I. BACKGROUND

The parties have been litigating since early 2022, when Netskope filed a complaint for declaratory judgment that it did not infringe upon several of Fortinet’s patents. Dkt. 1. Fortinet counterclaimed that Netskope was infringing. Dkt. 36. A detour to the Patent Trial and Appeal Board for inter partes review stayed the case in 2023 and removed several patents from contention. Dkts. 73, 75, 83. The stay lifted in late 2024, dkt. 84, and the Court entered a case management order providing a year for discovery, set to close on November 7, 2025. Dkt 90. In January 2025, Fortinet served its Infringement Contentions (“ICs”) and claim charts on Netskope pursuant to Patent Local Rule 3-1. In February 2025, Netskope sent Fortinet a letter asserting that the ICs and claim chart did not comply with the Patent Local

Rules because they were insufficiently specific and did not put Netskope on notice of Fortinet’s theories of infringement for the accused patents. Dkt. 226-3, Ex. B. Fortinet wrote back with arguments to the contrary. Dkt. 226-4, Ex. C. The parties’ disagreement carried forward as they proceeded through discovery. Netskope produced technical docu- mentation and source code related to certain products in March 2025 and again in June 2025. Dkt. 211-13, Ex. 22. Over the summer, the Court granted in part Netskope’s motion to dismiss several of Fortinet’s counterclaims against Fortinet’s patents, leaving only three pa- tents remaining at issue (’825, ’601, and ’968), and issued a claim construction order. Dkts. 143, 160.

In August 2025, the parties brought a broadly defined discovery dispute before the Court. Dkt. 161. The undersigned ordered the parties to meet and confer and attempt to narrow their dispute, particularly regarding the ’825 patent. Dkt. 171. Netskope produced additional technical documentation and source code in August, September, and October, but the parties continued to spar over discovery and filed a second joint discovery letter regard- ing the ’825 patent and production of source code related to Netskope’s Cloud Firewall prod- uct. Dkt. 197. In the letter, Fortinet asserted that it would soon move to amend its ICs based on Netskope’s recent productions. The undersigned denied Fortinet’s request to compel production of the source code primarily based on an assessment that Fortinet’s claim chart for the ’825 patent was insufficiently specific to put Netskope on notice as to how the Cloud Firewall product infringed upon the ’825 patent, in violation of the Patent Local Rules. Dkt. 203. The undersigned noted that if Fortinet demonstrated good cause to amend its ICs, it could conceivably be entitled to more discovery on the Cloud Firewall product. In November, the Court issued a revised case management order, extending the dead-

line for discovery by a month, to December 8, 2025. Dkt. 206. The order stated: “No new contentions allowed.” Roughly three weeks prior to the new close of fact discovery, Fortinet moved to amend its ICs. Dkt. 213. Hoping for the Court to decide the motion prior to the close of fact discovery, Fortinet also filed an administrative motion to shorten the time to brief the motion. Dkt. 215. The Court referred both motions to the undersigned and specified that “a further extension of fact discovery is not necessary regardless of whether Fortinet is granted leave to amend” because Fortinet represented it was merely supplementing existing ICs rather than moving for leave to add new ICs. Dkt. 216. Netskope opposed proceeding on an expedited

basis. Dkt. 217. The undersigned agreed that Netskope should be given full time to oppose the motion and ordered briefing according to the typical schedule under Civil Local Rule 7. Dkt. 219. Netskope opposed Fortinet’s motion to amend, and Fortinet filed a reply in sup- port. Dkts. 226, 234.

II. LEGAL STANDARDS

The Patent Local Rules apply in this district to all civil actions which allege infringe- ment of a patent. Patent Local Rule 3-1 requires the party claiming patent infringement to

serve, in the early stages of discovery, a “Disclosure of Asserted Claims and Infringement Contentions,” or ICs. The ICs must state, separately for each asserted claim, each accused product or method (“Accused Instrumentality”) that allegedly infringes; the “identification shall be as specific as possible.” Patent L.R. 3-1(b). The ICs must include a chart “identi- fying specifically where and how each limitation of each asserted claim is found within each Accused Instrumentality,” known as a claim chart. Patent L.R. 3-1(c). The ICs must also state whether “each limitation of each asserted claim is alleged to be literally present or

present under the doctrine of equivalents in the Accused Instrumentality.” Patent L.R. 3-1(e). The “degree of specificity under Local Rule 3-1 must be sufficient to provide rea- sonable notice to the defendant why the plaintiff believes it has a reasonable chance of prov- ing infringement.” Shared Memory Graphics LLC v. Apple, Inc., 812 F. Supp. 2d 1022, 1025 (N.D. Cal. 2010) (cleaned up). The purpose of requiring the party claiming infringement to file ICs early in the pro- ceedings is twofold: to (1) “allow[] the plaintiff to develop facts to support the theory of the complaint” and (2) “allow the defendant to pin down the plaintiff's theories of liability . . . , thus confining discovery and trial preparation to information that is pertinent to the theories

of the case.” O2 Micro Int'l Ltd. v. Monolithic Power Sys., Inc., 467 F.3d 1355, 1365 (Fed. Cir. 2006). By mandating early notice of the accusing party’s theories of infringement, the “rules thus seek to balance the right to develop new information in discovery with the need for certainty as to the legal theories.” Id. at 1366; see also Atmel Corp. v. Info. Storage Devices Inc., No. 95-1987-FMS, 1998 WL 775115, at *2 (N.D. Cal. Nov. 5, 1998) (“The [patent local] rules are designed to require parties to crystallize their theories of the case early in the litigation and to adhere to those theories once they have been disclosed.”). The ICs also serve to structure and streamline discovery and to take the place of interrogatories. See Apple Inc. v. Samsung Elecs. Co., No. 12-0630-LHK-PSG, 2013 WL 3246094, at *1 (N.D. Cal. June 26, 2013). Patent Local Rule 3-6 provides that a party may amend its ICs “only by order of the Court upon a timely showing of good cause.” An example of a “circumstance[] that may, absent undue prejudice to the non-moving party, support a finding of good cause” is “[r]ecent discovery of nonpublic information about the Accused Instrumentality which was not dis-

covered, despite diligent efforts, before the service of the Infringement Contentions.” Patent L.R. 3-6(c). The party moving to amend bears the burden to show good cause. Karl Storz Endos- copy-Am., Inc. v. Stryker Corp., No. 14-0876-RS-JSC, 2016 WL 2855260, at *3 (N.D. Cal. May 13, 2016). In determining good cause, the court considers whether the moving party acted diligently and whether the opposing party would suffer undue prejudice. Id. Assessing diligence entails a two-part inquiry: diligence in discovering the basis for amendment and diligence in seeking amendment once the basis was discovered. Id. If the moving party was not diligent, the court need not consider whether the opposing party is prejudiced. Apple v.

Samsung, 2013 WL 3246094, at *1. Prejudice is typically found when amending contentions will disrupt the case schedule or other court orders. Karl Storz, 2016 WL 2855260, at *3. “Unlike the liberal policy for amending pleadings, the philosophy behind amending claim charts is decidedly conservative, and designed to prevent the ‘shifting sands’ approach to claim construction.” Atmel Corp., 1998 WL 775115, at *2.

III. DISCUSSION

Fortinet has not demonstrated good cause to amend its ICs. First, the amended ICs

advance several new infringement and doctrine of equivalents theories. Second, Fortinet was not diligent in moving to amend its ICs. Finally, Netskope would suffer prejudice if Fortinet were allowed to amend due to the advanced stage of the case.

A. New Theories Fortinet’s original claim charts for the ’825, ’968, and ’601 patents comprise 32, 113, and 607 pages, respectively. The amended claim charts comprise 69, 182, and 674 pages. Fortinet did not provide a redline version highlighting the differences between the two ver- sions of each claim chart. The Court previously instructed that any request from Fortinet that “seeks to go beyond supplementing existing contentions and moves for leave to add new infringement conditions” is restricted. Dkt. 216 at 2.1 Netskope asserts that the amendments are a “systematic overhaul” forbidden by the patent local rules and provides several exam- ples of allegedly new theories to illustrate its point. Fortinet contends that the amendments merely supplement its existing ICs and do not introduce new theories.

Fortinet made many amendments to its ICs. Not every change introduces a new theory; many are citations to source code or technical diagrams that standing alone would merely supplement existing claim contentions. But even a few instances of new theories are not allowed, and we are in a take-it-or-leave-it situation: Fortinet has not provided a clear path for granting leave to amend some of its contentions but denying leave for others.2 Re- gardless, because Fortinet has not demonstrated good cause, leave to amend only part of the ICs would not be granted, even if such a refined request were before the Court. In any event, having reviewed the original and amended claim charts, the undersigned finds that the amended ICs propose at least some new theories of infringement, detailed below.

1 Pin cites to the record, here and throughout, refer to ECF-generate pagination.

2 In its reply brief, Fortinet suggests that the Court “can strike Fortinet’s [doctrine of equivalents] allegations while permitting the rest.” Dkt. 234 at 15 n.4. This suggestion was raised for the first time in reply and can therefore be disregarded. First, the original claim chart for the ’825 patent defines the Accused Products to: include any Netskope products and services that provide micro- services performing different security functions, including the Netskope Security Cloud and Netskope One Platform products and services, such as the Netskope One Security Service Edge (SSE) and Netskope One Secure Access Service Edge (SASE), which utilize the Netskope Zero Trust Engine and Netskope “microservices.”

Dkt. 213-8 at 2. The amended claim chart defines the Accused Products as: the Netskope products (specifically, Private Access products (a.k.a., “Zerto Trust Network Access” and “ZTNA”) (collectively, “Private Access”), Secure Web Gateway (a.k.a., “SWG”, “Next Gen Secure Web Gateway,” “NG-SWG”,” “Netskope for Web”) (collectively, “SWG”), Cloud Access Security Broker and Cloud Inline (a.k.a., “CASB” and “Cloud Inline”) (collectively, “CASB”), and Cloud Firewall (a.k.a., “CFW,” “Firewall as a Service,” and “FWaaS”) (collectively, “CFW”)) that leverage Netskope’s platform (specifi- cally, NSProxy) in conjuncture with “microservices” and Partner Cloud Proxy. Alternatively, to the extent the ’825 Accused Products are limited to just SWG (as Netskope has contended for months), the theories disclosed in the original contentions and further clari- fied herein based on confidential source code and technical docu- ments recently produced by Netskope are equally applicable to just

SWG.

Dkt. 211-6 at 2. The Accused Product descriptions are different enough to imply a new theory. The undersigned previously found that the original ICs were, in violation of Patent Local Rule 3-1, insufficiently specific to put Netskope on reasonable notice of Fortinet’s theory of how the Cloud Firewall product infringed the ’825 patent. Dkt. 203 at 5. The Accused Products originally did not contain a cognizable theory of how Cloud Firewall infringed, but now explicitly include Cloud Firewall as an infringing product, so this change, at least, constitutes a new theory of infringement. Second, also related to the ’825 patent, the original claim chart mapped both the “first network security appliance” and “second network security appliance” described in many of the claim 1 limitations to Netskope’s “microservices.” The amended claim chart now theorizes that NSProxy is the first network security appliance and that a Partner Cloud Proxy is the second network security appliance. Dkt. 211-6 at 6. This cannot be a mere supplement of the original theory because the second network security appliance is now defined as a third-partner proxy whereas previously it was defined as a software capability within Netskope’s systems. Fortinet blames the modification on the lack of public material availa-

ble to substantiate its new theory, but that merely begs inquiry into the sufficiency of Fortinet’s grounds for its original infringement contentions. Third, regarding the ’968 patent, the amended ICs include a new doctrine of equiva- lents theory not detailed in the original ICs. See Patent L.R. 3-1(e). Specifically, regarding the “network interface device” described in several claims, the amended ICs assert that “mul- tiple gateways, switches, and/or servers perform substantially the same function, in substan- tially the same way, to achieve substantially the same result as a single network interface.” Dkt. 211-5 at 30. The original ICs for the ’968 patent do contain a theory of doctrine of equivalents, but discuss only the equivalency between physical and virtual devices; the the-

ory that the “network interface device” can be one device or many devices working in con- cert is not present. Fortinet argues that it would have been “impossible” to develop this theory prior to receiving discovery, but that supports Netskope’s argument that the theory as stated in the amended ICs is new. Implicitly conceding the point, Fortinet acknowledges that it “refined” its doctrine of equivalents theories and belatedly suggests that the Court can strike those amendments while permitting the rest. Dkt. 234 at 12 n.4. Fourth and finally, regarding the ’601 patent, claim 1 describes a “network system comprising . . . a plurality of selectable service area systems for processing data, each service area system comprising one or more selectable service area sub-systems.” In the original claim chart, services are described as capabilities such as “traffic steering, threat detection, advanced threat detection, URL filtering, SSL decryption, and DLP profiles[,]” each of which “has several sub-services[.]” Dkt. 213-7 at 26. In the revised claim chart, some of those same capabilities (“data loss prevention (‘DLP’), threat protection, . . . , URL filter- ing”) are described as “selectable service area sub-systems.” Dkt. 211-4 at 66. URL filter-

ing, e.g., was previously theorized to be a system, but is now theorized as a sub-system. This a new theory of infringement, particularly in light of Fortinet’s previous, successful argu- ment that systems and sub-systems are distinct terms with distinct meanings. See Dkt. 160 at 23. As the examples show, the amended ICs for each accused patent contain at least some new theories of infringement and the motion to amend could be denied on that basis in light of the Court’s prior orders. See Dkts. 206, 216. For thoroughness, the undersigned also analyzes whether Fortinet demonstrated good cause to amend by acting diligently and show- ing that Netskope would not be prejudiced by amendment.

B. Diligence To demonstrate diligence, the party moving to amend must show that they were 1) diligent in discovering the basis for amendment and 2) diligent in seeking amendment once the basis for amendment was discovered. Karl Storz, 2016 WL 2855260, at *3. Fortinet’s argument for its diligence is that the basis for amendment was primarily found in the large volume of source code files and technical documents produced by Netskope in September and October. See Dkt. 211-13 (listing dates and volumes of production). Fortinet contends

that it diligently pursued the basis for amendment by seeking this discovery but that Netskope improperly delayed, and that there was no other way the ICs could have been fleshed out further without looking under the hood of Netskope’s products. Once Fortinet received these productions, it argues that it acted quickly to analyze the productions and revise its ICs with the newly available information. Netskope argues in opposition that the bases for many of Fortinet’s new theories have been available in the public domain prior to Fortinet serving its original ICs, so they

could have used that information to amend and did not need to wait to receive confidential information to do so. Netskope also counters that any delay in discovery was due to Fortinet’s improperly vague infringement contention which did not accuse specific products of infringement, and that Fortinet alone is responsible for prosecuting its case. Both sides make fair points on this issue. The parties have engaged in a protracted discovery dispute stemming from their initial disagreement in February regarding the suffi- ciency of Fortinet’s original ICs. Despite the sparring, the parties appear to have made con- sistent progress throughout discovery, with a flurry of productions in September and Octo- ber. Fortinet has certainly been diligent in pursuing various discovery disputes.

However, the undersigned finds that Fortinet was not diligent in seeking amendment once the basis for amendment was discovered. Fortinet included with its reply exhibits con- taining for each patent a table describing the supplemental materials cited in each amended claim chart. Dkts. 233-6, 233-7, 233-8. The tables name certain non-public material pro- duced by Netskope, list the date of production, describe the reason for addition, and show each claim and limitation where the non-public material is cited in the amended ICs. By the undersigned’s count, almost half of the new citations to non-public material refer to material produced in August or earlier.3 Further, each amended claim chart cites to multiple

3 For the ’825 patent, there are 15 non-public materials listed; 5 were produced in August or documents produced in March or June. Thus, although Fortinet protests that it wasn’t until September and October that it began to receive source code and technical documentation in earnest, its amended ICs rely significantly on productions from earlier in discovery. Fortinet asserts that the earlier pro- ductions were technically deficient in various ways and that later productions provided nec-

essary context. But the undersigned is not persuaded that Fortinet acted diligently by waiting to move to amend until after the original fact discovery deadline, when it admittedly had by August nearly half the material it relies on now to amend its contentions. The Patent Local Rules require use of infringement contentions and claim charts to force litigants to crystalize their theory of infringement early in the case and to guide discovery accordingly. Netskope has asserted from the beginning that Fortinet’s ICs were deficient and based many of its discovery disputes on that central contention. As soon as Fortinet could have amended its ICs, it should have—both because it is required to under the Patent Local Rules and because doing so likely would have helped to smooth discovery. See CBS Interactive, Inc. v. Etilize,

Inc., 257 F.R.D. 195, 202 (N.D. Cal. 2009) (the “best practice” for amending ICs is “imme- diately after” obtaining new information). By waiting until after the original deadline for fact discovery, Fortinet thus failed to demonstrate diligence, the first part of the good cause assessment. That said, the consistently contested nature of the discovery process in the case does provide Fortinet with reasonable arguments in support of its diligence. The undersigned

earlier. For the ’986 patent, there are 54 non-public materials cited; 28 were produced in August or earlier. For the ’601 patent, there are 68 non-public materials cited; 31 were produced in Au- gust or cited. Across all three patents, about 47% of the newly cited materials in the amended ICs were produced in August or earlier. therefore assesses the prejudice faced by Netskope from amendment.

C. Prejudice To rehash the timeline of the case: the stay was lifted in November 2024, Fortinet served its original ICs at the end of January 2025, the Court issued a claim construction order on August 11, the original close of fact discovery was in November 2025, and the revised close of fact discovery recently passed on December 8, 2025. Netskope argues that it would be prejudiced by amendment because it would necessitate a modification of the Court’s schedule to permit additional fact discovery and because Netskope would have approached the claim construction hearing differently if it had known about Fortinet’s new theories of infringement in advance. Fortinet originally contended that Netskope would not suffer prej- udice because discovery remained open and because even a short extension would cure any

timing concerns. Dkt. 213 at 6. In light of the Court’s order that no additional discovery extensions will be permitted, Netskope contends now that no additional discovery is needed because its amended ICs do not contain any new infringement theories and that any prejudice to Netskope is of its own making. Allowing amendment at this late stage in the case would prejudice Netskope, for multiple reasons. First, as described above, Fortinet has included at least some new theories of infringement in its amended ICs. Given these new theories, Netskope persuasively argues that it would have proposed different terms and constructions for the claim construction hearing had it been given earlier notice of Fortinet’s amended ICs. Dkt. 226 at 28 (listing terms linked to the new theories in the amended ICs); see Google, Inc. v. Netlist, Inc., No.

08-4144-SBA, 2010 WL 1838693, at *3 (N.D. Cal. May 5, 2010) (finding prejudice when amendment would “potentially require additional claims construction proceedings”); Dynetix Design Sols. Inc. v. Synopsys Inc., No. 11-5973 PSG, 2012 WL 6019898, at *3 (N.D. Cal. Dec. 3, 2012) (“Because the proposed amendment would leave [Netskope] prejudiced by its previous claim term choices, . . . the prejudice here requires that the request by [Fortinet] be denied.”). Fortinet cites to In Wisk Aero LLC v. Archer Aviation Inc. for the proposition that

there is no hard and fast rule against amending ICs after claim construction. No. 21-cv- 2450-WHO, 2022 WL 5007912, at *5 (N.D. Cal. Oct. 4, 2022). True enough, but in that case, the court expressly invited the party claiming prejudice based on claim construction strategy to move to amend the claim construction order if needed to relieve prejudice. Id. Such remedy has been foreclosed in this case by the Court’s scheduling order. Dkt. 206. Secondly, filing the motion for amendment so close to the end of fact discovery sug- gests prejudice for Netskope. Other courts in this district have denied leave to amend based on prejudice resulting from similarly minimal amounts of time remaining in discovery. See Unwired Planet, LLC v. Apple Inc, No. 13-4134-VC, 2014 WL 5396735, at *4 (N.D. Cal.

Oct. 23, 2014); CyWee Grp. Ltd v. Apple Inc., No. 14-1853-HSG-HRL, 2016 WL 7230865, at *3 (N.D. Cal. Dec. 14, 2016). Had Fortinet amended its infringement contentions earlier, Netskope “could have spent that time investigating the new theories, taking discovery, pre- paring responses, and developing invalidity arguments within the new claim scope.” Flu- idigm Corp. v. IONpath, Inc., No. 19-5639-WHA, 2020 WL 5073938, at *5 (N.D. Cal. Aug. 25, 2020). Because it lost the opportunity to shape its discovery strategy in these ways, Netskope would be prejudiced by the amended ICs. * * * * * The Patent Local Rules require the party asserting infringement to crystalize its theory of infringement in the early stages of litigation. The purpose of the rules is to balance the need for the accusing party to obtain discovery into the functioning of the accused prod- ucts with the need for the accused party to understand the legal theories brought against it. The rules thus emphasis early adherence to specific theories of infringement and only permit amendment upon a showing of good cause. The rules are designed to prevent an accusing party from proffering non-specific infringement claims, building up theories of infringement throughout discovery, and only late in the game providing specific notice of their theories. Fortinet has not shown good cause to allow its late-breaking amendments.

IV. CONCLUSION

For the reasons above, Fortinet’s motion for leave to amend its infringement conten- tions is DENIED.

IT IS SO ORDERED.

Dated: December 22, 2025 Alex United States Magistrate Judge