builder. head of growth @Silicon_Data and @computeexchange
San Francisco
Joined May 2022
- Tweets40
- Following200
- Followers108
- Likes763
Dmytro Lokshyn retweeted
Only index provider at NVIDIA’s GTC analyst Q&A, so of course I asked @JensenHuang about GPU residual values. 😆
His take: newer is better, but older GPUs matter when everything is sold out.
He knows exactly what we do. Then he asked: “How did you get your data, and how can we help?”
Surreal. 💚 Also love running into Alison!
We planned the Compute Trading Happy Hour for a small group. 850+ people requested a spot and the Chotto Matte rooftop was packed.
Thanks to everyone who came and to our co-hosts @robotventures and @ani_pai!
#SFTechWeek
2
130
Come and join us. Feel free to dm me
We’re looking for an exceptional Product Manager for @computeexchange
Salary is $150-$230k depending on experience level / background. Comes with equity
Preference to be based in the Bay Area or NYC
To apply DM me or @computeexchange your background. Let us know why you’d be a good fit working with Compute Exchange, and share your CV/Linkedin
1
74
Dmytro Lokshyn retweeted
If you're in SF for Tech Week, we're hosting a happy hour with @carmenli & @stevehou to discuss compute trading, hedging, and where this is all going.
We’ve seen 0->1 products emerge here, and we'd like to bring together the founders & investors actively building in this area.
Join us:
We believe compute will be the fastest-growing asset class of the next decade.
During SF Tech Week, we're hosting a happy hour with @silicondata on compute trading and hedging.
What products should be built here?
If you're interested in joining, sign up below:
looking forward to seeing everyone
Compute Trading Happy Hour
Tue, Oct 6, 5 to 8pm PT
Chotto Matte, Union Square
For trading desks, quants, Neo-Cloud operators, GPU buyers, and investors.
Hosted by Silicon Data and @robotventures and @ani_pai with @computeexchange
#SFTechWeek
luma.com/y0ki9bun
3
122
The hardest question in GPU procurement is still "what's available and what does it cost." 31 clusters, published prices, real go-live dates.
Try it out on: compute.exchange/
Today we're launching Live Inventory: a second way to buy GPU capacity.
Providers post their clusters directly — model, count, location, term, price per GPU/hour, and go-live date.
31 clusters listed today:
32x H200, US Northwest, $2.81/GPU/hr, Sep 30
48x H100, Asia Pacific, $2.70, Nov 24
256x B200, Europe West, $4.68, Nov 30
72x B300, US, $3.95, Jan '27
1,152x GB300 NVL, US West, $6.50, Feb '27
Every listing comes from a verified provider, with transparent pricing and specs. Live now on the Compute Exchange Marketplace.
Try it out: compute.exchange
1
1
286
just one example. there's no golden rule indeed.
Midjourney founder owns 100% of the company.
I’m so bored of “optimal” founder virtue signaling and larping. There are no rules to starting a company. You either win or you don’t.
If you need a book of rules that tell you you’re doing a good job, stay in college or stfu
70
My first Zoom call with @silicondata, a year ago: the whole team was at an offsite in Singapore, and I suddenly found myself on a screen with 15 people I’d never seen, all waving from one room on the other side of the world.
Twelve months later: an index family that keeps growing, a product suite that ships faster than I can update the website, and now a Series A with this investor list. What a ride, and we’re just getting started.
Huge thank you to @carmenli for building a team where the bar is high and the trust is higher, and to everyone who shipped, reviewed, corrected, and shipped again. The best part of this milestone is who it’s shared with.
We are happy to announce a $30.5 million initial closing of our Series A, led by Valor Atreides AI Fund (@valor @Atreidesmgmt), with investments from @CMEVentures*, @DRWTrading, @FPrimeCapital, @SamsungNext, @vaneck_us, @further, @jumptrading, Tectonic Ventures, and @wintermute_t, and participation from @breed_vc, @hack_vc, @blank_vc, @SancusVentures, and @sogalventures.
The infrastructure underneath the round: daily GPU pricing benchmarks built from roughly 100 rental platforms in more than 40 countries, over 150,000 verified pricing records a day, continuous history since September 2024, alongside the GPU Forward Curve, the Silicon Data Token Index, the RAM Index, and SiliconMark performance benchmarking.
Our raise comes as @CMEGroup prepares to use Silicon Data benchmarks as the reference price for its planned cash-settled GPU futures market (pending regulatory approval) - a regulated instrument that will allow market participants to manage changes in GPU rental prices against a published, daily benchmark.
The Series A will fund expansion across four areas: benchmark pricing, performance measurement through SiliconMark, institutional market and alternative data, and risk infrastructure for derivatives, insurance and credit markets.
* Correction: in a previous post, we had mistakenly listed @CMEGroup as an investor instead of @CMEVentures.
1
139
Dmytro Lokshyn retweeted
We are proud to partner with @CMEGroup on the first compute futures contracts — H100 and B200, launching October 5!
AI has shifted from a technology scaling problem into a capital (allocation) problem. As training gives way to inference at scale, compute becomes a core production input.
Futures let providers and buyers manage that cost volatility and give builders the confidence to plan — the same role oil and crop futures played for energy and agriculture.
Post-trained open models, exploding inference demand, and broader enterprise adoption are hitting at once. Inference compute is becoming increasingly fungible. This is the start of a real market for the physical layer of AI.
Start trading the infrastructure of tomorrow with Compute futures, launching October 5. View contract specifications to see how the product is structured, including trading hours, tick increments and settlement mechanisms powered by Silicon Data’s real-time GPU benchmarks. spr.ly/6013B1MVMf
spent most of this year building software to run my marketing work. a lot of it just broke. scrapers died quietly and i wouldn't notice for days.
honest take on ai automation: nobody got replaced by an agent. somebody just got a new job babysitting software.
2
76
this is getting interesting
Our LLM Token index continues to decline while the debate over open vs closed model rages on. So we decided to take a closer look at how closed and open models respectively contributed. Interestingly, we are seeing some convergence btwn them.
The effective prices paid for proprietary models have decreased sharply over the last few weeks as OpenAI releases powerful frontier models at lower prices. Open models otoh have in fact seen their effective prices increase as more powerful near-frontier Chinese models like GLM 5.2 and Kimi K3 are released and served at higher prices.
Overall, this should be unsurprising. Econ 101: competition is up and prices are down. This is good for consumer and enterprise users of AI (agents) and promotes much wider and faster AI adoption.
89