Varun Sivaram

🏆 @EmeraldAi_ is so honored to be named one of the @TIME 100 Most Influential Companies of 2026! We have a long journey ahead to truly earn our place alongside the extraordinary titans on this list, but we have our sights set on nothing less than a moonshot to turbocharge the AI revolution, unlock the energy bottleneck for compute, and protect energy affordability and reliability through flexible AI Factories. 🔗Read the full list here: ti.me/100companies 🔗And read Emerald AI's entry here: time.com/collection/tim… as well as our announcement here emeraldai.co/blog/emerald-a… It's inspiring and humbling to join the world's most impactful companies, including our close partner, major investor, and the most valuable company on Earth, @NVIDIA. Thank you for the support to dream big—that we can unlock hundreds of gigawatts of power around the globe for AI Factories that can flexibly control their power use while protecting the performance of high-value AI workloads. The rewards are massive: unlocking trillions of AI investment, accelerating life-saving scientific discoveries, and reducing energy bills for local communities as the next generation of AI data centers serves as grid allies and lucrative customers to anchor generations of revenue for public power systems. Emerald AI aims to make AI the hero, not the villain. And we stand on the shoulders of giants, such as pioneers in this field like @Google, whom we're so honored to join on this list. Achieving this transformation will require bringing the entire ecosystem together to fundamentally change how the energy system and the burgeoning network of AI infrastructure interact. That's why partnerships are at the center of our strategy, and we're delighted to see our partner and leading power company @nexteraenergy on this list. 🙏 A massive thank you to our team, so many of whom come from wonderful firms on this list and have chosen to build and launch our own rocketship. We've come together because of our shared mission to change the world and align the AI revolution with benefits for innovation, local communities, and global challenges. Thank you to our investors and incredible partners across the ecosystem. And thank you to @jsibo, CEO of TIME, and the fantastic @JustinWorland for writing the entry about Emerald AI! 🚀

@deanwball wow we independently posted the same tweet: x.com/vsiv/status/20…

This, from Olah, plainly contradicts the encyclical, which confidently asserts that AI does not have, and never will have, “real” thoughts or feelings. It’s disappointing to see Anthropic align itself with a document that violates their own moral and intellectual principles.

Fascinating. @Pontifex Pope Leo XIV writes in his AI Encyclical that AI basically can't be conscious because it's a machine. Then he invites @AnthropicAI co-founder @ch402 to speak at the launch, who argues we should take seriously AI's potential consciousness (functionalism)

#EEI2026 News: We are officially 12 days away from #EEI2026, and we are thrilled to announce that Arizona Governor @katiehobbs is joining our speaker lineup!

The world’s power grid is a ghost town. ⚡️ Massive infrastructure is being built that sits empty 70% of the time. Transmission lines are utilized at only 30% capacity, yet we struggle to connect new AI factories. The problem isn't just capacity; it’s unpredictability. @vsiv explains why the current "Peak Load" design is failing the AI era and how @EmeraldAi_ is turning "unpredictable" loads into flexible assets for the grid. The future of AI scaling isn't just more power—it's smarter power. 👇 Watch the full breakdown on the evolution of the grid now on YT! Link in replies.

Who is Varun Sivaram? CEO of NVIDIA-backed Emerald AI makes TIME 100 Most Influential Companies list of 2026 ebx.sh/5x0seu

Excellent take in @ForeignAffairs @edwardfishman: "US leverage" over China hinges on "If US can sustain its lead in AI," rather than on securing rare earth processing (hard but surmountable). Great citation of @RushDoshi @CFR_org calling this the "decisive decade" for US-China

@radicalvcfund @EmeraldAi_ @molwelch @_RobToews Thanks for having me on!

Our latest episode of Radical Talks is live! Tune in to hear from @vsiv, Founder & CEO of @EmeraldAi_ on the future of AI infrastructure and energy. Varun joins Radical Partners @molwelch and @_RobToews to unpack why powering AI is the defining challenge of the AI revolution and why the solution isn't just building more, but building smarter.

Since half the folks I read here, and folks up to Dario as well, are so down, maybe this is worth saying: the best evidence and theory is that AI will be very consequential but also very positive for human wealth, health, and flourishing. 1/3

The hyperscalers have already outspent the most famous US megaprojects

@ArushiSF We are so grateful to have you as our Executive Advisor! (And I love that special edition vest!)

Every company I work with is doing something that is integral to my value thesis for building a better power system: creating demand-supply balancing architecture updates, crafting new system rules, and occasionally changing power market design, one piece at a time, so that grids can serve people AND critical functions like #AI. Congratulations to @EmeraldAi_ on their seed extension with @EnergyImpact_ Proud to be your power & utilities executive advisor and help you shape flex design advancement- everywhere!

In this week's #RadicalReads, @vsiv, Founder and CEO of @EmeraldAi_, shares what a recent UK trial with @nationalgrid, @EPRINews, @nebiusai, and @nvidia revealed about turning AI's energy demands into a grid asset: radical.vc/turning-ais-bi…

@ShanuMathew93 @claydumas Guys! 😳 (If not for my complexion, I would be blushing). I'm grateful for both of your friendship and partnership—this big mission needs a whole ecosystem to achieve 🚀

@claydumas @vsiv I agree, it’s @vsiv’s superpower!

Near the top of @vsiv's (almost annoyingly) long list superhuman abilities, is how smoothly he can enlist people to a very big mission. That cause is turning data centers intro flexible assets, unlocking capacity and affordability right right when the world is desperate for both.

Nvidia and @EmeraldAi_ announced a collaboration with AES, Constellation, NextEra, Invenergy and Vistra to develop "flexible AI factories" that adjust power consumption based on grid conditions. Pairs Nvidia's reference architecture with Emerald software to modulate compute workloads dynamically. No firm official project commitments yet but eagerly watching...

🚀 Excited to launch our partnership at CERAWeek with @nvidia, @TheAESCorp, @ConstellationEG, @VistraCorp, @nexteraenergy, @nscale, and @InvenergyLLC! That incredible group includes 5 Fortune 500 companies, ~$5 trillion in market cap, and the 5 leading US independent power producers—all committed to flexible AI Factories based on NVIDIA's DSX Reference Architecture!

Nvidia, Emerald AI Partner With Power Companies on New AI Factories on.wsj.com/4t4iEaV

Nvidia, Emerald AI team with energy companies on "flexible" data centers trib.al/RVsuIBZ

@brexton 🙋🏽‍♂️we're in tech and have a story for relieving energy bills with AI as hero and not villain!

100% agreed and the pushback will only accelerate Does not help that the anxiety around AI severely outweighs the excitement for it Things will get much much harder to build

@dr_alphalyrae Congratulations Vega and NVIDIA team!!

Today we launch NVIDIA’s Nemotron Super 3, a 120B param open model designed to run agentic AI systems across scientific, enterprise and industrial applications. Partners working with us include Dassault Systèmes, Palantir Technologies, Lila Sciences and Edison Scientific Key highlights: - Nemotron 3 Super uses a hybrid mixture-of-experts (MoE) architecture with three core innovations, delivering up to 5x higher throughput and 2x higher accuracy than the previous model. - Hybrid architecture: Combines Mamba layers for ~4× higher memory and compute efficiency with transformer layers for advanced reasoning. - MoE efficiency: Only 12B of 120B parameters are active during inference. Activates 4 expert specialists for the cost of one to generate the next token, improving accuracy. - Multi-token prediction: Predicts multiple future tokens simultaneously, enabling ~3x faster inference. Read the tech blog👇