Liquid cooling solutions | backed by @ycombinator and @e14fund
San Jose, California
Joined August 2021
- Tweets285
- Following139
- Followers137
- Likes528
Ferveret retweeted
A data center is a black box. Power and capital go in. Tokens come out, along with a bunch of crap (CO2, heat, water consumption, and other externalities). Anything you improve inside that black box, whether cooling, power distribution, silicon, or networking, ultimately affects token output. The question is just by how much. Tokens are directly tied to data center revenue, and that is what operators actually care about. So instead of expressing benefits in thermal metrics, we express them in terms of how much more token output the same power and capital can produce. In our case, we also eliminate water and reduce the other crap along the way. Tokens per watt (and tokens per dollar of capex) is increasingly the metric everyone benchmarks on. It is the only one that maps directly to revenue and ROI.
Ferveret retweeted
AI is pushing data centers to their limits, and cooling is becoming one of the biggest bottlenecks.
@Ferveret is taking a nuclear reactor-inspired approach to fix it. Their system outperforms state-of-the-art liquid cooling by 15% and delivers 35% more tokens from the same power— with zero water consumption. news.mit.edu/2026/nuclear-in…
Ferveret retweeted
Data centers are projected to account for anywhere from 9 to 17 percent of total electricity usage in the U.S. by the end of the decade. Today, around a third of data center electricity is devoted to cooling the chips that run AI models. That’s the process Ferveret is working to make more efficient. news.mit.edu/2026/nuclear-in…
ALT A computer chip with bubbles rising from it.
What started as a friendship at @MIT has quickly become a mission: helping data centers generate more AI compute with less power and zero water consumption.
Today, Ferveret was featured in MIT News.
As AI infrastructure scales, cooling is no longer a supporting system. It's a strategic constraint. Our Adaptive Phase Cooling technology adapts the subcooled boiling physics used in advanced nuclear reactors to cool AI chips. The result:
• Up to 35% more AI tokens from the same power budget
• Zero water consumption
• Modular deployment that works with existing infrastructure
We're grateful to MIT News and Zach Winn for telling the story, and proud of what the team has built so far. The future of AI depends on the infrastructure beneath it. And we're just getting started.
- Read the full article here: lnkd.in/gAgRhUQk
- Check out our website: ferveret.com/
@nvidia @Azizian_Reza @CleanSpark_Inc
Every GPU you buy comes with a hidden tax. It's called cooling overhead. And with standard direct-to-chip liquid cooling, it consumes 17% of your total power budget before a single token is generated.
Ferveret cuts that tax to 3%.
That's not a rounding error. For a hyperscaler building at gigawatt scale, it's billions of dollars in recaptured revenue.
Ferveret retweeted
When power is the constraint, cooling becomes capacity. Every watt we save on cooling is a watt that runs AI.
If you're interested in getting 35% more compute out of your fixed power envelope in your AI data center, come talk to us at @ @ferveret .Thanks to Shane Snider and @datacenter for the great feature.#AIInfrastructure #DataCenters #Cooling #Energy
@ycombinator
@e14fund
@climatecap
@aramcoventures
datacenterknowledge.com/infr…
Yotta 2024 is upon us in just 1 month!
If you are passionate about the next generation of #datacenter infrastructure and #AI development, you should join us in Vegas from October 7th - 9th! @ferveret's team, led by our CEO @Azizian_Reza, will be in attendance to talk about data center challenges and @ferveret solutions.
Are you interested in the latest cooling innovations? Just DM us if you want to connect!
Great article in the Wall Street Journal on Data Center Cooling!
Some key highlights for data center practitioners:
🔹 Liquid cooling is going mainstream: @nvidia and @Supermicro_SMCI are now shipping significant volumes of compute and racks with liquid cooling.
🔹 Energy efficiency and sustainability: These continue to be major drivers and key differentiators in the industry.
🔹 Nuclear power plant expertise: An interesting mention in the article highlights how this expertise is crucial for effective liquid cooling engineering. At @ferveret, our co-founder and CTO brings over 10 years of experience in nuclear power plant cooling to the table.
🔹 Water leakage concerns: While this is still a significant risk for many, @ferveret's tech eliminates this issue by design, immersing the compute in a dielectric fluid.
🔹 Noise reduction: Even noise is a big consideration. @ferveret's two-phase immersion cooling technology is virtually noise-free.
💡 Interested in cooling high-density racks? Let's connect and explore how @ferveret can help!
wsj.com/tech/novel-ideas-to-…
#DataCenters #Sustainability #LiquidCooling #Innovation #EnergyEfficiency #Ferveret
Are you headed to Nashville this week for @TheBitcoinConf? Are you interested in a two-phase cooling solution that actually works? Connect with our co-founder & CTO Matteo Bucci to talk about our new modular #crypto solution.
Ferveret retweeted
@ferveret is revolutionising liquid cooling solutions and can save your data center's energy by 96%!
Founded in 2021, Ferveret’s innovative cooling solutions can significantly impact applications in data centers, artificial intelligence, cryptocurrency mining, and more while reducing their carbon footprint.
Read the full article here: newzchain.com/transforming-d…
#newzchain #startup #Innovation #Technology
Last week, @ferveret's CEO @Azizian_Reza took to the stage at @IgniteGTM's AI Infra Summit alongside other cutting-edge companies like @Supermicro_SMCI and @APLDdigital. He discussed the latest developments and products from Ferveret aimed at supporting AI infrastructure.
What made this event so special was the equal attention placed on every component required to make an #AI application.
Far too often, we focus on the final product and what ends up in the consumers hands. We forgot about what goes on behind the scenes: logging frameworks, cloud services, even the infrastructure it runs on in #datacenters.
To have a conference where so many different disciplines got to speak, was something more than special. Thank you to the Ignite team and we hope to be back toward the end of the year!
Ferveret retweeted
You can’t invest in deep tech startups from behind a computer screen.
You have to get on the plane.
This bright line test will separate the true believers from the tourists. – at Manhattan, NY
Very interesting read by the @HudsonLabs team.
Cuz previously..."Riot Blockchain recognized $126M from crypto mining for the nine months ended September 30, 2022. However, during that time, only converted $52M into cash AND recorded an impairment of its bitcoin assets by $132M."
hudson-labs.com/post/ftx-ano…
Ferveret retweeted
Today the world is instant, except where it counts.
At @armada_ai, we're bringing AI to the remotest corners of the planet and the businesses that need it most.
Excited to finally share what we're up to :)
Ferveret retweeted
Love that NVIDIA is basically like “y’all can keep doing whatever you want, we’re just going to keep selling more GPUs”
Ferveret retweeted
Marathon just revealed the state of its operations in its latest quarterly report. How prepared is this mega miner for the #bitcoin halving?
Find out in this thread, and don't forget to read the entire article at the end.
Ferveret retweeted
Last day to apply for this tenure-track faculty search in Transformative Design!
xitter.cf/medialab/status/171721…
The Media Lab’s Program in Media Arts and Sciences invites applications for a tenure-track faculty search in Transformative Design. Applications are due November 1. Please share with your networks! media.mit.edu/posts/assistan…