Primis Protocol

194 posts

Primis Protocol banner
Primis Protocol

Primis Protocol

@primisprotocol

The pricing layer for compute

Metaverse Katılım Şubat 2024
3 Takip Edilen13.3K Takipçiler
Primis Protocol retweetledi
Khouba
Khouba@khouuba·
no more comparing dozens of providers. no more surprise bills. add one line to your workflow. query a rare, reserve it, attach it to a workload, and reconcile usage automatically. hours saved. thousands protected. @primisprotocol gives builders predicable compute pricing before workloads run. primis mode
English
5
13
41
2.5K
Primis Protocol retweetledi
Khouba
Khouba@khouuba·
primis beta update: we’ve been rebuilding primis compute around one clear lesson from builders in our closed beta: the problem is not discovering more compute. the problem is knowing what compute will cost, trusting that rate, and being able to build around it. when compute becomes a core production input, pricing errors, latency, and volatility are not small issues. they hit margins, reliability, and product velocity. this is why we’re rebuilding the rate engine from the ground up. from the builder side, open beta is being shaped around one simple flow: query rate -> reserve rate -> attach workload -> reconcile usage no provider noise. no new workflow to learn. no extra complexity for builders. primis handles the rest behind the scenes. predictable compute pricing inside the stack builders already use. primis mode.
English
7
18
48
1.8K
Primis Protocol retweetledi
Khouba
Khouba@khouuba·
our current next steps with @primisprotocol : - we're waiting on the call and conversation with the @BagsApp team - waiting for the first wave of @BagsHackathon winners to receive the prize money, us included - working our gtm plan for our open beta release - do bd work with some ai teams to onboard them seamlessly primis mode
English
3
11
54
2K
Primis Protocol retweetledi
Khouba
Khouba@khouuba·
great article from the people @variantfund @firstc0in @AlanaDLevin. this is exactly why @primisprotocol exists. not because compute markets are already perfect, but because they’re clearly evolving in that direction. once the conversation shifts from “who has gpus?” to: - how is compute priced? - what is the unit? - how do you hedge it? - how do you standardize it? - how do you settle it? you’ve already admitted compute is becoming a real market. that’s the world Primis is building for, as we’re turning fragmented, volatile compute into something builders can actually price, plan around, and scale with. primis mode.
Caleb Shack@firstc0in

x.com/i/article/2054…

English
5
12
50
3.1K
Primis Protocol retweetledi
Khouba
Khouba@khouuba·
another massive one for @primisprotocol. cme launching compute futures means the market is starting to admit what we’ve been saying: compute is becoming a financial primitive. when an input gets benchmarked and hedged, it stops being “just infra.” it becomes something that needs pricing layers, access layers, and risk layers around it. that’s exactly where primis fits. primis mode.
matthew sigel, recovering CFA@matthew_sigel

*CME TO CREATE FUTURES MARKET FOR COMPUTING POWER BACKING AI *CME, SILICON DATA PARTNER TO LAUNCH FIRST COMPUTE FUTURES

English
10
21
60
6.3K
Primis Protocol retweetledi
Khouba
Khouba@khouuba·
we won the @BagsHackathon with @primisprotocol but the hackathon was never the finish line. the reason we’re building primis is simple. builders should not have to absorb compute price risk just to ship products. and this is a hard lesson i learned while building @dolos_diary and trying to ship dolion. today, teams compare fragmented rates, chase volatile pricing, and plan margins around an input they cannot properly forecast. that works when compute is a small ops expense. it breaks when compute becomes the primary production input. primis is not another cloud. not another gpu marketplace. primis is the pricing layer for compute. one interface to query normalized rates, generate quotes, reserve predictable pricing, and track usage inside existing workflows. the goal is not cheaper compute as a gimmick. the goal is making compute financeable. open beta next. primis mode.
English
9
15
69
3.8K
Primis Protocol retweetledi
Khouba
Khouba@khouuba·
forward compute market is the future. primis mode
English
3
6
41
1.8K
Primis Protocol
Primis Protocol@primisprotocol·
Congratulations to all the builders ! This is only step 1 for us. Primis Mode
Bags Hackathon@BagsHackathon

Winner #3: Primis Protocol @primisprotocol BlackRock just called compute futures the next big asset class. Primis has built the infrastructure and markets to make this possible. Now with Primis, companies are turning fragmented and volatile GPU prices into structured rates that allow them to scale. As compute becomes a core economic input, Primis becomes the layer that prices, routes, and coordinates compute for real markets. x.com/khouuba/status…

English
17
20
107
12.8K
Primis Protocol retweetledi
Khouba
Khouba@khouuba·
compute is becoming a market. @primisprotocol is building the pricing layer for it. primis mode
Khouba tweet media
English
7
20
54
2.5K
Primis Protocol retweetledi
Khouba
Khouba@khouuba·
wake up. larry fink saying compute will become an asset class with its own futures market is exactly why the @primisprotocol thesis matters. once compute becomes strategic enough to hedge, the opportunity is no longer just owning compute. it’s building the layer that helps the market access, price, and coordinate compute economically. that’s what @primisprotocol is building. primis mode.
AI Edge@aiedge_

Larry Fink just outlined the best opportunity of the next decade: "The U.S. is short on power, compute, and chips - I believe a new asset class will be buying futures of compute." Those who get in on the AI compute gold rush will print money.

English
12
18
57
5K
Primis Protocol retweetledi
Khouba
Khouba@khouuba·
you can expect more and more updates as the open beta launch approaches. primis mode
English
10
10
43
1.7K
Primis Protocol
Primis Protocol@primisprotocol·
Our new landing page is live. Check it out and join the waitlist for the beta. Primis Mode
Primis Protocol tweet media
English
3
11
41
4.4K
Primis Protocol
Primis Protocol@primisprotocol·
in the primis compute open beta, builders will be able to: query normalized compute prices generate quotable rates attach pricing to existing workflows track usage against those quotes forecast compute spend before it becomes a surprise we’re not building another cloud. we’re not building a gpu marketplace. we’re building the pricing layer for compute. open beta soon.
Primis Protocol tweet media
English
7
9
37
4.1K
Primis Protocol
Primis Protocol@primisprotocol·
We’re building the pricing layer for compute
Khouba@khouuba

i’m convinced @primisprotocol will become one of the most obvious ai race plays in hindsight. the @BagsHackathon is just starting line. every builder is competing on models, agents, infra, latency, and margins. but underneath all of it sits the same primitive : compute. and today, compute pricing is still a mess. the old way : compare fragmented providers Chase volatile rates Refresh opaque quotes Maintain countless spreadsheets play a guessing game with what your workload will actually cost our way : query one pricing layer for compute. @primisprotocol normalizes fragmented supply into predictable, quotable rates builders can use inside their existing workflows. if compute is the new oil, then pricing is the market infrastructure around it. that’s what we’re building. primis is the pricing layer for compute. primis mode.

English
1
3
25
1.7K
Primis Protocol retweetledi
Khouba
Khouba@khouuba·
this is exactly why the @primisprotocol thesis matters. when frontier labs lock in massive new compute supply, the question is no longer just who can access compute. it becomes who can access it economically. more capacity doesn’t remove the need for a pricing layer. it makes it more important. @primisprotocol is built for the next step : not just more compute, but better economic access to compute through pricing, routing, and eventually capital-backed stability. primis mode
Claude@claudeai

We’ve agreed to a partnership with @SpaceX that will substantially increase our compute capacity. This, along with our other recent compute deals, means that we’ve been able to increase our usage limits for Claude Code and the Claude API.

English
13
14
47
5K