we crossed $6m in volume in May.
our biggest month so far.
all organic.
real users moving money around the world.
and the next few launches are built to make Avici even more useful, visible, and used every day.
we’re just getting started.
it is silly to abide by the existing framework of laws around how financial exchanges work when we are creating a completely new system from the ground up
@HyperliquidX would not exist if they attempted to fit into the existing system, neither would many other crypto protocols, there is a reason that they have to come up with new definitions in the Clarity Act for protocols that previously they did not have the sufficient regulatory definitions for, makes more sense to believe that the new laws will be structured to best accommodate the most successful crypto startups, of which Hyperliquid is clearly at the forefront, Jeff & co have said that they have had ongoing conversations with US lawmakers and continue to be surprised with how much they are learning how Hyperliquid works, the ICE has also made similar comments that have been positive
yes they are closed source, and yes they started with a small colocated validator set, but it is objectively true that they would not have been able to have as much success as they have so far without those core decisions at the beginning, it also is not as if they are not attempting to move to a more permissionless and decentralized construction as they've added more validators to the valset that did not exist when they first started that are not directly a part of the Hyperliquid team, i believe it's around ~25? but i know it isn't all from just their team and there are other people running validators
what people are underdiscussing currently with the future optimistic case regarding Hyperliquid, is that for the first time we have a pro-crypto president and a pro-crypto admin who both understand the importance of being early and supportive of these technologies but also understand how much money can be made if it is architected properly, now if you look at how Hyperliquid is designed with builder codes and HIP-3 it will benefit from a large # of exchanges and frontends being built on top of its core infrastructure and utilizing the deep liquidity and work that they've done creating a top of class exchange, the reason this should be important to Trump and the US govt if they understand it properly:
if Hyperliquid is successful and they become the AWS of liquidity for financial exchanges supporting perpetual futures, while USDC is the underlying stablecoin on the exchange, by default they will be creating a huge amount of demand for USD and extending the dominance of the dollar hegemony globally
they want everyone using US dollars, and the future state of the world a lot of that is going to be digital, the reason Clarity Act has been able to actually get support in Congress is because they understand the importance of stablecoins in the future and how they will continue to become more popular, in the same way that Trump has been outwardly supportive of companies that he believes are doing great innovative work in their respective industries like Dell, Ford, Intel etc. it is very likely that he will be as supportive with crypto companies if they are able to convey the value prop of Hyperliquid doing well effectively, I have confidence in @chameleon_jeff & @HyperliquidPC to correctly illustrate this opportunity to lawmakers
if I can see this thesis clearly for how building a decentralized perps exchange that many other countries and regions can build their own perps exchanges on top of that ALL use USD-based stablecoins underneath the hood, then i imagine it will be just as easy for others to see as well, we've already seen Trump support similar financial infrastructure with the Trump accounts through Robinhood, so not that farfetched IMO
ethereum:0x232ce3bd40fcd6f80f3d55a522d03f25df784ee2 token is heavily backed by both Robinhood and Coinbase and it is registered in United States. No brainer at this point
We're building one of Europe's largest AI factories in Lappeenranta, Finland, and we're looking for the team to run it.
Latest-generation NVIDIA accelerated compute. First capacity live in 2027, powered by low-carbon energy.
From shift patterns to tooling, this is a chance to write the runbook from day one across data center operations and data center IT.
Bring it online with us: https://t.co/ojPiuxzG9Y
did some digging so you don't have to, and how crazy is this?
One little AI accelerator touches dozens of hyper-specialized companies across three continents, and we’re talking thousands of engineers, chemists, and machinists who’ve never met each other but have to hit their exact tolerances or the whole thing doesn’t work.
If you map every layer of an AI accelerator package and name the under-followed supplier at each layer, the map looks like this:
The fiberglass that goes inside ABF substrates: Nittobo Industries (Japan, private), Asahi Kasei (3407.T), Owens Corning (OC) on the lower-end grades. Unimicron just warned that low-CTE supply is the next constraint.
The ABF film itself: Ajinomoto (2802.T) holds 98% of the IP and licenses it to five substrate makers.
The substrate fabricators: Ibiden (4062.T), Unimicron (https://t.co/ysHqK2LKVS), Nan Ya PCB (https://t.co/CZuj06w7zb), Shinko (7820.T), Kinsus (https://t.co/a1oVl1GoyF). Ibiden's share is 85% to 55% over three years.
The interposer and CoWoS process: TSMC (TSM), with outsourced spillover to ASE / SPIL (https://t.co/wC05wori4M), Amkor (AMKR), Powertech (https://t.co/dGu3NZXbUO).
The hybrid bonding step: BESI (BESIY / https://t.co/P7EXczR8Nt), with KLIC (Kulicke and Soffa) as the secondary equipment vendor.
The 3D metrology for HBM stacking: Onto Innovation (ONTO), Camtek (CAMT), Nova (NVMI).
The encapsulation chemistry that prevents the package from melting: Sumitomo Bakelite (4203.T), Shin-Etsu (4063.T), Nitto Denko (6988.T).
The PCB the package sits on: TTM Technologies (TTMI), Unimicron (cross-listing), Ibiden (cross-listing).
The 800G/1.6T optical chipsets: MACOM (MTSI), AAOI, Marvell (MRVL), Credo (CRDO).
The lasers feeding optical engines: Lumentum (LITE), Coherent (COHR), Sivers (https://t.co/Y78wp2HOGU), IPG Photonics (IPGP), nLight (LASR).
The InP substrate underneath the lasers: AXT (AXTI), Sumitomo Electric (5802.T / SMTOY), IQE plc (IQE.L), Freiberger (private).
The wafer-level burn-in tooling: Aehr Test Systems (AEHR).
The retimers that keep the signal alive: Astera Labs (ALAB), Credo (CRDO).
The CXL controllers for system-level memory: Astera Labs (ALAB again), Marvell (MRVL again), Samsung Electro-Mechanics (009150.KS), SK Hynix (000660.KS).
The HBM itself: SK Hynix (000660.KS, 62% share), Micron (MU), Samsung (005930.KS).
The cooling at rack scale: Vertiv (VRT), Modine Manufacturing (MOD), LiquidStack via Trane Technologies (TT), CoolIT now inside Ecolab (ECL), Boyd (private).
The power semis converting 800V to 1V on the GPU rack: Power Integrations (POWI), Innoscience (https://t.co/yaXvTyRSv9), Navitas (NVTS).
The high-voltage transformers feeding the substation: Siemens Energy (https://t.co/vB88aXEM2D), GE Vernova (GEV), Hitachi (6501.T), Eaton (ETN), Virginia Transformer (private).
That is the supply chain
The names most generalists could pronounce are the GPU vendor and three or four ETF-level large caps. The actual binding constraints sit in the smaller names two and three layers below.
Gavin Baker had a really good take on the AI market recently on @InvestLikeBest podcast, and I think it gets at why this whole thing is starting to feel a little insane:
“I think cross-sectionally, if you look at the valuations for all these AI names, they just, they can’t all be accurate. You have memory makers at, you know, 3 to 5 times PE. You have NVIDIA at a really low PE. You actually have some other accelerator companies at reasonable multiples. And then you have everything else, everything in power, everything in cooling… power, cooling, even probably some of the optical names. These are discounting very different things.
If the multiples on the power, cooling, optical names are correct, Nvidia, memory, they’re going up a lot. If the multiples on NVIDIA and memory are correct, everything else is probably going to underperform. The AI market is cross-sectionally inefficient right now.”
The funny thing is that cards are not even the most interesting part of @AviciMoney
What blew me away is the combination of cheap on/off ramp (1$ per transaction, no exchange fees!!!!) with easy access to defi yield
Revolut is just waaaay behind
This feels a lot like when Chainlink started announcing partnerships left and right and then became the standard. Similair to what oracles did for defi $SERV Reasoning is doing for the ai industry. Turning expensive, innaccurate, hallucinating AI workflow into affordable, accurate, production ready AI workflow for enterprises accross the board.
$SERV standard loading.
BREAKING: President Trump says an agreement with Iran has been “largely negotiated” and will be announced shortly.
Trump says the Strait of Hormuz will be reopened under the agreement.
Long read, but this is one of the best articulated reviews on @NEARProtocol recently, covering all innovations (Privacy, AI / Agents, Intents, Abstraction, etc.) and proposing a new framework and metrics beyond old traditional ones.
I got into crypto because I was promised immutable capital that I could use to trade any asset, at any time, in a fully transparent manner on level playing field.
The right to transact is a fundamental, inalienable human right.
Hyperliquid is more than an asset.
Friendly reminder to those who own hype and have made considerable amounts of money now dunking on other companies.
Let’s not be smug and alienate future users of the platform. An exchange that is not growing is an exchange that’s dying.
Hyperliquid is for EVERYONE.
total blockchain revenue over past year compared to their circulating market caps
$HYPE: $788M | $13B
$SOL: $528M | $49B
$TRON: $470M | $34B
$ETH: $425M | $255B
always forget how crazy 2021 was, eth made nearly $2B off tx fees in one month that November
I’ve held HYPE for 1.5 years and haven’t sold a single token.
In that time I’ve faced massive port L’s including dumb leverage and on-chain plays
I even had to sell my long term stock portfolio from when I was 18 years old just to keep the content and streaming afloat
Where HYPE was at one point 20% of my port, becomes 30%, 50%, then 80%
So it would be deeply irresponsible for my to continue holding my net worth in one single asset
But I don’t care I’m holding anyways
Hyperliquid