Karthick Iyer

How does a startup accelerate development with limited time and resources? At Akeana, co-founder Nitin Rajmohan reveals how Cadence's like Xcelium, Verisium Debug, and Palladium platforms helped them go from simulation to emulation seamlessly, speeding up their development cycle and delivering high-quality IP solutions. Read more: https://ow.ly/20nE30sP3aO

8

As part of the IndiaAI Mission, Avataar is building UAI (Unified AI Interface) - a foundational DPI layer with thin, durable rails for trust and interoperability, connecting models, data and compute into a cohesive, discoverable ecosystem. We are privileged to collaborate with Prof. Ramesh Raskar (MIT) and Project NANDA: Architecting the Internet of AI Agents in shaping a broader “Internet of AI Agents” vision - exploring how a discovery layer for India’s evolving digital public infrastructure can unlock openness, innovation and responsible scale together. Equally grateful to Aafreen Siddiqui Sherwani (United Nations) for anchoring the dialogue, and to Pradeep Gupta (NVIDIA), Tanvi Lall (EkStep Foundation), Vineet Singh (Digital Green) and Jibu Elias (Mozilla Foundation) - for a thoughtful, forward-looking exchange on what it truly takes to build open AI platforms for India. The real opportunity ahead is not just technical, it is architectural. How we design these rails now will shape how intelligence compounds across citizens, enterprises and governance systems for decades. What principles should guide the design of AI infrastructure at national scale? #IndiaAI #DigitalPublicInfrastructure #ViksitBharat

183

9 Comments

*** Cross-die links introduce skew, coherence issues, and latency that don’t appear in monolithic SoCs. *** Chapter 5 of my book on chiplets details die-to-die interconnects (UCIe, BoW, AIB, and others) and the hooks needed to validate them. Now available on Amazon in hardcover, paperback, and Kindle. Links in the comments.

19

1 Comment

Efficient LLM Inference: Bandwidth, Compute, Synchronization, and Capacity are all you need In the world of large language models (LLMs), memory bandwidth and capacity are pivotal for efficient inference workloads. Discover how optimizing these elements can significantly boost per-user throughput and drive system-level efficiency. Key Takeaways: - Serving LLMs demands 100s of GB per server for model instances, highlighting the importance of memory capacity. - High memory bandwidth is crucial for maximizing throughput. - DRAM-based designs offer superior efficiency in terms of cost and power usage. These insights, drawn from a comprehensive study by NVIDIA Research, highlight the core performance bottlenecks and potential hardware advancements that can redefine LLM deployment strategies. Read the full article here: https://lnkd.in/gNhFm8Ki #AI #LLM #MemoryBandwidth #InferenceWorkloads

18

1 Comment

On-device AI becomes powerful only when it works reliably on real hardware. In this Melange Walkthrough, Yeonseok Kim and Sun-Q Kim break down how the full on-device pipeline works, from model selection and quantization to real-device benchmarking and SDK deployment. Select. Benchmark. Deploy. This is how production-ready on-device AI should operate. Watch the full walkthrough below. #OnDeviceAI #EdgeAI #MobileAI #AIInfrastructure #ZETIC #Melange

26

I am excited to welcome Sankara Venkateswaran to the Ayar Labs team as our new VP of Engineering. As we move into 2026, the industry is hitting a critical inflection point. The "copper wall" is a real-world bottleneck for hyperscale AI, and co-packaged optics has become an infrastructure necessity in future roadmaps. Sankara joins us at a pivotal moment as we move towards mass production. His deep expertise in  compute and networking acceleration silicon from his time at Intel, Qualcomm, and Atheros perfectly complements our leadership in co-packaged optics. He understands the discipline required to deliver production-grade silicon that integrates into the world's most complex packaging flows. With Sankara leading our engineering organization, we are doubling down on our commitment to: - Scale our TeraPHY™ optical engines for high-volume deployment. - Break the power and bandwidth barriers of traditional interconnects. - Deliver predictable, high-quality CPO solutions that meet the aggressive timelines of our partners for the delivery of hyperscale AI infrastructure. Please join me in welcoming Sankara to the team!

89

6 Comments

From "General-Purpose GPUs" to "Custom Silicon": A Hardware Inflection Point Indian Startups Shouldn’t Miss Broadcom recently announced it has received over $10 billion in orders for custom AI racks and accelerators. That’s not just a procurement milestone. This marks a decisive shift in AI infrastructure—from broad-spectrum GPU reliance to purpose-built silicon engineered for high-efficiency, scalable performance. This surge reflects growing demand from frontier AI developers like OpenAI and others building large-scale foundation models—who increasingly require specialized infrastructure that goes beyond general-purpose GPUs to meet the scale, efficiency, and latency needs of next-generation training and inference. General-purpose GPUs, while versatile, are increasingly being outpaced by custom silicon when it comes to inference-heavy workloads. ASICs and other specialized architectures offer better power efficiency, reduced latency, and tighter integration for predictable, high-volume operations. And Broadcom isn’t alone. I see, Google continues to evolve its Tensor Processing Units (TPUs). Amazon has deployed Inferentia and Trainium across its cloud stack. Microsoft is building its own Azure Maia and Cobalt chips. There are startups that are reimagining compute from the ground up—focusing on memory bandwidth and wafer-scale innovation. But my main point is, this shift opens a rare window for Indian startups. With the right partnerships and design focus, there’s a clear opportunity to: - Build domain-specific silicon for EV, Automotive, AgriTech, Space, wearables, audio, and other edge AI domains - Leverage fabless models and global foundries to bypass capital-heavy barriers - Create IP-led differentiation in smart devices. There’s no need to replicate legacy architectures. The real opportunity lies in out-specializing them—especially in segments they won’t prioritize. This isn’t just a hardware story. It’s a blueprint for deep-tech sovereignty. Let’s not just observe the AI infrastructure race. Let’s build lanes of our own. Thoughts? #AIHardware #CustomSilicon #IndianStartups #SemiconductorIndia #DeepTech #StrategicAutonomy #TechLeadership #FutureProofIndia

40

2 Comments

https://lnkd.in/gf9PRggT 🚀 Understanding the Hidden Impact of Onboarding Failures Many product launches face a silent setback: customers who try but then disappear. Months of development can be lost in a confusing first experience, impacting revenue and growth. Building well-crafted onboarding strategies turns initial confusion into a structured plan for success. 💡 Key Insight Effective onboarding can boost activation and willingness to pay by Between 12 and 21. Most users delete apps because they do not understand how to use them. Clear, helpful first impressions help retain 86 of customers. 🔍 How to Evaluate Onboarding Solutions Use tools with built-in analytics, segmentation, and targeted messaging. Look for features that measure behavior, automate engagement, and integrate knowledge bases. Invest in platforms that provide measurable impact and continuous improvement capabilities. ✅ Key Recommendations - Simplify first-run guidance to reduce churn. - Use checklist-driven flows for high completion rates. - Select platforms with analytics and feedback loops. - Deliver personalized guidance to increase engagement and revenue. - Before purchasing, evaluate pricing, security, and review real-world case studies. 🎯 Who Should Use This Guidance Founders, product teams, growth managers, and customer success leaders aiming to improve activation and adoption. Teams deciding whether to build or buy will find clear criteria for fast deployment with minimal engineering. Larger enterprises benefit from governance, integrations, and centralized insights. 🔑 Critical Success Factors - First-time clarity shortens the path to value and reduces drop-offs. - Checklists and guided flows improve notable adoption. - Platforms that analyze actions and enable automation yield continuous improvements. - Personalization elevates engagement levels and customer willingness to invest. 📊 Why Onboarding Matters for Long-Term Success Early user guidance influences long-term loyalty through faster time to value. Shortening the activation journey increases feature adoption, retention, and revenue growth. Teams that track these metrics explicitly can see direct benefits such as increased customer expansion and higher engagement. 🚧 Recognizing User Drop-offs and Confusion Analytics reveal where users stall — signups, first steps, or feature underuse — guiding targeted interventions. Many users delete because they do not know how to proceed, making interactive checklists and contextual guidance essential. 🤖 Modern Onboarding Technologies 🛠️ Integrating AI-driven guidance, behavioral analytics, and targeted messaging creates a connected growth engine. This stack continuously interprets signals, adjusts paths, and personalizes experiences at scale. Significant components include guidance platforms, analytics tools, and communication channels.

2

Samsung Semiconductor India Research is transforming chip design with Cadence Cerebrus AI Studio, achieving a PPA improvement of 8% to 11% on SoC subsystems. By utilizing features like Smart Model Replay and advanced data analytics, they've enhanced productivity and streamlined workflows. Discover the future of efficient chip design! https://ow.ly/RyNa30sMrSy

41

How much of the AI hype is real vs fad? Decide for yourself: https://lnkd.in/gjnanw85 is a sharp look at AI’s limits in predicting human behavior. From https://lnkd.in/gst_HSeE (full paper at https://lnkd.in/gh5NUzfz): LLMs Struggle to Use Novel Insights for Problem Solving AI models performed better on knowledge-heavy problems—ones that can be solved by stitching well-known templates, as the requisite problem-solving patterns appear ‘verbatim in training data’. Even on logic-heavy problems, which require a patterned way of thinking, these models performed well. However, they performed poorly on observation-heavy problems, whose solutions hinge on the discovery of novel insights — “something that cannot be retrieved from memorised snippets alone”.

7

Really important announcement from Qualcomm today which sees my former colleagues drive their impressive edge AI capability into the cloud. Two critical and potentially industry-changing items stand out for me: > No HMB - this system uses LPDDR for lower cost and increased capacity > Architecture unifies edge and cloud with same NPUs now driving Qualcomm AI in phones, PCs, Automotive and data-centre. Well done Q! Link to News article: https://lnkd.in/eMbbHcJH

18