About
I’m driven by the desire to build delightful products - elegant, genuinely helpful…
Articles by Avinash
-
Design to Cost: Why 80% of Your Product Cost Is Decided Before You Prototype
Design to Cost: Why 80% of Your Product Cost Is Decided Before You Prototype
We spend months optimizing supply chains……but the truth is, by then 70–95% of your cost is already committed in…
4 Comments -
CQL - Community Qualified Leads
CQL - Community Qualified Leads
Communities are the new growth frontier and at the helm of Product-led & Developer-led growth. However, quantifying the…
2 Comments -
Building Meaningful Engagements in this Low Touch Economy
Building Meaningful Engagements in this Low Touch Economy
The dynamics of building & nurturing relationships has changed forever. Doing business remotely is becoming a norm and…
1 Comment -
Developer Engagement as a Growth Strategy
Developer Engagement as a Growth Strategy
In the new API & Platform economy, the real consumer & influencer in buying Tech products is a developer or a technical…
4 Comments -
The 7S Framework for Enterprise BaaS (Backend-as-a-Service)
The 7S Framework for Enterprise BaaS (Backend-as-a-Service)
As Mobility becomes an integral part of the business & technology strategy, an mBaaS (Mobile Backend-as-a-Service) or…
6 Comments -
Leveraging mobile to improve customer service – A checklist
Leveraging mobile to improve customer service – A checklist
Customer service has always been an important factor contributing to the success of a business, and the realization…
3 Comments
Activity
10K followers
-
Avinash Harsh shared thisAI is already changing how software gets built. Hardware has a different kind of complexity. Kudos to Benjamin Chia and Hardware FYI team for bringing together Kinetic — a gathering for the new wave of hardware thinkers, operators, and builders. Rajesh, Jason Trachewsky, and I will be there from Wizerr AI. Every hardware product is ultimately shaped by thousands of component decisions. At Wizerr, we’re building the AI-native decision intelligence layer behind those decisions — connecting design, engineering-grade component intelligence, lifecycle, sourcing, and supply signals into one continuously evolving intelligence graph. If you’re attending Kinetic next week and working on AI tooling, datacenter hardware, robotics, edge systems, or complex electronics, a few of us are planning to connect around a small informal 'Hardware AI Huddle' on May 13. DM me if you’d like to join the conversation https://lnkd.in/gDW-7PGt
-
Avinash Harsh shared thisSpeed is table stakes. Compounding is the moat. Breakout AI-native companies won’t just ship faster - they’ll build systems that improve continuously. Systems that learn from usage, refine decisions, and get harder to compete with over time. Sharing this in case it’s relevant. 👇 AIBoomi (formerly SaaSBoomi) is hosting a Bootcamp in the Bay Area (May 18–20). Small, curated group (~50 AI-native founders). Founder-led. High signal. No fluff. Format is intentionally tight — 3 focused afternoons: • May 18: GTM + dinner • May 19: OpenAI Codex Day (at OpenAI HQ) • May 20: 0→1, fundraising + open sessions No conference noise. Just builders. Details: https://lnkd.in/g7wHwDip If this feels like your kind of room, or someone you know should be in, please apply below. Avinash Raghava Sandeep Todi
-
Avinash Harsh shared thisWe crossed 1,000,000 component inferences on Wizerr AI this month. Not searches. Not parametric table lookups. Each one: 20–30 engineer-grade passes. ↳ Electrical match ↳ Mechanical form-fit ↳ Package geometry ↳ Multi-source availability ↳ Price checks ↳ Datasheet-level reasoning, not a database query The kind of judgment a senior engineer applies before signing off on a part. Running in seconds. A million times over. — We're not a distributor. We're not a database. We're the intelligence layer that brings engineering constraints, supply chain reality, and design tradeoffs into a single pane of glass, so the teams building hardware make better decisions, faster. — Something significant is coming this month.
-
Avinash Harsh shared thisAlongside NVIDIA GTC, we introduced Decision API from Wizerr AI An intelligence layer for hardware, now accessible via API. Most APIs return data. This returns decisions. For OEMs: → best alternate - and why → multi-source entire BOM - with my preferred suppliers → optimize for functional depth + cost + availability For Suppliers: → OEM request in → my best alternates + why, in minutes Evaluates like an engineer. Returns decision-grade outputs with reasoning. Already powering internal AI initiatives, engineering and supply chain workflows at stellar teams. Now accepting the next set of partners. A short form is in comments. We’ll review and reach out.
-
Avinash Harsh shared thisAI is built on GPUs. Hardware programs are built on decisions. Data center Hardware. Robotics. Edge. Aerospace. Components. Supply. Cost. Risk. Power budgets. Thermal limits. Design constraints. Today, we’re opening: 𝐃𝐞𝐜𝐢𝐬𝐢𝐨𝐧 𝐀𝐏𝐈𝐬 for hardware programs. Not data. 𝐃𝐞𝐜𝐢𝐬𝐢𝐨𝐧𝐬. Request early access: https://lnkd.in/gnk65gwE Learn more https://lnkd.in/gg5_xEEE Wizerr AI #GTC2026 #AIforHardware
-
Avinash Harsh shared thisAI runs on GPUs. The physical world runs on decisions. At NVIDIA GTC this week. As AI moves into real-world systems, a lot of what determines outcomes comes down to hardware programs - and the decisions behind them: • component choices • alternates • lifecycle and supply signals • cost-performance tradeoffs Wizerr AI team is around this week connecting with engineers, builders, and hardware teams thinking about this layer. If you’re at GTC this week and thinking about how AI connects to real-world systems, would love to compare notes. (Something we’ve been working on - more on Thursday.) #GTC2026
-
Avinash Harsh shared thisThe next intelligence layer in hardware is emerging across the lifecycle! Just back from DesignCon 2026 - sharing some observations. Beyond discussions around 224G signaling, AI infrastructure, advanced packaging, and SI/PI, a deeper signal stood out. As hardware systems grow more complex, new opportunities are emerging to bring intelligence into the decisions that connect every stage of the hardware lifecycle. A few shifts that stood out. 1. Hardware complexity is creating a new intelligence-layer opportunity. Modern hardware programs span chip design, packaging, SI/PI, board design, manufacturing readiness, sourcing, and lifecycle management. Each domain runs on specialized platforms from Cadence, Siemens EDA, and PCB workflows like Altium® and they are going their own AI transformations. The emerging opportunity is connecting decisions across the lifecycle. 2. AI-native engineering platforms are emerging. A new generation of AI-native companies such as Circuit Mind is helping engineers explore and evaluate design decisions earlier in the process. And companies like Wizerr AI, where we are building the decision intelligence and inference infrastructure for entire hardware programs, helping teams reason across design, sourcing, and lifecycle decisions. 3. Signal integrity and power integrity are now system constraints. At 224G and beyond, SI and PI span silicon, packaging, board, and interconnect. Platforms like Synopsys Inc are helping engineers model these system-level interactions. 4. AI infrastructure is reshaping hardware roadmaps AI clusters led by platforms like NVIDIA are pushing bandwidth, interconnect density, and power delivery to new limits. 5. Packaging is becoming a performance frontier With chiplets and 2.5D/3D integration, performance increasingly depends on the packaging layer. Leaders like Intel Corporation and AMD are pushing advanced packaging to the center of system design. 6. Interconnect ecosystems are becoming critical infrastructure Companies like TE Connectivity and Amphenol power the physical backbone of AI systems, where high-speed connectors and cables must preserve signal integrity at extreme data rates. 7. AI implementation in semiconductor companies may follow hybrid models An interesting point raised by Dr. Olena (Jianfang) Zhu from Intel Corporation was that semiconductor design data is extremely sensitive. Because of that, many companies may end up using hybrid approaches combining smaller private models with carefully controlled external AI capabilities. 2026 may turn out to be a pivotal year for hardware. The next phase of hardware innovation will depend on how intelligently we navigate decisions across the lifecycle. At Wizerr AI, we are building the intelligence and inference layer for hardware programs, connecting decisions across design, sourcing, and build. #AiforHardware
-
Avinash Harsh shared thisDesignCon 2026 ⚡ Hardware development is entering a new phase. Decision Intelligence. Continuous inference embedded directly inside engineering and sourcing workflows. Built to help hardware teams make faster, smarter, trusted decisions. At Wizerr AI, we are building an AI-native intelligence layer for electronics. One that reasons across engineering specs, pin diagrams, graphs, lifecycle, cost, and supply in real time. What does that look like in practice? BOM optimization is entering a new phase. Engineering grade intelligence that evaluates every part decision in context. FFF alignment. Cross references. Lifecycle exposure. Supply and cost impact. Reasoned together, not reviewed separately. That is the shift. If you are at DesignCon, come meet us at booth 1266 (Aisle 1200). Let’s add a teammate that helps you move faster and focus on the real engineering work. #AIforhardware #designcon
-
Avinash Harsh shared thisDesignCon brings the hardware ecosystem together. This year may be a defining year for AI in hardware. Tonight we’re hosting a small private dinner with engineering and supply chain leaders on 'AI for Hardware Programs' - separating signal from noise. A strong group. Candid conversation. Excited to compare notes on how engineering and supply chain teams are operationalizing AI this year. If you’re at #DesignCon this week, Wizerr AI team is continuing the conversation at Booth 1266 in Hall B. #AIinHardware #DesignCon #HardwarePrograms
-
Avinash Harsh liked thisAvinash Harsh liked thisThe world around us relies on hardware engineering, from cars and nuclear reactors to the device you’re reading this on. Yet the system it was built with is still stuck in the dark ages. Walk into one of the best hardware engineering organizations in the world (think Space X, Tesla) and you will find teams of brilliant engineers manually moving between Excel spreadsheets, CAD tools, simulation environments, PDFs, and compliance documents. It's extremely manual. While software development has undergone a dramatic transformation through AI-assisted coding and autonomous agents, physical engineering has largely been left behind. As a mechanical engineer himself, this problem drew Pari Singh to develop Flow Engineering. With Flow, we are entering a world where intelligent agents will not just write software, but also design mechanical systems, run simulations, validate requirements, coordinate testing, manage regulatory workflows, and integrate across the full hardware development lifecycle. The cost of software creation collapsed once engineers could delegate implementation to AI systems. The same transition is now coming for the physical world. Flow is built for this shift. It exists to eliminate the operational bottlenecks between engineering intent and execution by creating an intelligence layer across complex hardware programs. It enables hardware engineering organizations to use agents to do the vast majority of their engineering work. The result is a future where building a rocket, humanoid robot, or energy system starts to look less like traditional hardware engineering and more like programming software. This incredible value proposition has led Flow to partnerships with companies like Rivian, and Flow is now looking to scale their engineering team drastically in the next few months. If you’re interested in changing how the physical world is built, come meet Pari at the next Founders You Should Know showcase! https://luma.com/zdy79tpp
-
Avinash Harsh reacted on thisAvinash Harsh reacted on thisI rarely get nervous. But I'll admit I have been a bit nervous for this post. LinkedIn is so great for all of us sharing our intellectual, career and "shiny" sides. And if anyone here has seen my shoes or outfits, I definitely have a shiny version of me! It's real and I love it. I have never been more motivated, career-wise, health-wise, as a friend or as a father than I am right now. But we all have other parts of us too. After the death of my daughter in February, I learned this first hand in connecting with hundreds of people who have experienced loss. Loss is a (regrettable) part of us but, as Anderson Cooper says, loss is the most common human experience. So in a highly un-shiny direction, I started a substack about my own processing of several losses across my personal life. Writing the first post definitely helped me get more clarity on what I was feeling. So with no further ado, I present to you the way Humpty Dumpty and the Japanese art form of kintsugi fit together in my strange mind: https://lnkd.in/gkh9gapQ
-
Avinash Harsh liked thisAvinash Harsh liked thisWhere are VCs actually writing seed checks right now? I analyzed dozens of seed rounds from top VCs over last few months. The breakdown: 1. Vertical AI agents in legacy Industries: 38% 2. AI infrastructure and agent platforms: 22% 3. Frontier deep tech: 16% 4. Consumer AI and creator tools: 12% 5. Fintech and capital markets: 6% 6. AI-Native cybersecurity: 6% What's missing is just as interesting. Horizontal AI agents (assistants, search, etc) and department-level agents (coding, customer service, IT, finance, etc) that are already well funded at this point in the cycle. Climate tech, crypto, biotech, and SaaS have fallen out of favor. And models or robotics? Well, most of those rounds are $100M+...are they really seeds?
-
Avinash Harsh liked thisAvinash Harsh liked thisB300 behemoth. Standing next to a Blackwell GPU rack is both intimidating and exciting at the same time. A few years ago, seeing a 50 kW rack in a data center was considered dense. Today, a fully loaded Blackwell rack can: • Weigh more than 3,500 lbs • Consume well over 150+ kW of power • Circulate massive amounts of liquid cooling continuously through the system • Push enormous east-west network traffic across ultra low latency fabrics • Deliver compute performance that previously required entire rows of infrastructure Standing next to these systems makes you realize how fast this industry is changing. What used to take an entire data hall is now compressed into a single cabinet powered by GPUs, networking, thermals, power engineering, and software orchestration all operating together in real time. Excited to welcome another addition to our growing Blackwell family at DigiPowerX as we continue building next-generation AI factories and AI infrastructure. The scale of compute being deployed today is unlike anything the industry has seen before. #jjsmusings #Blackwell #AIInfrastructure #GPU #AIFactory #GPUaaS #AIDataCenter #NVIDIA #DataCenters #DigiPowerX #SMCI
-
Avinash Harsh liked thisAvinash Harsh liked thisIf you are in San Francisco next week, product design engineer Sam Holland will be presenting something awesome at Hardware FYI's Kinetic Conference! If you're in the area, swing by Studio 45 on Monday, May 11 for the pre-conference party they're hosting with SendCutSend! Link to RSVP 👇 in the comments.
-
Avinash Harsh reacted on thisAvinash Harsh reacted on thisBig news from the a99 family. 🎉 Asha Venugopal has been promoted to Partner at a99 VC. Asha has been with us since the very beginning: since the first cheques, the first conviction calls, the first bets on Indian manufacturing when not everyone was paying attention. She came in with 20+ years of experience across investment banking and consulting, someone who has spent years stress testing businesses, reading between the lines of a P&L and backing conviction. She knows the difference between a business that looks good on paper and one that can actually be built. That rare combination of investing rigour and on the ground execution instinct? That's just Asha. Beyond the deals, Asha has been deeply hands on with our portfolio, rolling up her sleeves with founders to help them with execution and not just capital. She has also been one of the most important people in shaping our young team, mentoring the next generation of investors at a99 and raising the bar for what good looks like. Our founders trust her instincts. Vignesh will tell you she's the boss. We won't argue. We announced this at Foundry, our closed door event with DBS Bank and it felt like exactly the right moment. Because Foundry is about backing builders and Asha is one of ours. Fund III just got stronger. Watch this space. Welcome to the partnership, Asha. This one's been a long time coming. #a99 #Foundry #TeamNews #VentureCapital #IndiaManufacturing #FundIII
-
Avinash Harsh liked thisAvinash Harsh liked thisThe AI buildout has largely treated community resistance as a permitting nuisance; it's now an operating problem. Data center project cancellations quadrupled from 2024 to 2025, and resistance isn't slowing down. Industry can either strong-arm communities and hope for the best, or it can take the pragmatic approach: find the solutions to the pushback that are economically viable, and in instances, venture-backable. Here's what's central to the resistance: On energy: wholesale electricity costs 267% more near major data center clusters than five years ago. Under existing rate structures, the cost of new grid infrastructure gets socialized across all ratepayers. On water: the same mechanism is playing out, just more quietly. Serving data center water demand through 2030 will require $10B-$58B in new municipal infrastructure. Columbus, Ohio already raised water rates 18% this year in part because of it. Many are also planned to be built in already water-distressed regions (like my home state of Utah). This isn't an argument against building. It's an argument for building differently. For us at Dynamo, that means backing founders working on the infrastructure layer underneath (e.g. advanced cooling, waste heat recovery, energy sourcing & management) that make data centers cheaper to run and easier to build. The AI wave isn't going anywhere, and neither is the backlash. For the right builders, that tension is the opportunity. For more, see the link below. And if you're building a solution in this space, we want to hear from you! hello@dynamo.vc
Patents
-
System and method for multi-channel delivery of transformed and augmented messages in real-time
Issued US 9,491,125
See patentIn a system of interconnected enterprise apps, a business data object maintained by a server undergoes a life cycle event which triggers a message notification.
All messages between client applications and the server are intercepted, and a configurable rules engine and message processing filters govern the transformation and delivery of each message according to each recipient's role, application, and login state.
Messages can be enhanced by adding contextual information and…In a system of interconnected enterprise apps, a business data object maintained by a server undergoes a life cycle event which triggers a message notification.
All messages between client applications and the server are intercepted, and a configurable rules engine and message processing filters govern the transformation and delivery of each message according to each recipient's role, application, and login state.
Messages can be enhanced by adding contextual information and details from other messages and/or information retrieved from enterprise back-end systems according to the rules. The transformation and delivery of messages occurs in real-time across multiple channels, platforms, and users.
Honors & Awards
-
Stanford GSB LEAD: Intellectual Contribution Award - Business Model Design & Analysis
Stanford GSB
The Stanford LEAD Program: Corporate Innovation
is pleased to recognize AVINASH HARSH
on being the top nominee by their peers for having a profound
influence on the learning experience of their colleagues in the
Business Model Analysis and Design.
Recommendations received
2 people have recommended Avinash
Join now to viewOther similar profiles
Explore more posts
-
Michael Skarzynski
NVIDIA Omniverse is a platform of APIs, Services, and Software Development Kits (SDKs) that enable developers to build generative AI-enabled tools, applications, and services for industrial digitalization. Nvidia Omniverse is generally totally free to use and develop on. Companies including Accenture, Ansys, Cadence, Databricks, Dematic, Hexagon, Omron, SAP, Schneider Electric with ETAP and Siemens are now connecting Omniverse to leading software tools, extending the platform's reach across industries.
-
Carlotta "Lotti" Siniscalco
The next frontier of AI isn't just about choosing the right model. It's about building moats through implementation. Across the Emergence Capital portfolio, companies like Bland and Federato are scaling faster by embedding Forward-Deployed Engineers. These engineers bridge the gap between AI and reality, ensuring real results in production. Read more from my colleagues Gordon and Kyle on why FDEs are the key to unlocking AI's full value. https://lnkd.in/gcsRRGX2
3 Comments -
Parag Madhani
storage underpins AI by enabling efficient data handling, model training, deployment, and real-time processing, directly impacting performance, scalability, and cost. Advanced storage technologies. ScaleFlux technology (CSD, SSD, Inline Compression) is at right time at the right place👍
-
Pat Gelsinger
Snowcap Compute Inc. is rethinking compute architecture from the ground up and tackling one of the hardest challenges in AI infrastructure: How to dramatically improve performance and efficiency as workloads scale. By advancing superconducting technology from the lab into real-world systems, they’re pursuing a true step-function innovation in how compute is built. Through Playground Global, I’ve had the privilege of backing teams redesigning foundational systems. Snowcap reflects that mindset: disciplined hardware engineering, deep technical rigor and the experience to challenge long-held assumptions in computing. The AI and Quantum eras will demand breakthroughs across silicon, systems, software, and energy. Increasingly, that future will be a cryogenic temperatures. Snowcap is making it really cool to be really cold! https://lnkd.in/gu35jNh8
7 Comments -
Phyllian Kipchirchir
Santa Clara-based Upscale AI has launched with over $100 million in seed funding to build an open-standard AI networking platform. Upscale AI is a networking company creating hardware and software solutions to make AI more accessible. The company challenges today's closed, proprietary systems by pioneering open-standard networking technology, providing high-performance connectivity for AI training, inference, and cloud deployments. The round was co-led by Mayfield and Maverick Silicon, with participation from StepStone Group, Celesta Capital, Xora, Qualcomm Ventures, and others. The funding will support the launch of its new suite of open AI-native infrastructure solutions this fall. Congratulations to founders Barun Kar and Rajiv K., and the Upscale AI team. Tech Funding News: https://lnkd.in/dxmJNicd Deals like this are reshaping the AI landscape. For weekly analysis on the funding, valuations, and strategic moves that matter, subscribe to the Charted Growth newsletter: chartedgrowth.beehiiv.com #AI #Networking #Infrastructure #OpenSource #SeedFunding #DeepTech #VentureCapital
-
Salil Deshpande
As AI workloads eat up global computing supply, DRAM prices are surging like never before—creating major challenges for teams trying to plan out their data center spend. One of my portfolio companies, Mext, is addressing the largest cost component in the datacenter: server memory (DRAM). From being on the board of Redis for the last twelve years, I’ve learned a lot about the various aspects of this problem and approaches to solving them. Mext found a breakthrough, using new AI techniques, for dramatically reducing the amount of server-DRAM required to run applications, all while maintaining performance. It intelligently manages the server’s memory, keeping hot pages (i.e., those in use by applications) in fast DRAM and offloading cold pages (i.e., those less used) to a much less expensive memory tier (e.g., NVMe Flash). Key to the approach is ensuring cold memory pages that are about to be accessed by an application are back in DRAM before the application needs them or notices that they were gone. This is done without modifications to the application or the OS – so it can run in the cloud or on-premise. Mext thus allows applications to either run using less DRAM or keep their DRAM footprint but do more with it. They're hosting a webinar along with Fred Weber (former CTO of AMD) on Jan 22nd at 10am pacific time. You can register here: https://lnkd.in/gXe9GCeE
3 Comments
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content