.webp)
Introduction
Inside every cellular network lies a quiet engineering problem. Operators needed to squeeze more subscribers into the same physical space without expanding towers, buying spectrum, or tearing apart existing infrastructure. One solution involved redesigning how antennas beam signals. Instead of one symmetrical arc covering a standard sector, you split the sector, use two asymmetrical sub-beams, and reduce the handover zones where users momentarily lose stable connection as they move between cells. The critical coverage area stays intact whereas the capacity goes up. That, in compressed form, is the invention at the heart of Indian Patent No. 240893, filed by a Canadian company called Communication Components Antenna Inc (CCA).
In May 2026, the Delhi High Court in Communication Components Antenna Inc. v. Rosenberger Hochfrequenztechnik GmbH and Co. KG and Others, CS(COMM) 653/2019. The court upheld every claim of the patent, rejected every ground of revocation the defendants raised, found infringement established through computational simulation, and awarded damages of Rs. 152.32 Crore, approximately USD 18.28 million, at a 20 per cent royalty on total antenna sales. It is one of the largest patent damages awards in Indian legal history. It is also, technically and procedurally, one of the most rigorous patent judgments the Delhi High Court has produced.
The Parties
CCA is a Canadian company and the successor to M/s TenXc Wireless Inc., which originally filed the patent application in 2008 through the Patent Co-operation Treaty route. The patent was granted on 9 June 2010. TenXc ran into financial difficulties and sold its assets in December 2011. CCA purchased the patent portfolio in January 2012 and was formally recorded as the registered proprietor by the Indian Patent Office in November 2019. By the time trial commenced, CCA had sold over 30,000 antennas in India from its design facility in Ottawa, Canada.
The Rosenberger Group comprised four entities: Rosenberger Hochfrequenztechnik GmbH and Co. KG, a German multinational with operations across 22 locations in roughly 13 countries and over 11,000 employees; Prose Technologies India Private Limited, formerly Rosenberger Electronic Co. (India) Pvt. Ltd., an Indian entity that designed, manufactured, and sold the disputed antennas; Rosenberger Asia Pacific Electronic Co. Ltd., based in Beijing; and PROSE Technologies (Suzhou) Co. Ltd., incorporated in China.
CCA alleged that Rosenberger sold 1,27,974 infringing antennas in India, more than four times CCA's own sales volume, to buyers including Reliance Jio Infocomm Ltd. and Bharti Airtel.
The Patent: IN'893 and What It Claims
Indian Patent No. 240893, titled "Asymmetrical Beams for Spectrum Efficiency," comprises 25 claims. Claim 1 covers the method. Claim 10 covers the sub-sector antenna itself.
As of the priority date, 17 March 2006, operators were adding more antenna sectors to serve growing subscriber numbers. The unintended consequence was larger handover zones, where two beams overlapped, shrinking the dominant coverage area where traffic was actually carried. The invention replaces a conventional sector antenna with a split-sector antenna whose sub-sector beams are intentionally asymmetrical along a line of reference through the peak gain point in the horizontal plane. The total critical coverage area is maintained substantially equivalent to that of the antenna being replaced. Handover zones individually shrink rather than grow. More subscribers can be served from the same infrastructure.
The corresponding US patent, No. 8,311,582, was granted in November 2012 and survived an inter partes review before the US Patent Trial and Appeal Board in November 2016. A rehearing request was denied in March 2017. The EPO counterpart was withdrawn for commercial reasons, and a Chinese counterpart was invalidated under Chinese-specific law. The court treated neither foreign outcome as determinative of Indian validity.
Issues
The court framed three issues in July 2023. First, whether IN'893 was liable to be revoked. Second, whether Rosenberger infringed the patent. Third, if infringement was established, what reliefs CCA was entitled to and in what amount.
Issue One: Validity of IN'893
Novelty
Under Indian patent law, anticipation requires the whole of the subject invention to be present in a single prior art document. Prior art documents cannot be combined to construct a case for lack of novelty. The court applied this test to a collection of US and PCT patent applications Rosenberger cited. None of them individually disclosed the combination of intentional asymmetrical beam patterns paired with maintenance of substantially equivalent critical coverage area. The novelty challenge failed across all eleven cited documents.
Inventive Step and the Dartboard Model
This is where the judgment introduces its most influential conceptual contribution. Rosenberger cited prior art drawn from adaptive antenna systems, three-dimensional lens antennas, and underlay-overlay networks, fields that are not obviously connected to each other or to split-sector antenna design for cellular infrastructure. The defendants' expert witness admitted he lacked personal technical expertise in the relevant field and that he had endorsed the pleadings without independently analysing how the cited documents, in combination, would have motivated a person skilled in the art to arrive at the invention.
The court held that for multiple prior art references to establish obviousness, there must be a demonstrable link between them. That link may take the form of a common field of technology, a common problem the references address, or a shared conceptual source. Without that link, combining unrelated documents is not analysis. It is guesswork dressed in technical language.
To describe and censure this approach, the court formulated what it called the Dartboard Model. The subject invention sits at the bullseye of a dartboard. Each prior art citation is a dart thrown at the bullseye in the hope that one will land. The court observed, in terms that carry significant practical weight, that AI tools and Large Language Models can now generate large volumes of superficially plausible prior art citations very quickly, at minimal effort and cost to the party deploying them. The effect on patent holders, who must respond to each citation with substantive technical evidence, is serious. Courts may impose heavy costs on litigants who deploy the dartboard approach.
Sufficiency of Disclosure
Rosenberger argued that the patent was insufficient because it failed to disclose how to induce asymmetry, left key terms like "critical coverage area" and "substantially equivalent" undefined, and had been abandoned before the EPO after examination objections.
The court conducted a detailed comparison of the disclosure standards applicable in three jurisdictions. Under Section 10 of the Patents Act, 1970, as interpreted by the Supreme Court in Bishwanath Prasad Radhey Shyam v. Hindustan Metal Industries, claims must be read in the context of the complete specification. The EPO requires claims to be clear without reference to the specification. China requires all necessary technical features to appear within the claim itself. These are meaningfully different standards, and a disclosure found insufficient under EPO or Chinese rules may be perfectly adequate under Indian law.
On the term "critical coverage area," the court held it was understandable to a person skilled in the art as the dominant signal area of a sector, essentially the coverage area minus the handover zone, and did not need mathematical precision. "Substantially equivalent" simply meant that the replacement antenna continued to provide the same effective coverage as the antenna it replaced. "Asymmetry" referred to optimised or induced asymmetry along the line of reference, not to uncontrolled asymmetry arising from noise, distortion, or topography.
On the abandoned EPO application, the court noted that CCA's witness had explained the withdrawal was a commercial decision, since the technology is more commercially relevant in population-dense markets like India than in European markets. Rosenberger's senior counsel chose not to cross-examine that witness on the explanation. The court drew an adverse inference.
Other Grounds
Rosenberger argued under Section 64 (1)(i), (j), and (k) of the Patents Act that the invention was a natural phenomenon, a mental act, and involved a false representation to the Patent Office. The court found an internal contradiction in the defendants' own case, arguing that the invention is both a natural phenomenon and a mental act is logically incoherent, since those categories are mutually exclusive. Neither argument was supported by any simulation or independent technical analysis. All revocation grounds were rejected.
Issue Two: Infringement
The Evidentiary Problem
Rosenberger's antennas are customised industrial products, not available off the shelf. Neither party had access to the other's physical antenna during trial. Rosenberger had also removed its product brochures from the internet as the litigation progressed. The court described this withdrawal as "suspect to say the least," and noted that where a party withholds evidence it controls, an adverse inference is warranted.
MATLAB Simulation as Proof
CCA's expert witness, generated beam patterns by inputting the power and phase weightings disclosed in IN'893 into MATLAB software, deriving element spacing from the standard half-wavelength formula at the relevant frequency, and using a cosine function to represent the single element pattern. The resulting beam pattern outputs were overlaid on beam patterns from Rosenberger's admitted product brochures. The overlay showed near-identical patterns.
Rosenberger's expert, stated in cross-examination that element spacing can be derived from the cosine function exactly as PW-1 had done, and admitted that he had himself generated a MATLAB simulation but had chosen not to file it. That admission was decisive. If DW-2's own simulation would have contradicted PW-1's methodology, it would have been filed. Its absence allowed the court to conclude that PW-1's approach was technically sound and that the results were reliable.
The Gillette Defence Collapses
Rosenberger set up a Gillette defence, arguing that its antennas were built using the Butler Matrix design, a well-known prior art architecture dating to 1982, and that any asymmetry in their beam patterns was uncontrolled and accidental rather than intentional. DW-2 testified that the defendants applied the Butler Matrix to the antenna architecture of the Wästberg patent (US 6608591 B2).
Under cross-examination, the said witness admitted that the Wästberg patent is actually based on a Blass Matrix and a Nolan Matrix, not a Butler Matrix. The court extracted the Wästberg patent text and confirmed this. The foundational premise of the Gillette defence, that Rosenberger used prior art technology that predated and excluded the patent in suit, collapsed on the defendants' own evidence.
The Three-Way Comparison Document
The most commercially pointed piece of evidence came from the defendants themselves. Exhibit PW-2/10 was a document Rosenberger prepared for Reliance Jio Infocomm Ltd. comparing the beam patterns of Rosenberger, Ace Technologies Corp., and CCA. The document showed near-identical beam patterns across all three and stated that Rosenberger had optimised its side-lobes and suppressed its back-lobes beyond the standard specified in CCA's antenna.
The court's logic on this evidence is hard to argue with. If Rosenberger's antenna did not use the same technology as CCA's, there was no commercial rationale for preparing a document showing their beam patterns were essentially the same and sending it to a major buyer. The comparison document was made to win a sale, and winning the sale required showing equivalence to the patented product.
Infringement was found across all four features, i.e., split-sector antenna, asymmetrical beam pattern, maintenance of critical coverage area, and reduction in handover zones.
Issue Three: Damages
Why Lost Profits Were Rejected
CCA initially claimed damages on the basis of lost profits, quantifying its loss at approximately USD 140 million on Rosenberger's sales of 1,27,974 antennas. The Indian market for cellular antennas is an oligopoly. There are a small number of sellers and a small number of large telecom operators who buy from them. In that kind of market, a buyer's decision to choose one supplier over another reflects a complex mix of factors like pricing, supply chain reliability, after-sales service, side-lobe performance, relationships, and product range. CCA's own witness admitted in cross-examination that RJIL had communicated a preference for Rosenberger's antenna partly because its side-lobe performance was marginally better, and that pricing played a significant role. In an oligopoly, the court held, infringement alone cannot be assigned as the sole cause of a patentee's lost sales.
Royalty Method and the 20 Per Cent Rate
The court adopted the royalty method, which calculates damages as the royalty the infringer would have paid had it negotiated a licence rather than infringing. The starting point was the CommScope licence agreement, produced under sealed cover. CommScope had previously taken a global licence on the same patent, and the rate in that agreement served as the floor for the court's analysis. The court also noted that Rosenberger had filed no Freedom to Operate analysis before launching its products, which the court treated as evidence of bad faith and a disregard for CCA's patent rights. A bona fide licensee undertakes due diligence, negotiates, and concludes an agreement. A party that does none of that and infringes should not benefit from the rates available to a party that complied.
The court set the royalty rate at 20 per cent of total sales, expressly calibrated above the CommScope rate to reflect the defendants' bad faith conduct.
The Currency Conversion
The admitted sales figures included amounts in both USD and INR. The court applied the exchange rate prevailing on 30 March 2026, the date of the decree, in accordance with the Supreme Court's decision in DLF Ltd. v. Koncar Generators and Motors Ltd. (2024 SCC OnLine SC 1907), which holds that the date of the decree is the appropriate conversion date for foreign currency claims. The arithmetic produced a total damages figure of Rs. 1,52,32,36,783.90, or Rs. 152.32 Crore.
Reliefs Granted
The court granted four categories of relief:
- A permanent injunction was issued, restraining all four Rosenberger entities from manufacturing, using, distributing, selling, offering for sale, or importing into India any product infringing IN'893, covering the eleven identified infringing models and any future models that incorporate the protected technology.
- Damages of Rs. 152.32 Crore were decreed, payable by 30 June 2026, with simple interest at 7 per cent per annum from 1 July 2026 on any default.
- The court directed the defendants to file updated sales affidavits within four weeks, with additional damages at 20 per cent of any further sales to be paid by the same deadline.
- Taxed costs were awarded to CCA.
The court also granted a Certificate of Validity under Section 113 of the Patents Act, 1970, recognising that the patent had been contested and upheld after a full trial on the merits. That certificate carries procedural weight in future enforcement proceedings.
Why This Judgment Matters
The Dartboard Model is the first formal framework an Indian court has articulated for identifying and penalising unfocused prior art challenges. The explicit reference to AI tools generating citation lists is pointed. Any party considering a prior art challenge in Indian patent litigation now has to think carefully about whether the technical connection between its cited documents is clear and defensible, or whether it risks being seen as throwing darts.
The acceptance of MATLAB simulation as legally sufficient proof of infringement resolves a long-standing evidentiary gap. When accused products are not publicly available, when a defendant has removed its own product documentation during litigation, and when physical testing is therefore impossible, a computational simulation built from the patent's own disclosures and validated by the defendant's expert's admissions can discharge the burden of proof. This matters significantly for technology patents where the infringing component is embedded in a larger system and cannot be extracted and tested easily.
The royalty methodology, calibrated to a real licence as a floor with an uplift for bad faith, gives future courts a clear model for damages in oligopolistic industrial markets. The rejection of the lost profits approach in such markets, on the grounds that infringement cannot be isolated as the sole cause of lost sales, is equally important. It pushes damages analysis toward a principled royalty framework rather than speculative revenue comparison. This echoes the evolving approach visible in the SEP damages case involving Philips and Bansal, where the Division Bench similarly required an objective, evidence-based approach to royalty computation rather than assertions untethered from comparable licences.
If you want a broader view of how patent application and rejection procedures work in India, the principles from this post on patent hearing rights are a useful grounding. And for context on where India sits overall in the movement toward serious IP enforcement, the weekly IP digest covers the broader landscape.
Conclusion
The Delhi High Court's judgment in the above case is the most consequential patent damages ruling India has produced. Rs. 152.32 Crore awarded to a patentee who outlasted a fully contested multi-year trial, defeated every revocation ground, and proved infringement without access to physical product samples. That outcome should surprise no one who has followed the trajectory of Indian patent enforcement since the mid-2010s, but the scale and rigour of this particular ruling marks a genuine inflection point.
What strikes me most is not the damages quantum, significant as that is. It is the coherence of the court's reasoning across three separate problems, each of which Indian patent law had not fully resolved before this judgment. The Dartboard Model addresses the real and growing problem of AI-assisted scattershot prior art challenges in a principled way. The MATLAB simulation ruling gives patentees a viable evidentiary path when physical access to an accused product is denied or obstructed. The royalty methodology with a bad-faith uplift gives courts a tool to distinguish between infringers who ignored patent rights and parties who made genuine licensing efforts.
For foreign companies holding technology patents in India's telecom infrastructure market, this judgment removes any lingering doubt about whether Indian courts will actually enforce those patents with meaningful financial consequences. The absence of due diligence is not a neutral fact. It is evidence of bad faith, and it costs more.
Click here to read the basics of trademark law and copyright law.
~ Adv. Koushik Chittella
© All Rights Reserved
Disclaimer: None of the contents of this post constitute legal advice.