Prior Art Litigation Search in San Francisco.

Prior art search San Francisco work is technology work first: the patents fought over in this city protect software architectures, machine-learning pipelines, payment rails and fintech protocols rather than molecules or mechanical devices. San Francisco and the wider Bay Area concentrate the accused infringers that patent-assertion entities target most — cloud platforms, AI labs, neobanks and app companies headquartered south of Market and across the bay. When those patents are challenged, the fight runs through the U.S. District Court for the Northern District of California, the Patent Trial and Appeal Board, the Federal Circuit and, for accused importers, the U.S. International Trade Commission. PerspireIP builds the prior-art and invalidity record those forums demand, on the compressed schedules the Northern District’s Patent Local Rules impose.

Why San Francisco patent defendants fight over software, AI and fintech

Every San Francisco invalidity matter begins with the same question a Boston chemistry case asks in reverse: what kind of claim is this, and where does the art that defeats it actually live? In the Bay Area the asserted claims usually cover a software method, a distributed-systems architecture, a machine-learning model, a mobile interface or a payment protocol. The plaintiffs are frequently non-practicing entities that acquired the patent to assert it, and the defendants are the platform, AI and fintech companies clustered in the city and down the Peninsula. That mix shapes everything about the search that follows.

Software, AI and fintech claims are rarely defeated by other patents alone. The reference that reads on a claim is often an open-source commit, an arXiv preprint, a product manual, an archived release note or an internet standard that predates the priority date by years. A search that only crawls issued patents will miss exactly the material that decides these cases. San Francisco defendants therefore need a search discipline built for non-patent literature, not a patent-database keyword sweep repackaged as an invalidity analysis.

Why prior art search San Francisco defendants front-load the record

Patent suits filed in San Francisco are heard in the U.S. District Court for the Northern District of California, the court that authored the model patent local rules now copied across the country. Those rules replace open-ended discovery with a fixed disclosure sequence. Under Patent L.R. 3-1, the patent holder must serve its Disclosure of Asserted Claims and Infringement Contentions within 14 days of the Initial Case Management Conference. Under Patent L.R. 3-3, the accused infringer must then serve its Invalidity Contentions within 45 days of receiving those infringement contentions — identifying each item of prior art and stating whether it anticipates or renders obvious each asserted claim.

The practical effect is that invalidity comes due early, on a clock the defendant cannot negotiate away. A prior art search San Francisco defendant that begins looking only after the complaint arrives is already behind, because the L.R. 3-3 chart is the first substantive statement of the defense and it frames claim construction that follows. The art has to be found, dated and mapped to claim elements before the deadline, not discovered later in fact discovery.

• Element-by-element charts keyed to each asserted claim, ready to drop into L.R. 3-3 contentions
• Anticipation and obviousness theories separated and supported reference by reference
• Public-availability dates fixed for every non-patent reference before the priority date
• A search built once and reused for both district-court contentions and a parallel PTAB petition

Inter partes review: patents and printed publications only

Northern District technology defendants lean heavily on inter partes review at the Patent Trial and Appeal Board, and for good reason: the Board reviews validity before technically trained judges, on a statutory timeline, at a lower burden of proof than a jury demands. But the forum carries a hard scope limit. An IPR petition may challenge claims only under 35 U.S.C. §§ 102 and 103, and only on the basis of prior patents and printed publications. On-sale, public-use and § 112 theories cannot be raised there; they belong in district court.

That scope limit is why the quality of the printed-publication search decides the petition. The petitioner must show a reasonable likelihood of prevailing on at least one challenged claim just to be instituted, and a separate clock applies: under § 315(b), the petition generally must be filed within one year of being served with the infringement complaint. A San Francisco defendant that wants both the district-court and the PTAB route open has to build its prior-art record fast enough to serve the 45-day L.R. 3-3 deadline and the one-year IPR bar from a single body of work.

Because IPR estoppel later bars grounds that reasonably could have been raised, the search cannot be shallow. Every patent and printed publication that could support an instituted ground needs to be surfaced up front, or it may be lost for the rest of the case.

When the decisive reference is non-patent literature

For Bay Area software, AI and fintech patents, the killer reference is usually not a patent at all. The relevant art is the record of how engineers actually built things, published in the open, years before the application was filed. A rigorous prior art search San Francisco defendants can rely on has to reach into sources a patent examiner rarely consulted.

• Open-source repositories: commit histories, tagged releases and issue threads on GitHub, GitLab and SourceForge
• arXiv and conference preprints that disclose a model or algorithm before any journal or patent
• Product manuals, API references, SDK docs and release notes for shipped commercial software
• Wayback Machine and web-archive captures that fix the public-availability date of a page or download
• Internet and industry standards — IETF RFCs, W3C, ISO and IEEE drafts that predate the claim
• Mailing-list and forum archives where a technique was described and discussed publicly

None of these help unless a reference qualifies as a printed publication — meaning it was sufficiently accessible to the interested public before the critical date. That is an evidentiary question, not a keyword question. We pin every non-patent reference to a verifiable public date using archive captures, repository timestamps, indexed release records and library or standards-body metadata, so the reference survives a challenge to its status rather than collapsing under one.

How Alice eligibility interacts with the prior-art search

San Francisco software cases carry a second front that biotech cases rarely see: patent-eligibility under 35 U.S.C. § 101 and the two-step framework of Alice Corp. v. CLS Bank. Northern District defendants often raise eligibility early, sometimes on a motion to dismiss before contentions are even due, arguing that the claim is directed to an abstract idea implemented on a generic computer. Eligibility and prior art are legally distinct, but in practice they draw on the same evidence.

At Alice step two, the question is whether the claim adds an inventive concept beyond the abstract idea, or merely recites well-understood, routine and conventional activity. Prior art showing that the claimed building blocks were already ordinary in the field strengthens the eligibility attack and, at the same time, feeds the § 103 obviousness case if the patent survives § 101. A search organized to serve both defenses at once gives counsel a coordinated record rather than two disconnected ones, which matters when the same judge weighs the motion to dismiss and later the invalidity contentions.

The ITC’s Section 337 fast track for accused importers

Not every Bay Area dispute stays in district court. When the accused product is imported — consumer hardware, networking gear, AI accelerators or connected devices assembled abroad — a patent owner can file a Section 337 complaint at the U.S. International Trade Commission seeking an exclusion order that blocks the product at the border. For San Francisco companies that design in California but manufacture overseas, that is a real exposure, and the ITC schedule is far faster than a district court’s.

After instituting an investigation, the Commission sets a target date for its final determination within 45 days, typically landing 16 to 18 months out, with fact discovery often compressed into the first several months. Unlike an IPR, a Section 337 respondent can raise the full range of invalidity grounds, including on-sale and public-use theories — but it must marshal all of that art on an even tighter clock than the Northern District imposes. The premium on front-loaded searching is highest here, because there is simply no slack in the schedule to develop art late.

How PerspireIP builds a San Francisco invalidity record

We start from the claims, not the keywords. Each asserted claim is broken into elements, and each element is mapped to the art that reads on it, so the deliverable arrives already organized the way L.R. 3-3 invalidity contentions, an IPR petition or an ITC respondent’s case needs it.

• Element-by-element claim charts with anticipation and obviousness mapping
• Deep non-patent-literature retrieval: repositories, preprints, product docs, standards and archives
• Public-availability dating for every reference, evidenced and defensible
• Patents-and-printed-publications sets scoped for a PTAB inter partes review petition
• Alice step-two support showing the claimed elements were routine and conventional
• A written invalidity memo grading the strength of each reference, not just listing hits

We work as a search partner to your litigation counsel under confidentiality, to court, PTAB and ITC deadlines. This engagement dovetails with our patent invalidation work and with a defensive patent infringement analysis, so validity and non-infringement positions are developed from one consistent evidence base rather than two searches that never reconcile.

We are also candid about what we find. A prior art search San Francisco counsel can act on is one that grades the references honestly — because a defendant deciding whether to petition for IPR, press a Section 337 defense or push for settlement needs an accurate read of the art, not an encouraging one. Whether you are an accused platform in the Northern District, an IPR petitioner racing the § 315(b) bar, or an importer facing a Section 337 investigation, the work scales from a single patent to a portfolio-wide campaign.

IP Landscape & Resources in San Francisco

Key intellectual-property authorities and venues relevant to San Francisco:

• United States Patent and Trademark Office (USPTO) — grants the U.S. patents asserted against San Francisco companies and publishes their prosecution histories
• Patent Trial and Appeal Board (PTAB) — hears inter partes review petitions based on patents and printed publications under 35 U.S.C. sections 102 and 103
• U.S. District Court for the Northern District of California — the San Francisco federal venue for patent suits, operating under the model Patent Local Rules
• U.S. International Trade Commission (USITC) — hears Section 337 import investigations on a fast target-date schedule against accused importers

Request a Prior Art Search in San Francisco

Explore related PerspireIP services: Prior Art Litigation Search · Patent Invalidation · Patent Infringement Analysis.

Frequently Asked Questions