Pinned Tweet
Our @BagsHackathon demo submission
Primis mode
primis is the pricing layer for compute.
we help builders access compute through one unified layer instead of dealing directly with fragmented providers, unstable pricing, and infra complexity.
this is our @BagsHackathon demo submission.
6
9
59
34,552
// exciting start to my day regarding @primisprotocol
had a great call with a potential major compute data provider.
one thing became very clear: if we want to give builders predictable compute pricing, we must find a solution to how we hedge.
they’re introducing us to one of their partners to explore what the right risk-transfer infra could look like for our use case, including whether custom hedging instruments could make sense for primis.
still early, but a very exciting conversation. feel like an explorer reaching lands that no one ever been to before.
primis mode
11
8
43
5,731
// @primisprotocol update :
tbh we’re making good progress, just not as fast as i’d like.
the hardest part is giving builders prices they can actually plan around.
once primis gives you a workload quote, we carry the risk if the underlying market moves.
that’s why we’re investing so much into better pricing models, richer market data, and (hopefully soon) access to infras like @OrnnExchange to improve how we price and manage that exposure.
reliable quotes are hard. and we will make them the standard.
primis mode
10
7
39
4,505
We help teams save by turning the markets chaos into predictable rates.
most teams don’t overspend on compute because they’re careless. they overspend because the market is broken.
prices are fragmented. availability changes fast.
providers expose different rates. spot moves overnight. and builders are forced to make infra decisions with incomplete information.
@primisprotocol helps teams save by turning that chaos into workload-specific rates.
query rate → reserve rate → attach workload → reconcile usage
in the back, we score routes across price, availability, latency, reliability, and confidence and we also hedge prices for our builders.
so builders don’t just get “cheap compute.” they get compute they can actually plan around. that’s the unlock.
predictable compute economics for the compute markets.
primis mode.
2
3
26
7,721
this is very much @primisprotocol coded @blknoiz06.
we’re building onchain compute market infrastructure that lets builders access predictable rates, reserve capacity, and run workloads without directly eating every spike in gpu pricing.
primis mode.
where the fuck are the creative devs btw
how are kintara & papertrade the only cool things upcoming onchain rn
12
12
49
5,799
update //
appreciate all the feedback lately: product, comms, and token.
i’ll be honest: our communication with the community over the past month should have been better.
that’s on me.
i didn’t want to keep posting empty “soon” updates, but i also know silence creates uncertainty. going forward, i’ll be more consistent about sharing where we are and what we’re learning.
on product:
closed beta has been extremely valuable.
the teams using @primisprotocol right now are helping us catch edge cases, improve flows, and tighten the system before we open the gates more widely.
open beta is getting closer, and i’m very excited to have all of you onboarded.
on token:
our view has not changed. primis should align token holders, product, and company.
we’re always thinking deeply about how the token connects to the growth of the ecosystem over time, but the priority today is very clear:
build the product.
because the best thing we can do for long-term alignment is create something real that builders actually use.
and from now on,
less sorrys.
more shipping.
primis mode.
12
16
59
7,477
amazing post @0xfishylosopher, strongly aligned with our @primisprotocol thesis.
compute is too nonfungible for one generic market to solve everything.
builders need workload-specific rates:
gpu type, region, window, reliability, and usage profile.
then the layer behind that can manage exposure across providers, otc capacity, benchmarks, and futures.
that’s primis:
turn messy compute markets into rates builders can actually run against.
Lots of buzz recently on compute capital markets. But what might these markets actually look like? A few thoughts on its market structure from first principles:
> First, almost everyone agrees that compute has a nonfungibility quality. It behaves closer to electricity (temporal, nonfungible) than corn, oil, and gold.
> This nonfungibility creates several downstream corollaries:
(1) Reservations/capacity forwards are almost always bilateral OTC trades on particular SKUs and params (I want X hours of H200s in us-east-1 running Y model at 12pm on 8/1/2026)
(2) There is no transparent "one-size-fits-all" pricing model for "generic H200s" like there is for corn/oil/gold, hence no proper futures market used for hedging
(3) Most of the teams building in the space (eg. Silicon Data, Ornn, Compute Desk) are focusing on "standardization" indices/benchmarks, in preparation to create a liquid futures market.
> The short-side of compute markets fundamentally comes from neoclouds (Coreweave, Nebius, Lambda) and indepedent data centers (people with GPUs), while the long-side of compute markets comes from inference dev platforms (Fireworks, Modal, Baseten) and the agentic applayer (Cursor, Perplexity, Suno, Rime) that do not run datacenter fleets
> But these principals will never directly trade on general compute exchanges (eg. an H200 basket) because they require specific SKUs. Instead, they'll make their reservations/capacity forwards for specific SKUs with OTC dealers.
> These dealers in turn can "hedge" particular SKUs with exposure to the underlying generalized basket exchanges. So the folks actually using compute futures exchanges are going to be MMs/OTC desks/compute dealers on both sides. This creates an endgame market structure like below:
4
9
45
4,731
Slowly then all at once.
Primis mode
last thursday we opened the gates to the @primisprotocol beta by sending 150 invites to teams and individuals on our waitlist.
you can expect an in depth update tomorrow of how this week went with concrete numbers.
and the next step will be a full on open beta to everyone.
primis mode
2
7
31
4,401
this is a reminder that builders are exposed to more than just model quality and token pricing.
they’re exposed to provider risk, jurisdiction risk, and policy risk too.
the future won’t be built on a single model or a single provider.
it will require routing, optionality, and infrastructure that can adapt when parts of the stack become unavailable overnight.
primis mode
The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.
The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.
Access to all other Claude models is not affected.
We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.
Read our full statement: anthropic.com/news/fable-myt…
4
18
71
7,533
we’ve just sent 150 invites to people on the waitlist.
not a drill. not a hack.
we prioritized based on:
how long you’ve been on the waitlist. whether you signed up with a company email.
this is the final onboarding wave before open beta.
the goal is simple:
test everything under multi-team usage before opening the doors to everyone.
can you feel it coming?
primis mode.
11
15
70
11,980
we’re almost done. which means open beta is around the corner.
primis mode
been a little quiet, not because nothing is happening but the opposite.
first the market has been moving exactly where we thought it would: gpu availability tightening, prices repricing overnight, compute futures appearing and builders realizing “access” is not enough.
the missing layer is becoming obvious. and that’s what we’ve been building with @primisprotocol.
our closed beta that has been running for months is already processing ~$100k/monthly. and allowed us to understand what needed to improve for scale. which is what we’ve been working on.
all this, to say open beta has never been closer as we estimate we’re 1 big sprint away from opening the gates.
primis mode
11
22
67
15,372
budgeting becoming a “huge issue” means compute cost is no longer hidden behind flat saas pricing and subsidies.
as usage becomes more agentic, costs become more variable, harder to forecast, and harder to absorb.
the next question for every team becomes:
what will this workload cost?
can we plan around it?
can we scale it without margin shock?
that’s why @primisprotocol is building the pricing layer for compute.
predictable compute economics are becoming mandatory.
@sama let’s chat ;)
Sam Altman said AI budgeting has recently become a "huge issue" for some companies, something that "never came up" earlier this year. bit.ly/4uxIGnv
7
15
51
5,450
been a little quiet, not because nothing is happening but the opposite.
first the market has been moving exactly where we thought it would: gpu availability tightening, prices repricing overnight, compute futures appearing and builders realizing “access” is not enough.
the missing layer is becoming obvious. and that’s what we’ve been building with @primisprotocol.
our closed beta that has been running for months is already processing ~$100k/monthly. and allowed us to understand what needed to improve for scale. which is what we’ve been working on.
all this, to say open beta has never been closer as we estimate we’re 1 big sprint away from opening the gates.
primis mode
13
20
66
14,800
when a few actors control hardware, memory, cloud capacity, and frontier APIs, builders get squeezed from both sides:
hardware becomes unaffordable,
API pricing rises,
and the market loses optionality.
open models and cheaper inference are the counter-force.
but the real unlock is the layer that can route across all of it:
frontier apis,
open models,
cloud gpus,
local / edge inference,
alternative providers.
that’s the @primisprotocol thesis.
not one model.
not one cloud.
not one hardware stack.
pricing + routing across fragmented compute so builders can access the best economics for each workload.
primis mode
So what Big Tech did in the USA is they bought up all the computer hardware needed for AI, including HBM (High Bandwidth Memory), driving prices up 400+%.
Nvidia, meanwhile, doubled the pricing on its medium-range GPU, the GeForce 5090, which now retails at near $5000.
What this did is it made local AI hardware unaffordable for the vast majority of American consumers who might have been interested in running local AI. Now, instead, they have to "rent" AI through Big Tech and pay the rising monthly access fees to access overpriced AI models offered by the same tech giants that bought up all the hardware.
Fortunately, China came along and released DeepSeek, Xaomi, Qwen, etc., with per-token costs absolutely SLASHED to a fraction of U.S. tech company tokens. (Literally costs only about 1-2% to use DeepSeek, compared to using Anthropic Opus.)
And at the same time, China is beginning to catch up on microchip fabrication and High Bandwidth Memory manufacturing, which will result in their products driving down prices for consumers sometime in 2027 (on the hardware).
So U.S. Big Tech companies tried to create a near-monopoly to enslave the American people, and China actually broke those shackles and gave American access to affordable cloud-based AI right now, with more affordable local AI inference hardware coming soon.
Go figure.
12
13
63
5,849
we’re here to stay with @primisprotocol. we know the compute market is competitive but we will compete, because as i said i’m a perseverant mfer that will outwork anyone.
primis mode
12
13
76
4,212
bar for bar what we’ve been saying @primisprotocol and we’re building the pricing layer for compute. the goal is simple: turn fragmented, volatile compute pricing into rates builders can actually plan around.
everyday more validation on our thesis.
primis mode.
9
22
63
6,481
holyyyyyy!!
@OpenAI @sama basically saying what the market is starting to realize:
access to compute is not enough.
builders need guaranteed capacity and predictable planning for critical workloads.
that’s the @primisprotocol thesis too. but as a neutral layer across fragmented providers, not inside one model ecosystem.
predictable compute is becoming the product.
primis mode.
Introducing OpenAI Guaranteed Capacity: a new offering that enables customers to guarantee long-term access to OpenAI compute.
We’ve made long-term investments in infrastructure, partnerships, and capacity planning to help customers scale reliably.
Now, Guaranteed Capacity helps customers plan ahead for critical workloads in a compute-constrained world.
openai.com/guaranteed-capaci…
14
27
95
11,582