Kishan (@0xK885) header
Kishan (@0xK885) profile picture

president @voxchain, intern @Gemini

Joined October 2024
Photo by Kishan (@0xK885)Photo by Kishan (@0xK885)Photo by Kishan (@0xK885)
Simplicity wins in crypto. The best exchange is the one that needs no activation thought, because a new user never has to learn anything before they trade. Prediction markets show this. The technology is impressive, but the experience is what made it take off: yes or no, up or down. The benefits then sell themselves, such as no house setting the odds and a wide range of contracts. CT is in a bubble that makes it easy to forget the average person does not know Hyperliquid, Lighter, or even what a DEX is. That person is still the market everyone here is trying to capture. I believe the exchange that wins the tokenization and perps race will be the one that reaches people who have no desire to change how they invest today. That takes proving the value, which everyone is already hammering on, and making the product so simple there is no reason to say no. Most exchanges today are missing the second part, especially since most people's existing investing platform already offers the crypto they care about. @Gemini is one of the few platforms I see getting this right. You hold a credit card that earns crypto at no annual fee, buy crypto and equities, and trade predictions in one place.
1
8
235
GPU compute is becoming a tradable commodity, and two new instruments appeared this week. Hyperbolic hedged about $5M of B200 rental revenue with an OTC put built by Architect and Luxor, and Akash unveiled the design for Compute Forwards, a market to buy GPU time at a fixed price and take delivery months later. Until recently there was no standard price or contract for renting a GPU. AI teams could not lock in supply, and GPU operators could not hedge a falling rental rate. The first layer is price indices. Ornn tracks traded GPU rental prices and ICE plans futures on its index, while CME plans futures on Silicon Data’s H100 and B200 rental indices. Both are cash-settled, but neither has listed yet. ICE has not set a date, and the CFTC has pushed its review of CME’s contracts to January 8, 2027. The put works the same way, settling monthly in cash against the Silicon Data B200 index for one year. The second layer is physical delivery. Akash’s forwards are designed to settle in hardware. Each contract is one node of eight B300 GPUs in one region for one month, buyers are matched to verified nodes before the month starts, and validators check each lease hourly. It is only a design today, with a pilot planned for Q1 2027. The price case is clear. Silicon Data’s H100 index went from $2.00 in December to $2.97 on Nov 5, up about 48%. However, this is early as nothing trades at scale yet, and the indices do not see the large hyperscaler contracts. But, I believe this supports the long solana:rndrizKT3MK1iimdxRdWabcF7Zg7AR5T4nud4EkHBof thesis Dartmouth Blockchain pitched at the Franklin Crypto Research Competition. Compute is scarce enough that exchanges are building markets on it. solana:rndrizKT3MK1iimdxRdWabcF7Zg7AR5T4nud4EkHBof is a great example of this. Customers pay in dollars, and the network uses 95% of every dollar spent to buy RENDER on the open market and burn it. Over the past year, Ornn’s index shows the H100 rental price up 79% while solana:rndrizKT3MK1iimdxRdWabcF7Zg7AR5T4nud4EkHBof fell 25%. We think that gap exists because the market still sees Render as a 3D rendering network on flat demand. This is similar to the turn Nvidia made from gaming chips to AI chips. That demand is now turning. Monthly customer spending rose from $117K in August to $597K in October, the strongest month since December 2024, and burns covered 67% of new token issuance in October, up from 17% in August. Salad, a network of about 60,000 GPUs serving mostly AI customers, is routing a share of its revenue into the burn by around March 2027, which should roughly double the burn. Sources: Akash, Solana Breakpoint coverage, Luxor/Architect/Hyperbolic press release, CME, CFTC, ICE, Silicon Data, Ornn, Render Foundation dashboard, Allium, Blockworks
7
1
3
14
793
Unc
Apparently I've spent half my life on this app 🎉
1
3
345
Announcing the 9 finalists of the UBC 2026 Franklin Crypto Research Competition 🏆 Long >$250M $GRASS · UT Austin @txblockchain solana:rndrizKT3MK1iimdxRdWabcF7Zg7AR5T4nud4EkHBof · Dartmouth @Voxchain $PEARL · Oregon @oregonblock Long <$250M $SYRUP · @kublockchain $COW · EPFL @bsaepfl ethereum:0x31eea0edeb85b0cb65c1400bd2f4e487fd61b537 · Wilfrid Laurier Short >$1B $VVV · Cornell @CUBlockchain $PUMP · Yale @yaleblockchain $WLFI · St Andrews @USTABlockchain 100+ teams entered. These 9 pitch live Nov 20 at UT Austin. Thanks to @FTDA_US for the support, and to our judges @ev4nthompson @zackrosenblatt_ @br_slack @meta_rylan @brezshares and more!
6
10
3
38
2,925
Let's talk tokenized equities. The idea is to take a stock listed on NYSE, Nasdaq or LSEG and put it onchain so it trades 24/7. But that creates a problem, because the same stock now has more than one price. How is the price tracked? Most of the market uses a wrapper, where a custodian holds the real share and an SPV issues a 1:1 token against it. When the stock market is open, arbitrage keeps the token near the share price. When it is closed, the onchain price is the only live price. Why did this take so long? FTX launched tokenized stocks in 2020 and Binance in 2021, but Binance delisted within months after regulators pushed back, and Mirror's synthetic stocks went to zero when Terra collapsed. The problem was compliance, not tech, which is why xStocks relaunched in 2025 with real shares at regulated custodians. Size is now $4B+ onchain, up from $1.5B in June. What does this mean? In most wrappers you hold a claim on an SPV, so you get no vote, dividends by contract, and exposure to issuer and custodian risk. Most of the market also can't be bought by US persons, and only 2.6% of tokenized stock value is used as DeFi collateral. More than half of trading already happens outside regular hours, which suggests people want price exposure rather than ownership. The volume backs this up. In June, equity perps on Hyperliquid and Lighter did $67.8B in volume vs $4.2B for tokenized spot, about 16x. TradFi leans the same way, with US options volume hitting a record 15.2B contracts in 2025, the sixth straight record year. Why does this matter? SIFMA, Citadel, JPMorgan and Nasdaq warn that parallel tokenized markets will fragment liquidity, while crypto firms say the concern is overstated. The SEC's Sept 17 Innovation Exemption lets permissioned onchain venues trade tokenized stocks with the same rights as the underlying. DTCC went live Oct 28 and OKXICE opens Nov 23. OKXICE lists tokens without issuer consent, and 67 companies, including every Mag 7 name, have objected. I think that pushback is overstated. Wrappers backed by SPVs already exist offshore, and OKXICE tokens are backed 1:1 at a US broker-dealer with the same economic and voting rights. But I see the downsides, since liquidity can fragment and custody is weaker, with OKXICE saying SIPC coverage for self-custodied tokens is uncertain. I believe that moving onchain means migrating the existing system so that users notice no difference except the benefits. Longer hours alone are not the benefit, since Nasdaq and NYSE Arca launch 23-hour, five-day trading on Dec 6 on their existing infrastructure, with no blockchain. What onchain adds is 24/7 and weekend trading, instant settlement, perps and easier collateral. Sources: Coin Metrics, CoinGecko, Dune, RedStone, OCC, SIFMA and Citadel letters to the SEC, SEC Innovation Exemption, OKXICE public notice, DTCC, Nasdaq.
7
198
MSCI USA Index is dropping $MSTR despite @Strategy hitting $19.5 billion in profit last quarter (mostly unrealized $BTC gains). Why? MSCI changed its index criteria. Companies now need operating assets above 50% of total assets ($BTC is 92% of Strategy). Fail that, and MSCI runs five ratios (operating assets, expenses, cash flow, fair value swings, reliance on outside capital). Fail 4/5 and you get dropped. What does this mean? Passive index funds and ETFs that track MSCI benchmarks must sell their MSTR shares by the November 30 close. Strategy estimates MSCI-linked funds hold roughly 13 million shares (about 3.1% of its outstanding stock, worth ~$1.5 billion at recent prices). Why does this matter? @Strategy operates as leveraged BTC. When the stock trades at a premium to its BTC (mNAV), they sell stock and buy more BTC. The existing shares then have more BTC backing them. But YTD BTC yield, bitcoin per diluted share, is negative 3.7%. The edge over an ETF is low. Why this matters for BTC Strategy holds 850,364 BTC, about 4% of supply, and bought $14.33B of it this year. That bid depends on selling stock at a premium to its BTC. Losing index demand makes that harder. The buying has already slowed. Q3 holdings rose about 0.2%, with 7,218 BTC bought and 5,552 sold. BTC is roughly $65.4k against Strategy's $75.5k average cost. However they are not forced to sell yet: $5.07B of USD reserves covers 3.4 years of dividends and interest. The risk is losing the biggest marginal buyer, and a seller if capital dries up. $MSTR is not the only treasury facing these issues. Metaplanet (Japan's Strategy) and Yellow Cake (UK-listed physical uranium holder) were also removed, and about two dozen other treasury-style companies are on a watchlist. Sources: Strategy Q3 10-Q and 8-K (SEC), MSCI methodology via CoinDesk and Crowdfund Insider, TechTimes, BigGo, The Defiant
1
4
167
Kishan retweeted
fuck it. whole entire @Gemini intern interview
3
4
22
978
.@gemini Winter 2027 SWE and PM intern applications are live!!!!! see you guys on the spaceship🚀
21
7
3
245
35,569
Best intern team!
Welcome @Gemini Fall 2026 Intern Class !!! onwards and upwards 🚀
1
5
953
Happy National Intern Day to all my future interns!! Can't wait to see you on the spaceship 🚀
2
1
42
2,212
Meet Kishan Sreenivasan (@0xK885), a Stand With Crypto College Representative at Dartmouth! As the President of @Voxchain, Kishan is already driving major impact for the crypto community on campus. SWC college reps like Kishan are instrumental to our mission. If you’re a student and passionate about crypto, get involved: docs.google.com/forms/d/e/1F…
1
4
12
1,362
The launch of LoopedHype has brought an interesting way to use HYPE to the Hyperliquid ecosystem, but how does it work? To start, what is Looped Hype? Looped Hype or LHYPE is a liquid staking token granted once staking your HYPE on the Looped HYPE protocol, allowing you to maximize your HYPE staking yield and potentially put your HYPE to work in HyperEVM DeFi protocols as well via LHYPE. . Looped HYPE uses an automated loop strategy where the Auto Loop feature rebalances the user’s position daily by borrowing, staking, or unstaking tokens based on market conditions, such as staking rewards (APY), borrowing costs, and LTV ratios. For example: • If staking rewards are higher than borrowing costs, AutoLoop borrows more HYPE and stakes it to increase yield. • If borrowing costs rise or token values drop, it repays loans or unstakes to reduce risk. The protocol claims to offer yields around 10% APY on HYPE, compared to a lower 2.3% APY for basic staking on Hyperliquid. According to LoopedHYPE, Phase 1 saw over 325k HYPE deposited, concluding the Phase 1 cap. Will be interesting to see if a similar dynamic plays out for Phase 2. @Looped_HYPE
2
310
Absolutely great read thank you @0xlykt! So easy to overlook how these parts of the Hyperliquid ecosystem actually come together to make arguably the best overall crypto experience out there.
The Hyperliquid Thesis This is going to be a very long thread explaining why I am so bullish long-term on Hyperliquid, the Hyperliquid ecosystem and $HYPE, including some points which people might easily overlook.
1
369
Anyone else need @kinetiq_xyz to launch? @0xOmnia bless us up
1
2
159
Kishan retweeted
Hyperliquid is on another level. In the last 3 months alone, ~$550B in volume has been traded—that’s $5.6B per day on average. For a so-called “new player,” these numbers are insane. You aren't bullish enough.
10
5
4
140
20,126
Kishan retweeted
ETH Denver Takeaways - Hyperliquid addition: >Users are cautiously optimistic about the unorthodox EVM launch, but no one of size is rushing in to farm with no real use cases just yet >HL natives are excited to see more pushback to normalcy around DeFi being developed on HyperEVM—thinking from first principles around what is needed and where can we innovate (equities, options, retail UX, etc) >Everyone is still a bit confused about why no native USDC / what to do on HyperEVM >Camaraderie among HL builders is real. We PvP in the arena but understand all are pushing on the same mission. >Structured products that offer easy ways to onboard non-DeFi native capital to Hyperliquid should have some significant impact over the next 6-12 months with some great teams working on this Lmk if I missed anything here
11
6
1
101
5,986
Wish our boys luck as they go to @buildoneigen hackathon tmrw at ETH Denver! From the woods 🌲to the mountains 🏔️ @NSchaievitch @0xK885
1
1
6
420
One thing that will be crucial to monitor with the nascent HyperEVM ecosystem is ensuring you are not using protocols that rushed to deploy to capture early mindshare. Use protocols playing the long game here, not the 2-week trade. HYPE security matters most.
The HyperEVM is live. This is a major step toward the vision of housing all finance by bringing general-purpose programmability to Hyperliquid’s performant financial system. The initial mainnet release of the HyperEVM includes: 1. HyperEVM blocks built as part of L1 execution, inheriting all security from HyperBFT consensus. 2. Spot transfers between native spot HYPE and HyperEVM HYPE. As a reminder, HYPE is the native gas token on the HyperEVM. 3. A canonical WHYPE system contract deployed at 0x555...5 for defi applications. The source code can be found at hyperliquid.gitbook.io/hyper… Effective immediately, the bug bounty program will pay mainnet bounty amounts for reports within the scope of the points above. See hyperliquid.gitbook.io/hyper… for details. For API and wallet users: the mainnet HyperEVM has chain ID 999. A JSON-RPC server for the mainnet HyperEVM is hosted at rpc.hyperliquid.xyz/evm. Node operators and other builders are encouraged to host their own RPC servers. Tooling and analytics around mainnet HyperEVM may not be polished on day one. However, there are many talented builders working to solve these developer pain points. To help with these efforts, raw HyperEVM block data is streamed realtime to S3 so that running a node is not required to index the HyperEVM. More technical details can be found here: hyperliquid.gitbook.io/hyper… General ERC20 native transfers and precompiles will be enabled on a future network upgrade. Any ongoing feedback for these features on testnet is greatly appreciated. While these features are implemented on testnet, the mainnet releases are staggered for minimal disruption to existing users on the L1. The HyperEVM is composable with the L1 state while not affecting the low latency trading experience of existing users. Thank you to all the builders and users who have shared feedback on testnet so far. It will be exciting to see new applications leveraging and building upon the existing onchain financial system of Hyperliquid.
1
1
7
393
HyperEVM mainnet is live. Now it is time for applications to bring Hyperliquid to its next state. Billions of dollars in capital efficiency is about to be unlocked. Let's ensure we do so in a manner that embodies the Hyperliquid ethos of product > pundits and user safety above all. Time to rethink how we build out a dapp ecosystem that does the small things right from the start. Time for community ownership first. Time for HyperEVM. Hyperliquid.
Quick roll call: who’s ready to borrow against their HYPE?
11
7
1
140
11,621