '885) — notably the trade-press framing that "Samsung/Google lost their patent-invalidation bid against Cerence" refers to the '750 patent and the Fintiv denials against the '461/'750, not to the '600. Do not let an opposing party blur that distinction: a denial of institution on the '750 is not a validity ruling, and it certainly is not a ruling on the '600.

2. Estoppel landscape. Because no IPR was instituted on the '600, § 315(e)(2) estoppel does not attach to the '600 at all. A defendant being asserted against on the '600 can (a) raise any prior-art ground in district court, and (b) file a first-ever IPR against the '600 without any prior-raised-or-reasonably-could-have-raised bar. The only family-level estoppel that exists runs against Samsung/Google as petitioners on the '750/'461 — it does not bind an unrelated defendant and does not touch the '600.

3. Pattern signals.

• Repeat petitioners: Samsung and Google are serial challengers across this Cerence family (IPR2024-00978/-00979/-01267; IPR2024-01464/-01465; IPR2025-00457/-00458/-00459/-00460). The '600 is conspicuous by its absence from that list — worth investigating why (e.g., claim scope, priority/§ 112 vulnerability, or it simply wasn't asserted yet).
• Patent owner posture: Cerence fights institution hard — it won Fintiv discretionary denials on the '461/'750 and successfully opposed the mandamus petition. Expect a well-funded, aggressive preliminary-response and discretionary-denial fight if you petition on the '600.
• Defensive aggregator: No Unified Patents (or similar aggregator) involvement surfaced for the '600. The Unified Patents portal hosts a '600 profile page but shows no challenge activity. (Source: https://portal.unifiedpatents.com/patents/patent/US-[11676600](/patent/11676600)-B2)

---

Recommended next steps

1. State the absence plainly, and use it. No PTAB activity exists on US 11,676,600. For a defendant: you are not bound by any estoppel on this patent, and the full universe of prior art under §§ 102/103/112 remains available both in court and in a fresh IPR. There is no FWD to cite because none exists.

2. If you are contemplating an IPR on the '600: the gating risk is discretionary denial, not the merits. The Board has repeatedly denied Cerence-family petitions on Fintiv grounds and the Federal Circuit will not rescue you via mandamus (In re Google LLC, No. 25-144, 2025-11-06). If you are in parallel litigation, a Sotera-type stipulation may no longer be sufficient post-Rescission, and a later-filed petition invites a § 325(d) / General Plastic fight. File early and be prepared to argue the "true alternative" standard.

   • Trial-stage milestones if you do institute: institution decision within the § 314(b) window, then a statutory 1-year FWD deadline from institution (§ 316(a)(11)) — calendar the oral hearing and FWD issuance accordingly.

3. Verify the family mapping before you rely on it. Before using any of the sibling-proceeding citations above in a brief, pull the documents from PTAB E2E (https://ptacts.uspto.gov) and confirm (a) which patent each proceeding number targets, and (b) the claim-level disposition of IPR2024-00978's termination and any FWDs in the -00979/-01267 line. I flagged those mappings as unconfirmed above.

4. Rerun the ODP query periodically. The '600's absence from the PTAB docket is a moving target if Cerence expands its assertion campaign (e.g., into the Apple W.D. Tex. action, which currently pleads the sibling '885 patent). A newly filed petition on the '600 would change the defensive calculus materially and trigger a § 315(b) one-year clock analysis if you are already sued.

Sources: Structured ODP "PTAB proceedings on file" block; https://patents.google.com/patent/US11676600/en; PTAB E2E https://ptacts.uspto.gov; Google v. Samsung mandamus petition, https://www.reexamlink.com/wp-content/uploads/2025/09/Google-v-Samsung-Mandamus.pdf; CCIANet Joint Amicus, https://ccianet.org/wp-content/uploads/2025/08/Joint-Amicus-In-re-Google-.pdf; Patexia docket for IPR2024-00978, https://services.patexia.com/lawsuits/Samsung-Electronics-Co-Ltd-et-al-v-Cerence-Operating-Company-et-al-id-[201456](/patent/201456)/documents/; PatSnap, https://www.patsnap.com/de/resources/blog/litigation/in-re-google-llc-25-144-ipr-mandamus-denied-patsnap/.

I'll research the assignment chain. Let me start with targeted queries.

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

Assignment & Ownership Analysis — U.S. Patent No. 11,676,600 B2

Sourcing caveat (read first). I was unable to open the USPTO Assignment Center (https://assignmentcenter.uspto.gov/) or the Assignment Search mirror (https://assignment.uspto.gov/patent/index.html) directly — my environment only returns search snippets, not interactive queries. What follows is reconstructed from (a) the authoritative Google Patents legal-events block in the supplied full text, (b) corroborating SEC filings (Cerence 8-K Exhibit 10.3, June 17, 2020; Amendment No. 3 to Credit Agreement, April 15, 2024), and (c) third-party assignment indexes (PlainSite, Dimensions) that mirror USPTO reel/frame data. Where a reel/frame or correspondent could not be retrieved, I say so rather than guess. Two of my four reel/frame-targeted queries were cut off by a step limit, so this is not an exhaustive assignment-clearance pull.

---

Inventors

Confidence: moderate. The employer attribution follows from (i) the 2013 priority date predating the Nuance/Cerence separation (completed 2019-10-01) and (ii) the 2021 filing being made by Cerence Operating Co. I did not independently verify each inventor's individual employment record, so treat the Nuance→Cerence flips as inference from the corporate timeline, not from a personnel record.

Unusual patterns: None observed. There is no evidence of a mass inventor exodus within 12 months of filing. The four inventors are long-tenured Nuance/Cerence speech-technology personnel, and the same inventive entity recurs across the family (the sibling U.S. 9,361,885 shares the "Methods and apparatus for detecting a voice command" title and lineage). This is a continuation-family retaining entity, not a pre-fire-sale roster.

---

Original assignee

Cerence Operating Company (Burlington, Massachusetts) — named as original assignee on the issued face of the '600, per the Google Patents record.

• Corporate lineage: Cerence Operating Company is the operating entity within Cerence Inc. (NASDAQ: CRNC), the automotive-AI business spun off tax-free from Nuance Communications, Inc. effective 2019-10-01 (first day of regular-way trading 2019-10-02). The Nuance ↔ Cerence Intellectual Property Agreement dated 2019-09-30 obligated the parties to record patent assignments for assets allocated to the SpinCo group — this is the source of the recorded "Nuance Communications Inc. → Cerence" assignment instruments indexed at PlainSite (records id=9281678 and id=9966561). I could not retrieve the specific reel/frame covering the '600 family in the time available.
• Primary line of business: embedded automotive voice/NLU software (voice assistants, voice biometrics, gesture/gaze) licensed to automakers. Cerence publicly claims AI in 280M+ cars across 40+ languages (2019 announcements), with ~1,300 employees at separation.
• Does it ship a product embodying the claims? Yes — Cerence's own voice-assistant stack is the commercial embodiment; the '600's "always-on," triggerless, low-power voice-command detection is precisely the type of functionality Cerence licenses into head units.
• Current status: Operating. Public company (CRNC). No bankruptcy. Note the collateral facility history below (Wells Fargo 2020 → release effective 2024-12-31), which indicates refinancing activity, not distress-triggered disposition.

---

Assignment timeline

Two post-issuance USPTO records are surfaced by the Google Patents legal-events block. A pre-issuance inventor-to-assignee assignment was not captured (see note at the end of this section).

• 2024-04-15 (recorded; execution date not captured) — Reel 067417 / 0303 (reel/frame as cited in the 2025 release; see interpretation note)

  • Conveyance: Security Agreement (grant of security interest in patents)
  • Assignor: Cerence Operating Company
  • Assignee: Wells Fargo Bank, N.A., as Collateral Agent
  • Correspondent: Not captured. I could not retrieve the recording correspondent of record for this entry.
  • Context: Securitization — pledge of the '600 as collateral under Cerence's credit facility. The 2024-04-15 date aligns to the day with Amendment No. 3 to Credit Agreement (Cerence Inc., SEC accession 0001193125-24-095496), consistent with a routine supplemental "Notice of Grant of Security Interest in Patents" recorded as the '600 issued into the collateral pool.

• 2025-01-02 (recorded; effective 2024-12-31 per the parallel Cerence release record) — Release citing Reel 067417 / 0303

  • Conveyance: Release
  • Assignor: Wells Fargo Bank, National Association
  • Assignee: Cerence Operating Company
  • Correspondent: Not captured. I could not retrieve the recording correspondent for this entry.
  • Context: Lien release — the collateral agent released its security interest in the '600, returning clean title to Cerence Operating Company. The parallel release on a sibling Cerence patent (US 10,394,944) cites Reel 052935 / Frame 0584 (the 2020-06-15 Wells Fargo security agreement) and is its own reel 069797 / 0818; the '600 release cites 067417 / 0303, a different underlying security-agreement reel, which is why the '600 needed its own release citation. Functionally this is a refinancing-driven discharge, not a sale.

Interpretation note (flagged, not reconciled). Google Patents renders the 2025 entry as RELEASE (REEL 067417 / FRAME 0303). Applying the pattern seen on the sibling '944 patent — where the free-text names the released security agreement's reel/frame and a separate REEL/FRAME: field names the release's own reel — I read 067417/0303 as the '600-specific security agreement being released, with the release document itself recorded on a separate reel/frame that the truncated fetch did not expose. I cannot rule out that 067417/0303 is instead the release's own reel/frame. Verify at the Assignment Center before relying on it.

Pre-issuance inventor assignment — not surfaced. The Google Patents legal-events list begins at the 2021-08-09 filing by Cerence Operating Co. and does not display an inventor→assignee record. This is common where the operative assignment is a case-specific instrument recorded with a batch, or where title flows through employment/confidentiality agreements plus the 2019 Nuance→Cerence IP Assignment Agreement. Do not read its absence as "no assignment was recorded."

Facility-genealogy cross-reference (contextual, moderate confidence). A Cerence IP-security document names Barclays Bank PLC as Collateral Agent (with a "Notice of Grant of Security Interest in Patents"). That is a different collateral agent from Wells Fargo. One coherent reading: Cerence refinanced around the end of 2024, Wells Fargo's lien was released effective 2024-12-31, and a Barclays-led facility took a new IP security interest. If a Barclays security interest was recorded against the '600, it has not surfaced in the Google Patents legal events for this patent, and I could not confirm it. Flagging as a lead, not a finding.

---

Timeline diagram

timeline
    title Ownership of US 11676600
    2013 : Priority application filed by Nuance
    2019 : Nuance spins off Cerence
    2021 : Continuation filed by Cerence Operating Co
    2023 : Patent issued
    2024 : Security interest pledged to Wells Fargo
    2025 : Security interest released to Cerence

---

NPE / troll-pattern signals

1. Shell-entity transfer — Not present. The only non-owner party in the chain is Wells Fargo Bank, N.A., acting as Collateral Agent under a secured credit facility — a fiduciary lienholder, not a licensing vehicle. Title never leaves Cerence Operating Company, an operating company. No "IP/Holdings/Ventures" successor appears. (Reels 067417/0303, 2024-04-15 and 2025-01-02.)

2. Known asserter in the chain — Not present. Cerence Operating Company / Cerence Inc. matches none of the listed NPE directories (Acacia, Marathon, IV, IPNav, Wi-LAN, Mosaid/Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, etc.). Cerence is a publicly traded product company (formerly the Nuance automotive division). (No reel/frame needed; corporate-identity finding.)

3. Repeat correspondent across the chain — Unclear / not determinable. I could not retrieve any correspondent of record for either the 2024-04-15 security agreement or the 2025-01-02 release; the Google Patents legal-events block does not expose correspondent data, and my reel/frame-targeted queries were truncated before returning results. Because the signal is defined by recurrence of a specific attorney/firm, an absence of captured names cannot support even a weak finding. Recommended: pull the correspondent fields for reels 067417/0303 and the companion release reel directly from the Assignment Center.

4. Cascading transfers — Not present. Two events total, both facility-related, spread across ~9 months, with no chained LLC sequence and no shared-principal pattern. (2024-04-15 → 2025-01-02.)

5. Pre-litigation transfer — Not present. There is no transfer of ownership in the chain at all; the 2025-01-02 event is a lien release returning title to the original owner, which is the opposite of a transfer-to-asserter setup. Nothing in the record shows the '600 changing hands within six months of any suit. (Note for completeness: the release became effective 2024-12-31, roughly nine months before Cerence's 2025-09-04 W.D. Tex. complaint against Apple — but that suit asserts the sibling patents U.S. 9,361,885 et al., not the '600, per the prior section's litigation findings.)

6. Bankruptcy fire-sale — Not present. Cerence Inc. is a going-concern public company (CRNC); the 2024-12-31 release reflects refinancing, not a Chapter 7/11 disposition. No bankruptcy docket was identified.

7. Privateering — Not present. No transfer to any third-party assertion vehicle. Cerence's enforcement activity (Apple, Samsung, Amazon) is direct operating-company litigation against competitors, not a transfer-and-assert arrangement.

8. Defensive aggregator (anti-NPE) — Not present. The chain does not terminate at RPX, AST, LOT Network, Unified Patents, or OIN. The chain terminates back at Cerence Operating Company.

Aggregate signal count: zero strong signals, zero weak signals confirmed; one signal (repeat correspondent) unresolved for lack of data.

---

Verdict

Operating-company assertion.

Justification: the only recorded post-issuance events are a security-interest grant to Wells Fargo as Collateral Agent (recorded 2024-04-15, Reel 067417/0303) and the corresponding release back to Cerence Operating Company (recorded 2025-01-02) — a secured-financing pledge and discharge, with ownership never leaving the original operating assignee and no NPE, shell LLC, aggregator, or bankruptcy anywhere in the chain. Current owner Cerence Operating Company (Cerence Inc., NASDAQ: CRNC) ships automotive voice-assistant products embodying the claimed triggerless, low-power voice-command detection, and Cerence is itself an active plaintiff against competitor operating companies (Apple, Samsung, Amazon) on related family patents — though the '600 specifically is not confirmed asserted in any of those matters. The verdict would only flip toward "Insufficient data" if the reviewer insists on treating the un-captured inventor assignment as the sole record; the two secured-facility entries are nevertheless sufficient to exclude any NPE characterization.

Verify at: https://assignmentcenter.uspto.gov/ (search patent number 11676600) and the mirror https://assignment.uspto.gov/patent/index.html — retrieve the correspondent of record for Reel 067417/0303 and for the companion release reel, plus any pre-issuance inventor assignment, all of which remain unretrieved here.

I'll search for the USPTO record and the cited prior art for US 11,676,600.

The record confirms the patent but I need the actual citation list. Let me retrieve the "Patent Citations" and "Cited By" sections.

Let me pull the front-page "References Cited" list from full-text patent databases.

Scope note and a hard limitation I need to flag up front

Two things before the analysis:

1. Date discrepancy. My system date is 2026‑09‑30; the task header says April 26, 2026. I'm flagging it rather than silently picking one, because it matters for "as of" statements about legal status.
2. The citation list is not in the record I have, and my searches did not retrieve it. The authoritative Google Patents text supplied to me is truncated — it ends mid‑sentence in the description ("the acoustic input may be processed") and contains no "References Cited" / "Patent Citations" / "Cited By" section at all. My targeted web retrievals returned the '600's bibliographic data, abstract, and other patents' reference lists, but not the enumerated front‑page citations of US 11,676,600. So I cannot honestly do the literal task as framed ("for each patent citation … provide full citation, date, description, and which claim it anticipates") for the '600's actual citation set. I will not invent that list.

What follows is (a) what the record does verify about the '600 and its citation metadata, (b) the references actually surfaced in the record, correctly labeled as family vs. third‑party, and (c) candidate prior art with an explicit "not verified as on the '600's face" label and a § 102 mapping to the claim groups established in the earlier section.

---

1. Record verification for US 11,676,600 (confirmed)

These match the previously generated section; I'm not re‑deriving them.

2. Citation metadata — the only citation facts I can actually verify

From the Unified Patents page for US‑11676600‑B2 (the only retrieved source that quantifies the citation set):

• "Patent Art (95)" — i.e., the reference set (cited‑by‑examiner + cited‑by‑applicant U.S./foreign patent documents) numbers 95.
• "Non‑Patent Literature (8)."
• "Referenced By (0)" — per Unified's data, no later patent cites the '600.
• Portfolio metrics shown: PVIX 80.76, RNIX 79.83, SVIX 0.91, RUBY/CITX/BRIX all "D".

From the Google Patents record, the "Prior art keywords" field (the examiner's search concepts) is: mobile device; acoustic input; voice command; voice; acoustic.

Bottom line: 95 patent references and 8 NPL items exist on the face of the '600, but their identities are not in any source I retrieved. Any itemized list of them at this point would be fabrication.

3. References that surfaced in the record — and their correct status

These appeared in the searches, but they are not front‑page prior‑art citations. They are family members / "similar documents" / other cases' citations. Listing them accurately prevents them being misread as § 102 art.

4. Candidate prior art for the '600 (verified relevance to the claim groups; NOT verified as on the '600's face)

Because I could not retrieve the actual 95‑reference list, I am giving you the references that my searches show are actually in play against this exact claim language elsewhere, plus the obvious field art — each clearly labeled.

(a) Basye et al., U.S. Pat. No. 8,140,327 B2 — "System and method for controlling the operation of a device by voice commands" (Nuance Communications; tenants Basye/Shenhav/LeBeau).

• Verification status: Tentative identification. A PTAB petition (Google‑affiliated, at ptacts.uspto.gov/.../petitions/1556713) relies on a "Basye" reference (Ex‑1002) and on "Basye‑Shenhav" / "Basye in Combination With Shenhav and Parker" to challenge a Cerence patent (the "'750 Patent," Ex‑1001) whose claims recite: "a device configured to operate in a lower power mode and a higher power mode"; "performing automatic speech recognition (ASR), while in the low power mode"; "determining … whether the recognized … word or phrase matches a trigger word or phrase"; "requesting, via a network while in the low power mode, that at least one server perform recognition"; and "evaluating … at least one contextual cue selected from … time of day, motion of the device, location of the device, an occurrence of an event on the device, recent activity on the device, and state of a battery powering the device." That contextual‑cue group is verbatim the group in the '600's Summary of Invention. The petition snippet did not disclose the Basye patent number, so the number is my inference — do not rely on it without checking Ex‑1002.
• Dates: "Basye" predates 2013‑03‑12 (it is asserted art against a Nuance/Cerence family member), so it is squarely § 102(a)/102(b) (pre‑AIA) art if the effective filing date is the 2013‑03‑12 priority.
• Potential § 102 relevance: Group C (multi‑stage / two‑processor processing, server‑side recognition, lower‑vs‑higher power mode) and the contextual‑cue elements of Group B (contextual‑cue group of claim set). Note the '750's claims are trigger‑word based, whereas the '600's Group A independent claims are expressly "without requiring receipt of an explicit trigger" — so Basye's value against the '600 is likely § 103 in combination, not clean § 102 anticipation of Group A.

(b) "Shenhav," "LeBeau," "Parker" references relied on in the same PTAB petition.

• Status: surfaced as Ex‑references by label only; numbers not disclosed in the retrieved text. "Parker" is described in the petition as teaching sending "audio command information" to a server that performs speech recognition and returns "action execution instructions," and as teaching an immediate acknowledgment of the command before server processing completes (relevant to the '600's "initiating responding"/confirmation concepts). I cannot cite a number.

(c) U.S. Pub. US 2014/0365225 A1 (Broadcom) — "ultra‑low‑power adaptive, user independent, voice triggering schemes."

• Relevance: low‑power, handheld, always‑listening triggering, two‑dimensional HMM state machine, adaptive training.
• Date caveat: published 2014‑12‑11; filing appears to be after the '600's 2013‑03‑12 priority, so it is likely not § 102 art against the '600 at all. Relevant as background only.

(d) U.S. Pat. No. 10,373,615 B2 (Google Technology Holdings; Gunn et al.) — "Voice control user interface during low power mode."

• Relevance: low‑power mode voice control keyed to a predetermined command phrase; its own cited art includes US 2013/0289994 (Newman), US 2013/0325484 (Chakladar), US 2013/0325485 (Chittar), US 8,554,560 (Valsan).
• Caveat: this is the '615's citation list, not the '600's; its 2013 publications post‑date 2013‑03‑12 (only § 102(a)(2)‑style art if their filings predate the priority). Treat as a lead to check against the '600's actual face citations.

(e) U.S. Pat. No. 5,774,841 / Nuance and Nuance‑lineage "voice trigger" family (e.g., US 8,660,849‑type trigger detection).

• Status: I did not retrieve these from the '600's record and cannot confirm they are cited on it. I am not asserting them as the '600's citations.

5. § 102 framework you should apply (and why single‑reference anticipation is unlikely)

• Which § 102 applies. The '600's priority is 2013‑03‑12 (five days before the AIA's 2013‑03‑16 first‑inventor‑to‑file cutoff). If all claims' effective filing date traces to 2013‑03‑12, pre‑AIA § 102(a)/(b)/(e) governs (with § 102(e) for earlier‑filed U.S. applications/patents). If any claim lacks that priority support, AIA § 102(a)(1)/(a)(2) applies instead. This must be resolved claim‑by‑claim before any date analysis.
• Anticipation vs. obviousness. Each of the three independent claim families (A: triggerless detection; B: low‑power mode + ≥1 in‑low‑power stage + contextual cue; C: two‑processor staged processing with either stage able to trigger a response) recites a combination. Single‑reference § 102 anticipation of Groups B and C in particular is improbable; expect § 103 combinations. The most likely § 102 candidate for a narrow, single‑reference hit would be a reference that both detects a command in low power mode and evaluates a contextual cue.
• Family‑member exception. Any same‑family document (US 11,393,461; US 2022/0358924; US 9,361,885; US 12,548,565) is not prior art; watch for § 102(b)(2)(C) / common‑ownership arguments if an examiner cites co‑owned Cerence/Nuance art.

6. What I could not verify — and how to fix it

Unknown (no fabrication): the identities of the 95 patent references and 8 NPL items; exact claim numbering/count of the '600 (still missing from the truncated text); whether Basye = US 8,140,327; the case/patent number of the PTAB petition's "'750 Patent"; whether US 10,373,615 or US 2014/0365225 appear on the '600's face.

To get ground truth, pull the front page directly:

1. USPTO PatentCenter for application 17/397,522 → "References Cited" tab (this is the authoritative 95+8 list).
2. USPTO Full‑Text (PatFT) / Patent Public Search → the "Referenced By" and "References Cited" fields for 11,676,600.
3. Google Patents https://patents.google.com/patent/US11676600B2/en → scroll to "Patent Citations" / "Cited By" tables (the HTML my retrieval captured was truncated before these tables).
4. PTAB → locate the petition at the URL fragment ptacts.uspto.gov/ptacts/public-informations/petitions/1556713 and read Ex‑1001 (the '750 Patent) and Ex‑1002 (Basye) to confirm numbers, dates, and whether the '750 is in the '600's priority chain.

Once you send me the enumerated front‑page list (or the PatentCenter "References Cited" export), I can complete the requested per‑reference deliverables for all 95 + 8 items — full citation, publication/filing date, one‑line description, and the specific claim(s) each potentially anticipates under the correct § 102 subsection. I'd rather give you that complete table than a plausible‑looking one built on inference.

I'll ground this analysis by pulling the patent's actual claim set and its cited prior-art references, since the full text I was given was truncated before the claim and "References Cited" sections.

Good — I've located the published claim set and leads on the cited art. Let me pull the granted claim text and the specific "Patent Art" / NPL list.

Obviousness Analysis — U.S. Patent No. 11,676,600 B2

Under 35 U.S.C. § 103

---

0. Provenance caveat on the "Prior Art" section (read first)

You instructed me to use the results from the Prior Art section of this page. I have to flag a sourcing problem rather than paper over it:

• The authoritative full text I was given is truncated. It contains the abstract, Definitions, Summary, Description, and figure list — but neither the "What is claimed is" claim set nor a "References Cited" section ever appears in the fetched material.
• The Google Patents page itself (per the snippets I retrieved) exposes only the tags Prior art keywords: mobile device / acoustic input / voice command / voice, not an enumerated reference list.
• Unified Patents reports the counts — "Patent Art (95)" and "Non-Patent Literature (8)" — but the enumerated list did not come back in my retrieval (https://portal.unifiedpatents.com/patents/patent/US-11676600-B2).

So there is no verbatim "Prior Art section" from this patent's page available to me. I am therefore grounding this §103 analysis on two substitute sources, both flagged as such:

1. The claim text that was retrievable — the pre-grant publication claim 1 from Justia (https://patents.justia.com/patent/20210375285), which is the publication of the same application (17/397,522) that issued as the '600; and the Summary-of-Invention paragraphs in the authoritative text.
2. The prior art applied by the PTAB in the inter partes review against U.S. Patent No. 11,087,750 — Petitions 1556713 (Google + Samsung as co-petitioners), Schmandt Declaration, https://ptacts.uspto.gov/ptacts/public-informations/petitions/1556713/download-documents?.... This is the closest thing to an examiner/PTAB-grade §103 mapping that exists for this family, because (see §3) the '750 patent shares the '600's specification verbatim and the same March 12, 2013 critical date.

I am not representing that the '750 IPR references are the references cited on the '600's face. They are the best-evidenced candidate art for the same disclosure.

---

1. Governing framework

Effective priority date: March 12, 2013. The '600 is a continuation (app. 17/397,522, filed 2021-08-09; priority 2013-03-12; anticipated expiration 2033-03-12). All art below was assessed by the PTAB against a March 12, 2013 critical date, because the sibling '750 family carries the same date (the Schmandt declaration states he was "asked to consider … prior art available as of March 12, 2013").

POSITA (construct): a person holding a B.S. in electrical engineering, computer science, or equivalent, plus ~2–3 years' experience (or an M.S. plus ~1 year) in speech/signal processing and embedded/mobile system design, including power management of handheld devices. This is a conventional construct for G10L15/22 + H04W52/02 art.

Claim-scope versions analyzed. Because the granted claims are unknown to me, I analyze two candidate scopes:

• Scope A (as-published claim 1, likely closest to the granted independent device claim):

  "1. A mobile device comprising: at least one input configured to receive acoustic input … while the mobile device is operating in a low power mode; and at least one processor configured to perform, while in the low power mode, at least one processing stage on the acoustic input to evaluate whether the acoustic input includes a voice command, the at least one processing stage including transmitting at least a portion of the acoustic input to at least one server via a network for processing by the at least one server … wherein the mobile device is configured to initiate responding to the voice command when the at least one processing stage determines that the acoustic input includes a voice command."

• Scope B (Summary-of-Invention groups): (A) triggerless detection; (B) low-power mode + plurality of processing stages + at least one contextual cue; (C) first processor → second processor staged processing with response initiated if either stage finds a command.

Note: Scope A does not recite "without requiring an explicit trigger." That negative limitation appears in the Summary (Group A) but not in the retrieved claim 1. This bifurcation is decisive for the analysis (§4.5).

---

2. Prior art references to be combined

Labels below are the exhibit names used in the '750 IPR (Ex-1005 et seq.). I characterize each by what the petition and declaration attribute to it, not by first-hand reading — flagged accordingly.

Directly on point for the '600: the '750's claim 10/18 recites a Markush group — "time of day, motion of the device, location of the device, an occurrence of an event on the device, recent activity on the device, and state of a battery powering the device" — which is essentially the same contextual-cue list the '600 specification recites. The same spec is at issue.

---

3. Why the '750 IPR is probative for the '600

The '750 petition quotes the patent as disclosing:

"a hierarchical approach to monitoring and processing acoustic input that applies relatively low power processing to received acoustic information first and generally higher power processing as needed to evaluate the likelihood that the acoustic information includes a voice command." ('750 at 6:… , quoted in the petition)

That sentence appears verbatim in the authoritative '600 text ("some embodiments utilize a hierarchical approach to monitoring and processing acoustic input that applies relatively low power processing to received acoustic information first and generally higher power processing as needed…"). Same specification + same March 12, 2013 critical date ⇒ the '750 IPR art is directly transposable to the '600. The petition also identifies the parallel district case as Cerence Operating Co. v. Samsung Elecs. Co., Ltd. et al., 2:23-cv-00482-JRG-RSP (E.D. Tex.), which is consistent with the '750 and '600 both being Cerence family members.

---

4. Grounds of rejection

Ground 1 — Basye in view of Shenhav (Scope A; and Scope B, Groups B and C)

Basye discloses every element of Scope A, and Shenhav supplies the dual-processor staging:

Motivation (KSR): Basye and Shenhav are in the same field (embedded speech processing for power-constrained devices) and address the same problem the '600 identifies — the power cost of "recruiting the mobile device's relatively power intensive resources each time acoustic input is received." A POSITA seeking to keep the main CPU asleep had a clear design incentive to distribute the work between a low-power front-end processor and a higher-power processor, which is nothing more than "combining prior art elements according to known methods to yield predictable results." KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398, 416 (2007). Reasonable expectation of success: both references are directed to the same class of devices with interchangeable components.

Ground 2 — Basye + Shenhav + Parker

For any claim requiring network/server-side speech recognition of the voice command and/or user confirmation that the input was received, Parker supplies: (i) sending "audio command information" to a server that performs speech recognition and returns action instructions; and (ii) an immediate visual or audio acknowledgement. Motivation: Parker expressly identifies the time lag inherent in server-side recognition and its tendency to cause the user to repeat commands — a recognized problem that the '600 also attributes to network offload ("utilizing network resources … may result in a net savings in battery power … may justify the power consumption needed to transmit"). Combining Parker's acknowledgement/offload with Basye's low-power architecture is a predictable, two-way-compatible improvement.

Ground 3 — Basye + Shenhav + Lebeau (the contextual-cue claims — strongest ground)

Lebeau "monitors ambient sounds for voice input that indicates a request to perform an operation" and gates that monitoring on context, expressly including battery state (below ~25% charge) and power-cord connection, balancing "the potential convenience to the user of monitoring for voice input … against … potentially draining the battery."

This maps onto the '600's claimed cues almost one-for-one. The '600 recites cues including "time of day, the location of the mobile device, motion of the mobile device, recent activity on the mobile device, biometric information, status of the power source, information about the acoustic environment." The companion '750 claim 10/18 Markush group (same spec) was met by Lebeau. Under KSR, choosing among a finite number of identified, predictable additional cues (battery, motion, location) to bias a power-sensitive listening decision is routine optimization, and the '600 itself frames contextual cues as biasing stages rather than as a new mechanism ("Contextual cues may be used to bias one or more processing stages, aid in the selection and/or omission of one or more processing stages, operate as a processing stage itself…").

Ground 4 — Basye + Shenhav + Conroy

For any claim requiring the device to carry out a task without waking the main CPU / while remaining in the low power mode, Conroy teaches processing and playing audio "with little or no involvement of the CPU so that the CPU is free to enter a low power state which extends the useful battery life." This maps to the '600's disclosure that "some mobile devices are capable of rendering music while in a low power mode" and to voice commands such as "next track," "pause music," "increase volume." Motivation: extending battery life is the stated raison d'être of the '600 ("Attempting to generally optimize … while keeping power consumption to a minimum"), and Conroy teaches exactly the technique that achieves it.

Ground 5 — Treating the "without requiring an explicit trigger" limitation (Scope B, Group A)

This is the weakest point of the obviousness case and should be flagged as such:

• Basye, Parker, and the '750's own claim 1 all employ trigger/name/wake-word detection. The '600 defines an "explicit trigger" to include a spoken trigger phrase, so a wake-word system does not disclose a claim that affirmatively requires detection "without requiring receipt of an explicit trigger."
• To the extent a granted claim recites that negative limitation, a challenger needs (a) ambient/always-on listening art that detects commands with no wake word (Lebeau's "monitors ambient sounds for voice input that indicates a request to perform an operation" is the best candidate in this record), plus (b) an obvious design choice / "finite number of predictable solutions" argument, or the argument that the limitation is not entitled to patentable weight because it is a statement of intended use or a mere negative limitation that does not change the structural cooperation of the elements (the same hardware/method performs the same steps whether or not a trigger is spoken first).
• Alternatively, on Scope A, the argument is the reverse: because Basye requires more than Scope A recites (a trigger match), Basye discloses the broader claim; the trigger is surplusage relative to Scope A.

---

5. Dependent-claim vulnerabilities

---

6. Secondary considerations

None is evidenced. The record shows no unexpected results, no copying, no industry praise, and no nexus-bearing commercial success documented for the '600. The '600's own specification frames its contribution as an optimization trade-off ("strike an acceptable balance" between false positives and false negatives), which is the classic hallmark of an obvious design choice rather than a patentable advance. Cerence's enforcement activity (Cerence v. Apple, 7:25-cv-00400; Cerence v. Samsung, 2:23-cv-00482 and 2:24-cv-00181; the Amazon/ITC matters) is presumed-reasonable royalty / licensing evidence only and is not evidence of non-obviousness.

---

7. Confidence and limits

One deliberate false-positive warning: the NPL list appearing in the search results ("Speaker Recognition, Wikipedia … Jan. 30, 2014"; "Germain … Mar. 26, 2013"; etc.) comes from the file history of U.S. Patent No. 11,276,406, a different document. Several of those items are dated after March 12, 2013 and therefore are not prior art to the '600. Do not import them.

---

8. Bottom line

On the claim scope I can actually see (Scope A — low-power-mode listening with server-side command evaluation and conditional response) and on the Summary-level Groups B and C (low-power + plurality of stages + contextual cue; first-processor-then-second-processor), a prima facie §103 case exists on:

• Basye + Shenhav (core: low-power acoustic monitoring, staged/multi-processor evaluation, server-offloaded command recognition, conditional response);
• + Parker (server recognition and user acknowledgement), + Lebeau (contextual cues, including battery — a near-exact match to the family's own Markush group), + Conroy (task completion with the CPU in low power).

The single most defensible claim limitation is the affirmative "without requiring receipt of an explicit trigger" language — if it survives into the granted independent claims. That limitation is not met by Basye, Parker, or the '750's claim 1, all of which use a wake word. My plain reading of the retrieved publication claim 1 is that it does not carry that limitation, which would make the family's claims more exposed, not less.

Recommended verification before any reliance:

1. Pull the actual "What is claimed is" text for 11,676,600 from the USPTO PatentCenter/Global Dossier PDF (https://patents.google.com/patent/US11676600B2/en → Download PDF).
2. Retrieve the enumerated "Patent Art (95)" / "NPL (8)" list from https://portal.unifiedpatents.com/patents/patent/US-11676600-B2.
3. Confirm the '750 IPR status (IPR number, institution, FWD) at PTAB/PTActs and check whether the '600 was named in any IPR.
4. Verify the true bibliographic identities of Basye, Shenhav, Parker, Lebeau, and Conroy from the petition's exhibit list rather than from my summary.