Case Study: Semiconductor IP Due Diligence Case Study: Auditing a Fabless Chip Portfolio Before a $400M Acquisition

This semiconductor IP due diligence case study follows a private-equity buyer through the intellectual-property review that decided whether a $400 million acquisition of a fabless radio-frequency and analog chip designer would close on its original terms. In a sector where the U.S. Patent and Trademark Office granted more patents in semiconductor technology than in any other field for the third straight year, the patent estate is the target — and a portfolio that looks strong in a data room can hide chain-of-title breaks, licensing encumbrances and import-ban exposure that erase the thesis. The scenario below shows the exact sequence PerspireIP runs, the public data it is built on, and where the value moved.

The Challenge: A Chip Portfolio Worth More on Paper Than in Fact

The buyer, a mid-market private-equity fund, had signed a letter of intent to acquire a fabless semiconductor company whose value rested almost entirely on a 300-plus asset patent portfolio covering RF front-end modules and mixed-signal designs. The target’s management deck led with headline numbers: a portfolio filed across the U.S., EPO, China and Korea, and a claim that several families read on industry-standard 5G front-end architectures.

Three things made the diligence high-stakes. First, semiconductors are the most litigated technology at the U.S. International Trade Commission, where a limited exclusion order can block a chip line at the border within roughly 16–18 months — a remedy no damages award substitutes for. Second, the sector’s funding surge under the CHIPS and Science Act — $52.7 billion in incentives, including $39 billion in manufacturing grants — has pulled new entrants and new assertion campaigns into the same technical space, raising freedom-to-operate risk. Third, fabless designers routinely license third-party IP cores and rely on foundry design kits, so the ‘owned’ portfolio and the ‘usable’ portfolio are rarely the same set.

The fund needed to know, before it removed the financing contingency, whether the patents were clean, whether they were actually assigned to the entity being bought, and whether the target was itself exposed to an assertion that could halt shipments.

Our Approach: A Five-Layer Pre-Close IP Audit

PerspireIP structured the semiconductor IP due diligence as five sequential layers, each gating the next, so the buyer could stop paying for deeper work the moment a deal-breaker surfaced.

1. Portfolio reconstruction and asset census. We rebuilt the portfolio from primary records rather than the seller’s spreadsheet — USPTO Patent Center and PatentsView for U.S. assets, EPO Register and Espacenet for European families, and the CNIPA and KIPO registers for the Asian filings — producing an independent census of 312 live assets across 41 families.
2. Chain-of-title and encumbrance review. Every family was traced through USPTO assignment records and foreign register entries to confirm the target actually held title, and to surface security interests, prior licenses and employee-inventor assignment gaps.
3. Legal-status and maintenance audit. We verified fee status, term, and any terminal disclaimers or post-grant challenges (IPR, EPO opposition) that would shorten or weaken the assets being paid for.
4. Freedom-to-operate and litigation exposure. We mapped the target’s own products against live assertion campaigns and standard-essential patent pools in the RF front-end space, and screened ITC Section 337 dockets for respondents and technologies overlapping the target.
5. Value and integration readout. Findings were scored by severity and translated into a purchase-price and representations-and-warranties position the fund’s counsel could act on.

What the Audit Surfaced

The independent census matched the seller’s asset count within a few percent, but the layers underneath told a different story:

• Three chain-of-title gaps. Two U.S. families traced back to inventors from an acqui-hired startup whose assignments were never recorded at the USPTO, and one Korean family sat in a founder’s name rather than the company’s — a fixable but material defect that had to be cured as a closing condition.
• An undisclosed security interest. A venture-debt lender held a recorded lien across a slice of the U.S. portfolio, meaning the assets were not being delivered free and clear.
• Two ITC 337 exposure flags. The target’s flagship RF module read on technology at issue in active Section 337 semiconductor investigations of the kind non-practicing entities have driven to record levels — NPEs brought roughly a third of all Section 337 investigations at their peak, and 2025 saw dozens of new complaints instituted. An exclusion order against the product line would have removed the revenue the model depended on.
• Thin standard-essential coverage. Several families the seller marketed as ‘5G-essential’ were, on claim reading, implementation patents unlikely to survive an essentiality challenge — strong assets, but not the licensing annuity the deck implied.

The Outcome: A Repriced, De-Risked Close

None of the findings killed the deal — but together they moved both price and structure. The buyer used the audit to:

• Negotiate an $18 million purchase-price adjustment, reflecting the reduced licensing thesis and the cost of curing title.
• Make recordation of the three assignments and release of the lender’s lien conditions precedent to closing, so the entity delivered a clean portfolio on day one.
• Secure a specific IP indemnity and a holdback tied to the two ITC exposure flags, rather than relying on a generic representation.
• Redirect post-close integration spend toward the families that genuinely protected the product, instead of maintaining marginal assets worldwide.

The fund removed its contingency and closed with a defensible IP position — the difference between buying a patent count and buying a patent estate.

Lessons for Semiconductor Dealmakers

• Rebuild the portfolio from primary registers. A seller’s asset list is a marketing document; USPTO, EPO, CNIPA and KIPO records are the source of truth.
• Chain of title is where value leaks. Acqui-hires, founder filings and unrecorded assignments are the most common defects in venture-backed chip companies — and the cheapest to cure before closing, the most expensive after.
• Screen ITC 337 exposure explicitly. In semiconductors, the border remedy, not the district-court damages number, is the risk that reprices deals.
• Test ‘essential’ claims against the standard. A patent marketed as standard-essential is only worth the licensing story if the claims actually read on the standard.

Data Sources

The market and patent data referenced above comes from:

• Anaqua — 2024 Analysis of USPTO Patent Statistics — Semiconductor technology the #1 USPTO field for a third year; Samsung 10,427 and TSMC 4,127 granted patents in 2024.
• WIPO — World Intellectual Property Indicators 2025 — ~2.1 million patents granted worldwide in 2024 (+5.2%); semiconductor and digital-communication filings among the fastest-growing fields.
• USITC — Section 337 Statistics — NPE share of Section 337 investigations and annual complaint/institution counts; semiconductors the most frequently litigated technology at the ITC.
• NIST CHIPS Program Office — Incentives Funding — $52.7 billion in CHIPS incentives ($39B manufacturing grants, $13.2B R&D); grant and loan awards to date.

Discuss a Similar Matter

Frequently Asked Questions