Aditya Dhavala

Voice Products @ Sarvam

[Aditya Dhavala @dvs_aditya]

🥹🔥

[Sarvam @SarvamAI]

We're thrilled to announce that we have raised $234M in the first close of our $300M Series B at a $1.5B valuation.

@HCLTech and @BessemerVP have joined us in this round, alongside continued support from @khoslaventures and @peakxvpartners

For countries and companies, sovereign control on the AI stack is no longer an optionality. Sarvam will be the partner of choice for this aspiration. The capital allows us to accelerate our momentum towards this full stack of models, compute, and deployments.

A huge thank you to our customers, partners, investors, and the Sarvam team for your trust and belief in what we are building. We’re just getting started.

Read more: https://t.co/VmLtpnj8gx

[Pratyush Kumar @pratykumar]

Compute:
We were the first team in India to train at scale. We trained the sovereign models at the scale of ~3400 H100s. We are now putting serious capital behind the next step. India's first Blackwell cluster is now online and used by us, and we are building momentum towards operating 10s of megawatts in compute on Indian soil by 2027.

Models:
With Sarvam 105B, India's first sovereign model built from scratch, we showed that highly capable models can be trained here, independently. More importantly, the capability is now compounding across data, training, evaluation, systems, alignment, and deployment intuition. And we are scaling up to trillion-parameter class models, with larger runs built for coding, agents, and security. A coding model is coming soon...

Inference:
We already host our own models, with third party usage tripling in the last three months. We are soon taking live a production-grade token factory with the price, throughput, latency, reliability, and governance that banks, governments, enterprises, startups, and developers need for real systems.

Products:
Our products are now reaching India scale. Voice was our first wedge. It powers millions of interactions per day, doubling in the last three months, while we continue to optimise costs. Like voice, another modality at scale in India is documents, and we are hitting exponential growth of our new document intelligence product. Our fully managed agents product is live with enterprises and is being launched for all next month.

Deployment:
Most of the value in AI is unlocked in the last mile. We learned that by doing it across engagements in enterprises, government, and strategic sectors. Now we are turning that learning into a platform that allows every organisation to hill climb on its own use cases - whether it is building an agent, customising the harness, creating the data/tool backbone, or finetuning the model on custom data.

Talent:
Serious researchers are joining us across pretraining and RL, including people who have done meaningful work at the frontier. We are also starting our San Francisco office as the conduit for frontier AI ambition for India first, then the world.

[Pratyush Kumar @pratykumar]

Ok, so here is my take on the Fable ban, sovereign AI, Sarvam, etc.

The event is interesting as it has implications from many perspectives.

For AI users, it is clear that you should not confuse access with ownership, or adoption itself as advantage. And if the most significant tech differentiator you are leveraging has external control loops, then you have to accept you are vulnerable.

For AI talent, it is now a precedent that you would be *seen* aligning to national interests more than company interests. And even if its just a whim for now, this trend will be hard to reverse as the world gets more automated…

For AI labs, their offerings will be stratified - general purpose AI would be available as utility, but frontier AI would be gated. This is a fantastic business model for labs - *democratized* AI sucks in all the data liquidity of the world which is locked in higher margin frontier offerings.

I think for the world to be a better place, all three of the above are bad vectors. We need to have more countries and companies owning their own destinies. And in the post AI world, that means being able to use and improve AI systems within their own perimeters - what one may call Sovereign AI.

At Sarvam, Sovereign AI in India was the founding thesis a couple of years back, and continues to remain the core operating principle. From our vantage point, it is super clear that India will build, leverage, and create massive business value and societal impact with sovereign AI. The following is precisely how we at Sarvam are contributing to make that happen.

[Aditya Dhavala @dvs_aditya]

An AI native FDSE firm, and having internal workflows which are AI enabled, can outcompete the Wipros and TCSs of the world and build a services company larger than these firms.

[Y Combinator @ycombinator]

AI-Native Service Companies
@gustaf

The total spend on services is many times larger than the spend on software, and a lot of those services are already outsourced, which makes them easier to replace with an AI-native product.

We're excited about companies that don't sell a tool to help you do the work: they just do the work.

[Aditya Dhavala @dvs_aditya]

In the old world of product management, you had 2 deterministic systems (the product/app and the measurement layer: events, SQL, analytics) interacting with 1 stochastic system: users.

In the new world, all three are stochastic. The product is non-deterministic. The measurement (evals) is non-deterministic. And users are still unpredictable.

You can't just instrument your way to insight anymore.

Product intuition and taste are more important than ever. You can no longer fall back on numbers and metrics alone.

[Aditya Dhavala @dvs_aditya]

@kamath_sutra Great launch, and some really well articulated insights! Congrats @kamath_sutra 🔥

[Aditya Dhavala @dvs_aditya]

@0xThoughtVector @SarvamAI Hey do apply through the link

[Aditya Dhavala @dvs_aditya]

We're hiring Product Managers at @SarvamAI (Bengaluru, on-site).

Looking for builders with 3-7 yrs of PM experience who've gotten their hands dirty with AI, have built agents, thought about evals, shipped real products at scale.

If you light up talking about what you've built, we'd love to chat.

https://t.co/Tnr0fngHMQ

[Pratyush Kumar @pratykumar]

📢 Open-sourcing the Sarvam 30B and 105B models! Trained from scratch with all data, model research and inference optimisation done in-house, these models punch above their weight in most global benchmarks plus excel in Indian languages.

Get the weights at Hugging Face and AIKosh. Thanks to the good folks at SGLang for day 0 support, vLLM support coming soon. Links, benchmark scores, examples, and more in our blog - https://t.co/DcCG3zlN8p

[Aditya Dhavala @dvs_aditya]

Airports feel like portals into new chapters of your life.

There’s just something about stepping off a flight.

[Sarvam @SarvamAI]

We're excited to partner with the Government of Maharashtra. Through this partnership, we aim to strengthen the state’s sovereign compute capacity and deliver tangible impact across healthcare, education, agriculture, and governance.

[Aditya Dhavala @dvs_aditya]

Important Update about Indus

[Sarvam @SarvamAI]

The response to Indus has been incredible. People are using it across languages and asking questions that require depth, context, and real reasoning.

We’ve been listening closely. A lot of what we’ve shipped comes directly from your feedback.

Recent updates:

1. Indus is now available across the United States, Canada, Singapore, the United Arab Emirates, and the United Kingdom on web. We will be live on Android and IOS within the next 2 days.
2. Dark mode is live on all platforms. Please update the app to experience it.
3. Stability and performance improvements across the stack.
4. Improved reasoning for maths, logic, and science questions with more structured, step-by-step answers.
5. We’ve increased context windows and rate limits on Indus to support longer conversations

We will keep iterating. What would make Indus better for you?

[fatih kadir akın @fkadev]

My son asking me a lot of questions. It’s a distillation attack obviously.

[Aditya Dhavala @dvs_aditya]

Smaller models trained on enough data are pretty good at chat use cases. Who would have thought.

[Aditya Dhavala @dvs_aditya]

@Krishnaik06 Making this happen didn’t just require the models but a lot of engineering ingenuity. Take a bow @meenu_stardust @kingekartik !

[Krish Naik @Krishnaik06]

Tested Sarvam AI conversational Agent and its just amazing.
#sarvamai #Sarvam

[Dharmendra Pradhan @dpradhanbjp]

Tried the Indus App by @SarvamAI today — impressive experience. The LLM responses were sharp, context-aware, and thoughtfully structured, and the product design feels intuitive and user-friendly.

What excites me more is what it represents: a step toward India’s #SovereignAI — models built, trained, and aligned within India, on Indian languages, datasets, and governance frameworks.

As digital public infrastructure transformed payments and identity, sovereign AI could power secure, culturally rooted, multilingual intelligence at scale — from governance to MSMEs to education.

If tools like Indus are any indication, India isn’t just adopting AI — we’re building it. @pratykumar

[Ministry of Information and Broadcasting @MIB_India]

Following the launch of Sarvam-30B and Sarvam-105B models at the India AI Action Summit, co-founder Vivek Raghavan discusses the company’s focus on developing full-stack AI solutions for India.

📽️ Watch

#IndiaAIImpactSummit2026

@GoI_MeitY @AshwiniVaishnaw @JitinPrasada @IndiaAiExpo

[Aditya Dhavala @dvs_aditya]

It's actually 1B* active

[Aditya Dhavala @dvs_aditya]

Since the start of February, @SarvamAI has been showcasing everything we’ve been working on.

We launched global benchmark–beating Vision, Speech Recognition, and Text-to-Speech models. We announced government and private partnerships. We demonstrated the scale and impact enabled by the application layer we have built. We unveiled India’s fully pre-trained, sovereign 30B and 105B models. And just yesterday, we launched Indus.

Among all of these celebratory moments, the highlight for me personally was getting to showcase what our smaller 30B model can do on stage at the AI Impact Summit. We demonstrated “Vikram,” an AI assistant built on top of the 30B model, accessible through a simple phone call. To make the point unmistakable, we did the demo on a Nokia keypad phone.

[Aditya Dhavala @dvs_aditya]

Our 30B model with just 2B active parameters, it delivers high throughput at extremely low latency and cost, built for population-scale inference. "Vikram" is connected to the internet, so users don’t need a data plan or smartphone. They just need the ability to make a phone call. No apps. No downloads. No internet required on their side.

[Aditya Dhavala @dvs_aditya]

@ParthVa53135321 @DwivediAbhi Good workaround for now, we will try to launch outside India as well soon.

[Aditya Dhavala @dvs_aditya]

Proud to announce the launch of Indus!!

Chat with our 105B model on India, do DM or reply with feedback.

Please do remember this is limited beta and there may be bugs. Please bear with us.

[Pratyush Kumar @pratykumar]

Drop 14a/14: Over the past few days, we’ve been putting our models, products, and research out there. The positive feedback has helped more people agree with our long held belief that #IndiaCan be a builder in this space.

Today, we are rolling out Indus - a chat interface to experience Sarvam 105B. Here is what all you can do: