<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Ajay Kumar on Medium]]></title>
        <description><![CDATA[Stories by Ajay Kumar on Medium]]></description>
        <link>https://medium.com/@ajay.gahlawat?source=rss-c8a643ac61aa------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*7dPABXl4VsP_yIbY.</url>
            <title>Stories by Ajay Kumar on Medium</title>
            <link>https://medium.com/@ajay.gahlawat?source=rss-c8a643ac61aa------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 01:24:07 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@ajay.gahlawat/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[I Hold 25 Patents. Here’s Why I Want the Patent System Destroyed.]]></title>
            <link>https://medium.com/@ajay.gahlawat/i-hold-25-patents-heres-why-i-want-the-patent-system-destroyed-32f8805d6c4c?source=rss-c8a643ac61aa------2</link>
            <guid isPermaLink="false">https://medium.com/p/32f8805d6c4c</guid>
            <category><![CDATA[innovation-management]]></category>
            <category><![CDATA[patents]]></category>
            <category><![CDATA[vc]]></category>
            <category><![CDATA[copyright]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Ajay Kumar]]></dc:creator>
            <pubDate>Fri, 24 Apr 2026 06:16:36 GMT</pubDate>
            <atom:updated>2026-04-24T06:16:36.206Z</atom:updated>
            <content:encoded><![CDATA[<h3>A 25-year tech insider explains why AI innovation and intellectual property law are on a collision course — and why the only way forward is a radical reset.</h3><p><em>Part One of “The IP Squish: Why AI Innovation Demands a Hard Reset”</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tXzGfi0yYSdBnV1s6sB6AA.png" /></figure><p>I hold 25 United States patents.</p><p>I’ve spent 25 years building technology at Sun Microsystems, Microsoft, Adobe, Yahoo, and many other major companies. I’ve led engineering teams of 100 people, run high-transaction ML systems processing 100,000 requests per second, and managed technology transformations across media, hospitality, education, and enterprise distribution.</p><p>By almost any measure, I should be a staunch defender of the intellectual property system. I’ve benefited from it professionally, strategically, and often financially.</p><p>I’m here to tell you why it’s broken.</p><p>And not broken in the “needs some tweaks” sense. Broken in the “fundamentally designed for a world that no longer exists” sense.</p><h3>The Speed Mismatch</h3><p>The IP regime we have — 20-year patents, 70-year copyrights — was built for a world where ideas were scarce and human effort was expensive. We traded temporary monopolies for long-term public good because we needed to lure capital into the grueling, decade-long slog of R&amp;D.</p><p>But AI has broken the “effort” side of that contract.</p><p>Here’s what this looks like on the ground. I ran one of the most sophisticated advertising technology systems in the country. My team and I operated ML pipelines that process over a billion events daily. We iterate models from prototype to production in two weeks. We deployed updates to our systems twenty times a week. We ran 30+ concurrent A/B tests at any given time.</p><p>This is normal in the AI era.</p><p>Now consider the patent timeline. From the moment you file a patent application to the moment it issues and becomes enforceable, you’re looking at 1–2 years.</p><p>By the time an AI-derived invention actually gets patent protection, the invention is obsolete.</p><p>I’ve seen this play out across my entire career. At Microsoft, I built the first-party health and fitness application pre-deployed on Windows 7.5. At Microsoft’s Bing, I leveraged machine learning and information retrieval techniques to compete in local search. At various companies, I’ve built horizontal practices for UX, Machine Learning, Data &amp; Analytics.</p><p>In each case, the innovations that mattered weren’t protected by patents. They were protected by speed.</p><h3>What the Patent System Actually Costs</h3><p>Let me share what it costs to build and maintain a patent portfolio, because most people don’t see the inside of this.</p><p><strong>Filing a single patent:</strong> $20,000–$50,000 when you include attorney fees, international equivalents, and the time cost of engineers and architects who spend 2–4 weeks per patent in documentation and review.</p><p><strong>Portfolio maintenance:</strong> $15,000–$30,000 per year for a substantial portfolio.</p><p><strong>Litigation defense:</strong> $2M–$5M per lawsuit (on the defense side).</p><p>But here’s what the ROI analysis doesn’t account for: the engineering drag.</p><p>The collective groan that goes through a team when a “patent disclosure meeting” appears on the calendar.</p><p>The friction between legal and engineering over what can be open-sourced.</p><p>The “not invented here” syndrome that takes root because teams can’t use competitors’ IP.</p><p>I’ve encountered this balance across multiple organizations — at Pitney Bowes building horizontal practices, at RateGain leading technology verticals, at ClassGrid and 9.9 Media across education and media.</p><p>The pattern is the same everywhere: the administrative overhead of the IP system is cannibalizing the innovation it’s supposed to protect.</p><h3>The M&amp;A Room Reality</h3><p>Perhaps the most sobering view comes during acquisitions, where I’ve integrated technology teams for US acquisitions. The IP audit is the number one deal-breaker after financial terms.</p><p>And here’s what I’ve learned: patent portfolios are dramatically overvalued.</p><p>In my experience, <strong>5% of patents drive 90% of enterprise value</strong>.</p><p>The rest is window dressing — vanity metrics like “total patent count” and “international coverage” that look good on slides but create negligible strategic value.</p><p>And now there’s a new wrinkle: the AI attribution problem. When we’re conducting due diligence on a target company, we’re starting to ask a question no one has good answers for: “What AI-generated content is in your codebase? Do you have rights to use the training data? Who actually owns this code?”</p><p>The IP system can’t answer this question because it was never designed for AI.</p><h3>Portfolio Economics: The Math Doesn’t Work</h3><p>The economics are even worse when you run the numbers on AI-era invention.</p><p><strong>Pre-AI era calculation:</strong></p><ul><li>$1M research and development investment</li><li>$10M patent value</li><li>10x return on investment</li></ul><p>This made sense. The system rewarded substantial, sustained innovation with a reasonable monopoly period.</p><p><strong>AI era calculation:</strong></p><ul><li>$10K research and development investment (AI generates solution)</li><li>$1M patent value (same market value)</li><li>100x nominal return on investment</li></ul><p>But here’s the catch: to maintain a substantial portfolio in the AI era, you’re filing 10x as many patents. That $10K per invention becomes $100K in portfolio maintenance costs. You might hold 100 patents to protect your foundational innovations.</p><p>So: 100 inventions × $10K filing costs = <strong>$1M in filing costs</strong> for the same $10M in patent value. Your return drops from 10x to negative.</p><p>The math doesn’t work.</p><p>The IP system was designed when invention was expensive. It breaks when invention is cheap.</p><h3>The Foundation Pattern Recognition</h3><p>I’ve noticed a consistent pattern across every company where I’ve worked, especially in foundation-building periods:</p><p><strong>The Foundation Phase:</strong></p><ul><li>Early innovation generates significant patent activity</li><li>Patents provide strategic moats against competitors</li><li>IPO valuation heavily weighted by IP portfolio</li><li>Patent portfolio becomes business valuation anchor</li></ul><p><strong>The Velocity Phase:</strong></p><ul><li>Innovation speed dramatically exceeds patent protection timelines</li><li>Companies that cling to IP protection fall behind</li><li>The patent portfolio becomes a liability rather than asset</li><li>Competitive advantage shifts from IP to execution</li></ul><p>I’ve seen this at Sun Microsystems where web-scale infrastructure innovations became standard practice after patents expired. I’ve seen this at Microsoft where search and information retrieval patents provided limited practical protection. I’ve seen this across multiple acquisitions where due diligence revealed patent portfolios as vanity metrics, not actual competitive advantages.</p><p>The pattern is consistent: patents matter for foundation-building. They actively harm velocity-building.</p><p>AI moves us permanently into the velocity phase.</p><h3>The AI Attribution Nightmare</h3><p>The new frontier of IP problems is even more fundamental: when AI is involved in invention, who actually owns the patent?</p><p>This isn’t a theoretical problem. It’s happening right now in companies across Silicon Valley. Teams using GPT-4, Claude, and other AI tools to generate code, designs, and technical solutions. The question that keeps general counsel up at night: if your core patent was invented by AI, do you actually own it?</p><p>The current legal answer is a mess. Some cases suggest AI-generated content can’t be copyrighted. Others suggest it falls into gray areas. The US Patent Office has issued guidance that requires human authorship but doesn’t provide clear standards for mixed human-AI invention.</p><p>The result is paralysis. Companies are unsure whether to file patents on AI-assisted innovations. Legal teams are creating complex documentation requirements for every AI interaction. Engineering teams are frustrated by the uncertainty and bureaucracy.</p><p>The IP system has created an attribution crisis at exactly the moment when AI becomes the primary invention engine.</p><h3>What Comes Next</h3><p>This brings us to the core question: if the current IP system is broken, what replaces it?</p><p>In the next article in this series, I’ll propose seven concrete alternatives — models designed for AI-era innovation. Some are radical. Some are evolutionary. Some build on frameworks that already exist.</p><p>The Singapore Certificate of Entitlement system for automobiles. Bitcoin’s supply control. The Longitude Prize approach. Netflix-style licensing for patents.</p><p>Each addresses a different piece of the IP reform puzzle. Together, they point toward what I call “the IP squish” — a systematic compression of IP protection that aligns with AI innovation realities.</p><p><strong>The future doesn’t demand stronger IP protection. It demands smarter IP squish.</strong></p><p><em>This is Part One of “The IP Squish” series. Part Two, “Seven Models for AI-Era Innovation,” explores concrete alternatives to the broken patent system.</em></p><h3>About the Author</h3><p><strong>Ajay Gehlawat</strong> holds 25 US patents and has spent 25 years building technology at Sun Microsystems, Microsoft, Adobe, Yahoo, and other major companies. He’s led teams of 100+ engineers, run ML systems processing 1B+ events daily, and managed technical acquisitions across multiple industries. He currently operates sophisticated ad tech systems at scale. He writes about technology, innovation, and the intersection of AI and business.</p><p>Follow for more insights about AI innovation, intellectual property reform, and technology strategy.</p><p><strong>Liked this article?</strong><br> • <strong>Clap</strong> if you found this valuable 👏<br> • <strong>Follow</strong> for Part 2: Seven Models for AI-Era Innovation<br> • <strong>Share</strong> with someone interested in AI and innovation policy</p><h3>Tags for Medium Algorithm</h3><p><a href="#ArtificialIntelligence">#ArtificialIntelligence</a> <a href="#IntellectualProperty">#IntellectualProperty</a> <a href="#TechPolicy">#TechPolicy</a> <a href="#Innovation">#Innovation</a> <a href="#Patents">#Patents</a> <a href="#AILaw">#AILaw</a> <a href="#Technology">#Technology</a> <a href="#BusinessStrategy">#BusinessStrategy</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=32f8805d6c4c" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>