The Valuation Is The Defence

Two experts walk into a courtroom. Each carries a valuation of the same patent. Expert A arrives with a discounted cash flow on a single 50 per cent rate, a 20-year horizon, and a peak revenue that assumes the patent covers the entire product line. Expert B arrives with a relief-from-royalty model grounded in ktMINE royalty bands for the specific industry, a 12-year remaining life, an apportionment argument grounded in Lucent v. Gateway 2009 and VirnetX v. Apple 2015, and a sensitivity table that runs the number from the 5th to the 95th percentile of comparable deal rates. The judge reads both reports. The judge reads both experts' deposition transcripts. The judge rules on Daubert. Expert A's number is excluded because the 50 per cent rate conflates technical risk with financial risk and the apportionment is unproven. Expert B's number is admitted because each input has a source and the sensitivity bounds the dispute. The valuation you bring to court is the defence. The valuation you keep in a spreadsheet is a guess.

The Fraunhofer mp3 patent family cost roughly fifty thousand euros to prosecute. The licence revenue from the family runs in the high double-digit millions every year. The cost approach puts the patent at fifty thousand. The market puts it at multiples of that, every year, for the life of the family. The cost approach is useful in three places and three places only: an early-stage patent before any comparable exists, a blocking patent that has no direct revenue, and an internal transfer pricing exercise for tax purposes. In every other situation it is a floor reference. United States courts in the Georgia-Pacific analysis treat cost only as a check on the floor of the reasonable royalty. German courts accept it as the statutory minimum under Lizenzanalogie but prefer the income calculation when there is one. A valuation built on cost alone will lose to a valuation built on income the first time it meets a damages expert in deposition. The cost approach is the defence that concedes the case before it opens.

The income approach is the dominant method in professional practice. Big Four firms, the IVSC, AICPA and courts all prefer it when sufficient data exists. The reason is simple: damages must be grounded in cash flow, not in cost. Inside the income approach, the relief from royalty method is the most widely used for individual patents. The math is present value of royalty rate times projected sales times one minus tax, discounted over the remaining patent life. The method is a hybrid because the royalty rate comes from comparable licensing data, which makes it depend on the market method too. The critique lives in this overlap: commercial royalty databases often show large differences between sources and offer little detail on the contractual terms behind the rate. For pharma and biotech, rNPV is the gold standard. It adjusts cash flows for phase-dependent probability of success before discounting, so technical risk lives in the probability layer and financial risk lives in the discount rate. The IQVIA 2025 transition probabilities sit near sixty per cent from preclinical to Phase I, sixty from Phase I to II, thirty from Phase II to III, sixty from III to regulatory, eighty-five from regulatory to market, with a cumulative likelihood of approval near eleven to twelve per cent from Phase I onward. A standard DCF that folds all of this into a single high discount rate conflates two different risks and systematically undervalues long-duration assets. Real Options Analysis values the optionality of the patent: the right to commercialise later, in a new geography, on a new indication. A binomial lattice is preferred over Black-Scholes because the underlying asset is rarely traded, volatility is unobservable, and the assumptions break. Courts do not value what the patent holder might do. They value what has been done and can be proved.

About half of litigated US patents are invalidated, per the Lemer 2008 work in the Journal of Economic Perspectives. The PTAB institution rate on inter partes review runs near sixty to seventy-five per cent, with a final invalidation rate around fifty to seventy per cent of instituted cases. The numbers mean a granted patent is not a defensible patent. A granted patent that has survived a challenge is a defensible patent. The valuation inflection lives in the survival. A patent that survives a validity challenge is worth two to five times what the same patent was worth before the challenge, on the rule of thumb that survives-plus-probability-of-validity-post-challenge divided by prior probability of validity sets the premium. The published stage ladder runs from a granted patent at two to fifty million euros, to a licensed patent at ten to five hundred million, to an enforced patent at twenty million to one billion, to a global portfolio at fifty million to five billion. The valuation you build before the challenge is a forecast. The valuation you build after is a number.

A patent is worth what a court will enforce, and every valuation method is a different defence when the number has to survive cross-examination.

Germany and the Unified Patent Court sell the fastest, most predictable injunctions in the world. The UPC covers roughly fourteen trillion euros of GDP and three hundred million people since June 2023, with a first-instance decision timeline of twelve to fourteen months. Article 68 UPCA awards negative economic consequences plus appropriate aspects including lost profits, unfair profits and moral prejudice. The Lizenzanalogie runs alongside three damages heads: lost profits, infringer's profits and reasonable royalty. United States damages under 35 USC 284 run adequate to compensate, in no event less than a reasonable royalty, with treble damages for wilful infringement and the Georgia-Pacific fifteen-factor test for reasonable royalty. The apportionment requirement from Lucent v. Gateway 2009 and VirnetX v. Apple 2015 holds damages to the contribution of the patented feature, not the entire product. On a per-capita, GDP-adjusted basis, German and UPC patents are worth one and a half to three times their US equivalents because of enforcement speed. The published cost differential runs both sides at USD three to ten million in the United States against EUR two hundred thousand to one million in Germany. The defence that wins in Karlsruhe in twelve months is not the defence that wins in Texas in thirty-six months.

Three valuation defences lose in cross-examination on a regular basis. The first is the cost-only valuation. It loses because cost is a floor, and the floor concedes the ceiling to the other side. The second is the market-only valuation without a true comparable. It loses because the court asks for the specific comparable transaction, the date, the parties, the technology class and the contractual structure. A royalty band cited without a source is a number, not a defence. The third is the single-rate DCF that folds technical and financial risk into one discount number. It loses because the IQVIA probabilities and the Farre-Mensa signalling premium of plus forty-seven per cent VC probability, plus seventy-six per cent loan probability and plus one hundred and twenty-eight per cent IPO probability from a first patent are public, peer-reviewed and not in the model. The defence that wins combines all three approaches, names the method for each claim head, and bounds the dispute with a sensitivity table the judge can read in five minutes.