Sudheera Vanguri

A decade ago, Eclipse made a bet on the potential of AI. This was before LLMs, before GPU shortages, before the world fully grasped the scale of compute AI would require. We believed that the future of intelligence would demand an entirely new computing architecture. That’s what we saw in Cerebras. Not another semiconductor company, but a team willing to defy a 75-year assumption about how computing should scale. While the industry kept adding more chips, more clusters, and more complexity, Cerebras asked a more fundamental question: What if the architecture itself was the bottleneck? The answer was wafer-scale computing, an engineering challenge many believed was impossible to solve, let alone manufacture and deploy at scale. Today, Cerebras powers some of the most advanced AI workloads in the world across healthcare, energy, scientific research, sovereign AI, and frontier model inference. And now, they’ve reached another milestone: Becoming a public company. What excites me most isn’t the IPO itself. It’s what it represents. We are entering a new era where intelligence becomes core infrastructure for every industry. The limiting factor is no longer imagination — it’s compute. The companies that solve that constraint will shape the next generation of economic and societal progress. Congratulations to Andrew Feldman, Sean Lie, and the entire Cerebras team! This is just the beginning. Read more: https://bit.ly/49I7pxt

1,229

108 Comments

Google DeepMind under Demis Hassabis and University of Chicago alum John Jumper gave us AlphaFold and then Isomorphic Labs. NVIDIA, through Kimberly Powell and BioNeMo, is pushing into autonomous labs and AI-native biology infrastructure with companies like Thermo Fisher. AWS and Dan Sheeran are building Bio Discovery around biological foundation models and wet-lab workflows. Microsoft, through Hoifung Poon and Microsoft Discovery, is focused on scientific agents, molecular modeling, and AI systems for scientific reasoning. OpenAI is increasingly showing up in pharma and life sciences around scientific reasoning, knowledge work, and agentic research systems. Big tech is approaching life sciences the way it approached enterprise software: cloud infrastructure first, then AI services layered on top. But I don’t think the best life sciences companies will use AI that way. They will build it directly into instruments, workflows, scientific reasoning, and operational execution. That is where the market gets more complicated. Life sciences is a massive market, and the next generation of biology will require enormous amounts of compute, data infrastructure, automation, simulation, and scientific reasoning. Big tech wants a piece of that stack but servicing life sciences may not be enough. The companies that ultimately win may be the ones that invest in, partner with, or acquire biology-native platforms where AI is embedded directly into the scientific workflow.

4

There's a reason he's Uncle Bob.... "When both human and AI operate from the same scope, the same component architecture, the same interaction model, the same contracts, the cognitive load shifts from exhausted vigilance to collaborative refinement."

2

Flow is broken. Csikszentmihalyi's flow depends on uninterrupted immersion and is sustained by a sense of authorship. 1. Uninterrupted immersion is completely broken when waiting for LLMs/Agents to do their job. People look for hacks like keeping several parallel sessions which brings only a small sense of satisfaction. 2. While for authorship, things are more subtle. Prompting introduces a persistent uncertainty about whose thought it really is. When you instruct, you become a cause of a cause. Our sense of competence and self-efficacy is tied to direct causal contact with outcomes. I've been turning this over for weeks and still no answer. What's your take?

37

25 Comments

Google’s Willow Chip Is Here, And Education Must Move Faster! Google’s new Willow Chip, capable of solving in minutes what even supercomputers would take unimaginable years to do, isn’t just a tech milestone. For education, it’s a wake-up call. When technology moves at this speed, every field, engineering, management, design, healthcare, finance, must learn, unlearn, and relearn continuously. The pace of innovation means students can’t depend only on traditional learning anymore. They need exposure to AI tools, real-time problem-solving, and industry-ready skills from the very beginning. This is exactly the direction modern institutions must move toward: helping students prepare for a world where breakthroughs like Willow will become the new normal. Whether it’s integrating AI in classrooms, providing hands-on workshops, or guiding students in building strong digital profiles, the focus must remain on staying future-ready. The message is clear: Technology will not slow down. Education shouldn’t either. #AI #education #engineering #management

20

2 Comments

India just committed $1.1B to a state-backed AI venture fund. Blackstone took a majority stake in an AI compute startup for $600M. Every major AI CEO is in Delhi this week. Everyone's racing to own AI infrastructure. But most are buying the wrong layer. GPUs and compute are first-order infrastructure. Necessary, but commoditizing fast. The second-order infrastructure — the layer that controls how AI actually touches customers, enforces compliance, and compounds data — that's where durable value lives. Compute is the highway. We're building the toll booths. While $2B+ flows into GPU farms this week, the real moat is being built quietly: AI systems that own the customer engagement layer for entire industries. Which layer of AI infrastructure are you betting on? #SecondOrderThinking https://lnkd.in/esPXVG8P

7

Pumped to be able to announce our partnership with Adarsh and the Foundry Robotics team! Low-to-mid volume production of complex assemblies is a massive problem area. Foundry is bringing AI and robotics to the problem, building the Everything Factory to power the next century of American manufacturing 🔥 🚀

29

On the Futurist Podcast with Ian Khan The Futurist, recorded live at the Humanoids Summit in San Francisco, we explored what it will truly take for robots to move beyond impressive demos and into sustained real-world capability. The core challenge is not actuation. It is not perception. It is not scaling model size. It is reasoning under physical constraints. The physical world unfolds over time. Objects shift. Contacts accumulate. Errors compound. Intelligence in this setting cannot be a next token prediction loop. It must maintain internal state, understand geometry, respect physics, and make decisions that remain coherent across seconds and minutes. Humanoids make this tension obvious because they operate in human environments. But the lesson applies across logistics, manufacturing, agriculture, and beyond. We are entering the phase where Physical AI must mature from pattern recognition to structured world modeling. That transition will define the next decade of robotics. If you are building in this space, I would love to hear what you think is still missing. Full conversation here: https://lnkd.in/gqw48xhA

48

3 Comments

Can we navigate societal transformation on tech timelines? Vinod Khosla (legendary Silicon Valley VC) is saying that AI do 80% of economically valuable tasks within the next decade. He imagines a world where humans are freed from drudgery where work becomes optional and purpose is self-defined. But like many tech leaders, he’s long on the transformation of technology and short on the transformation of society. And maybe that’s no bad thing, because we shouldn’t rely on the tech bros to design the future social contract anyway! But…. no one else is currently doing that work at the scale and speed required. That’s why efforts like Deric Cheng’s AGI Social Contract project matter. Serious thinking about how to rebalance power, protect dignity and steer toward collective benefit in a post-AI world. But we need to 100x that (at least) to have a chance of keeping up with the tech timelines. #AI #AGI #FutureOfWork #SocialContract #VinodKhosla #DericCheng #Governance #Transformation #TechEthics #AIeconomics #UtopiaOrDystopia https://lnkd.in/eNa8TfdG

6

Day 1: NeurIPS Met up with the incredible Jigar Doshi. He is one of the few using AI not for profit or hype, but to tackle poverty in India. Here are some remarkable initiatives he has led: 1. Copilot for Front Line Workers: LLM-powered co-pilot into an existing learning and support application to improve the training of auxiliary nurses and midwives in India so they can better manage high-risk pregnancies (in native languages). One woman dies in childbirth every twenty minutes in India. Many maternal and infant deaths could be prevented by improving access to critical care information and ensuring that health workers can detect risk factors and treat complications early on. 2. CottonAce app: Wadhwani AI’s initiative to help cotton farmers fight pest infestations. It is easily accessible, works offline and is available in nine languages. Farmers install pheromone traps amongst their crops, which trap pests. They then upload images of the pests to the app and the AI algorithm identifies the pest and calculates how severe the infestation is. CottonAce generates instantaneous recommendations on which pesticide to use and how to best apply it. The app then shares the data with nearby farmers, creating an early-warning system for the area. https://lnkd.in/eYCZb7hz https://lnkd.in/eSwSFVHy https://lnkd.in/eh8HY6Sd

227

9 Comments

I enjoyed reading Honson Tran's analysis of US-India partnership on AI. He argues that instead of looking at AI as an arms race, we should look at it in terms of an infinite game. This framing allows for more opportunities to collaborate. #AI #USIndia #ORFAmerica #Fellowship https://lnkd.in/eZDfAyMP

18

1 Comment

Demystifying AI at Stanford Continuing Studies & the HAI AI Index A brief recommendation for two resources: 1. Stanford Continuing Studies - several courses offered online and in person address key demand for AI education for technical leaders, investment leaders and ethicists. https://lnkd.in/gk5MJ2PC. I've attended several rewarding classes here in the past year which expanded my professional network and my knowledge of AI ethics and my ability to track the evolution of the AI industry. In my latest class Demystifying AI, Nestor Maslej treats his class to week-by-week insight at an analytical level not typically available in public news sources. 2. The HAI AI Index - https://lnkd.in/geSqc6Bs - lies at the core of the last course I attended, led by the editor of the AI index (https://lnkd.in/gXu2HkqN). Nestor and the HAI team are looking at the industry in ways that enable meta-viewpoints on progress and capability. I have no affiliation with Stanford - just a grateful review of effective curricula and for gaining the skills needed to peer into & shape our AI future. And sharing an observation on the current era of AI - this is a good time for people who have both technical and legal training to contribute to our AI future.

15

"AI is about personalization" - Neeraj Sahejpal beautifully articulated during a recent fireside chat. We explored #PersonalAI and the future of elite wearables, emphasizing the importance of cross-platform integration and a robust partner ecosystem to develop use cases for consumers like you and I. This session was very informative, and I personally learnt more about the technologies driving the wearable revolution and how Qualcomm is building the foundation with their products for these advancements. For those interested, you can watch the full discussion here: https://lnkd.in/ga7vxnGd

31

1 Comment

Out of the many #AI solutions we build at onnyx electronisys for the traffic and mobility industry—deployed across major cities in India and designed for true industrial-grade scalability—there is one project that stands out as my personal favorite for its - deeply meaningful purpose. This system embodies our mission of “Saving life by Saving a minute.” At #ITSIndiaCongress2025, held at Indian Institute of Technology Hyderabad in association with the ITS India Forum and OMI Foundation on 7th November, we had the privilege of showcasing a live, successful demonstration of our Ambulance Priority System, witnessed by Hon’ble Justice A. M. Sapre. This solution is special—not just technologically, but because it directly supports emergency response, reduces delays, and ultimately helps save lives. Proud moment for the entire team and a milestone in our journey toward making India’s mobility ecosystem smarter, safer, and more responsive. #ITSIndiaCongress2025 #IITHyderabad #SmartMobility #AIinMobility #TransportInnovation #DigitalIndia #FutureOfTransport #MobilityTransformation

25

I fell into the “Token Trap.” And this is not just my story. Uber burned its entire 2026 AI budget in 4 months. CTO Praveen Neppalli Naga called it publicly: “back to the drawing board.” That’s not just a tech euphoria running wild but a Product Governance failure, and I’ve seen early versions of it in my own work. We’ve been equating “AI-first” with high token consumption. I’ve learned, experientially, that this is not real truth. When token spend skyrockets, the budget shock is just the visible symptom. There are two deeper problems we aren’t talking about enough. 1. The Environmental Debt that today we are not accounting for. Every 100-word AI prompt costs roughly 500ml of water for data centre cooling. By 2030, AI infrastructure will rival the carbon emissions of entire cities. I plan to now run an environmental viability check alongside the technical spec on every AI feature, and why to build it using AI for a non AI feature. I will now move away from just the COGS to adding what does it cost to the world outside my product roadmap. 2. The Atrophy of Product Thinking — the highest risk to our skill, the craft of thinking, though the impact will show at last. When we outsource reasoning, we stop building the product muscle that handles ambiguity. AI produces fluent, confident output , but lacks the conviction that only comes from wrestling with a hard trade-off innate to us. If our teams stop thinking deeply through the Why and the How, our products don’t just become less meaningful , they also lose their soul. And that’s a competitive edge we can’t buy back with a bigger model. AI can be an effective copilot if we do not let it steer our ability to reason, think and empathize. My framework for responsible product building, the three principles I’m now living by: → Precision Over Power. Match the model to the complexity. Using a frontier model for a routine task isn’t AI-first thinking, it is an optimisation failure. Look for precision for a specific task while choosing what model to use. → The 30-Second Rule. Before reaching for a prompt, spend 30 seconds in focused human thought. Can I answer this myself? Can my team? Essentially, protect the habit of independent reasoning before it quietly disappears. → Human-First Architecture. Use AI to augment the data, sharpen the options, accelerate the draft. But never let it replace the judgment. The last mile of every product decision must remain human. Token for now may appear just to be a unit of cost but it actually is a unit of strategic choice. Human prowess still lies in knowing when not to prompt. #ProductManagement #TechLeadership #ProductGovernance #UnitEconomics #Sustainability #GenerativeAI #ProductStrategy

29

4 Comments

"India should focus on applied AI" is a position mindlessly parroted everywhere. CEOs of a western companies who want to sell foundational AI APIs to India are expected to say this. But surprisingly same position is taken by "experts" of domestic IT industry. We had this same apathy in developing fighter jet engines to advanced battery technology. Result? Country of billion people can have aircraft program stalled because a foreign player decides to just stop selling the engines. Or a large company has to go pleading to Chinese startups to sell technology/IP. On the other hand, the same big companies with >10,000 crore *quarterly* profits say we can't build an LLM in-house and invest in R&D, instead choose to buyback shares or distribute dividend. Same companies are happy to become mere distributors of technology in India. LLMs as technology is much simpler IMO as compared to jet engines to hardware. But the same apathy remains! Geopolitics is destiny for the next few decades. Fortunately, leadership of the country is far more aware of this than private players.. https://lnkd.in/daENie_V

53

9 Comments

🌐 Aspora Secures $53M to Power Banking for NRIs and Immigrants Aspora, formerly Vance, has closed a $53 million Series B round to accelerate its mission of building modern financial solutions for the Indian diaspora and immigrant communities worldwide. 🔹 The funding was co-led by Sequoia and Greylock, with support from Quantum Light Ventures, Hummingbird Ventures, and Y Combinator. 🔹 Aspora has now raised more than $99 million, reaching a $500 million valuation. 🔹 The platform enables NRIs to send remittances, invest, bank, and access credit and insurance globally. 🔹 In the last six months, Aspora’s transaction volume jumped from $400 million to $2 billion, serving over 250,000 users. 🔹 The company is expanding from the UK, UAE, and EU to the US in July, and to Canada, Australia, and Singapore by year-end. 💡 Why It Matters 🔹 Aspora is bridging the gap for NRIs and immigrants seeking seamless cross-border financial access. 🔹 The company’s growth reflects a broader shift toward digital-first, global banking for diaspora communities. 🔹 Backing from top VCs and angels highlights confidence in Aspora’s vision and execution. 🔭 What’s next? 🔹 Aspora will soon launch bill payments, NRI fixed deposits, and a streamlined banking account. 🔹 Upcoming services will also support healthcare and family needs for users’ relatives in India. #Aspora #Fintech #Banking #Remittance #IndianDiaspora #Startups #Innovation #Sequoia #Greylock #GlobalFinance

9

From PoC → Production is where most AI journeys fail. At MLDS 2026, I’ll talk about building real-world financial AI agents that are reliable, scalable, and trustworthy. See you in Bengaluru.

56

Major developments are on the horizon for India's burgeoning tech and AI ecosystem. Nvidia CEO Jensen Huang is set to visit India next week, engaging with prominent tech leaders and government officials. This visit is strategically timed as India intensifies its focus on domestic semiconductor manufacturing and AI development. Nvidia, a global leader in AI chips, has been a critical enabler for Indian AI startups and data analytics firms. Discussions are expected to revolve around potential investments, R&D collaborations in AI, and bolstering India's semiconductor ambitions. This signifies Nvidia's growing commitment to the Indian market, positioning India as a potential global AI hub. For professionals in AI, semiconductors, and venture capital, this visit could signal new partnership opportunities, talent demand, and a significant acceleration in the local tech landscape. How do you see this impacting India's journey towards becoming a global AI powerhouse? https://lnkd.in/g2h5SiaG #IndiaTech #AIInnovation #Semiconductor

7