How to Trade CBRS (Cerebras) on Hyperliquid

What CBRS Actually Is

CBRS is the ticker for Cerebras Systems Inc., the Sunnyvale-based AI chipmaker founded in 2015 by Andrew Feldman that builds the Wafer-Scale Engine — a single piece of silicon that spans nearly an entire wafer with 900,000 cores and 2.6 trillion transistors. The pitch against Nvidia is that by keeping the entire model on one chip, Cerebras avoids the inter-chip communication latency that dominates GPU clusters; on Meta's Llama 4 Maverick benchmark, Artificial Analysis clocked the CS-3 at 2,522 tokens per second versus 1,038 TPS on the Blackwell B200. Revenue ran $510 million in 2025, a 76% jump that flipped the company from a $481M net loss to $87.9M of net income. Cerebras filed its S-1 on April 17, 2026 targeting a Nasdaq listing under the ticker CBRS at a $22 to $25 billion range, with Morgan Stanley leading the book alongside Citigroup, Barclays, and UBS.

Why the OpenAI Deal Is the Whole Story

The reason CBRS is the first blockbuster AI IPO of 2026 is one customer. In January, Cerebras signed a multi-year Master Relationship Agreement worth more than $20 billion with OpenAI for 750 megawatts of inference compute, deployed at 250 MW per year through 2028, with options to scale up to 1.25 gigawatts by 2030. The structure of that deal is what matters: OpenAI receives warrants that can scale to roughly 10% of Cerebras equity, and Cerebras gets ~$1 billion in OpenAI funding for data-center buildout. That single contract underwrites most of the forward revenue case the underwriters are pricing into the IPO range — concentration risk on one side, the most credible AI-inference demand pipeline in the market on the other. If you are trading CBRS, you are trading a leveraged view on that relationship holding up.

How the HIP-3 Pre-IPO Perpetual Works

CBRS lives on Hyperliquid as a HIP-3 builder-deployed market — meaning it is not a core Hyperliquid listing but a perp deployed by an outside builder, in this case trade.xyz, which acquired the CBRS ticker on April 24 for about $22,975 worth of HYPE. HIP-3 lets builders stake HYPE to operate their own perpetual contracts on Hyperliquid's matching engine and shared liquidity layer, and trade.xyz is using that primitive to run a Pre-IPO Perpetual suite. The mechanic is hyperp-style: there is no external oracle pre-listing, so price is discovered entirely on-book against an initial reference seeded from Cerebras' February 2026 Series H at a $23B valuation and S-1 disclosures. The convergence rules are explicit. If Cerebras lists on Nasdaq by May 30, the market converts to a standard externally-priced CBRS perpetual marked to the public stock. If the IPO slips, the contract enters a settlement window and pays out no later than July 30 against a TWAP of the perp's own price history. Either way, this is a cash-settled derivative — there is no claim on shares and no IPO allocation.

What Actually Matters for the Trade

At a $255 mark and 5x leverage, CBRS is a basis trade with a known catalyst date. The premium is the trade. Forge's secondary last marked Cerebras at $113.50, and Hiive cleared $166.37 on April 30; the perp is printing roughly 1.5x to 2.3x where private buyers and sellers actually transacted in the last week. Pre-IPO perps have historically held a premium to last-known private rounds because they offer 24/7 leveraged access in a market with almost no other on-chain venue, but premiums of this size are unusual when the convergence event is two to three weeks out rather than hypothetical. Three things to think about before sizing. One: liquidity is thin — the launch-day book did $1.9M in 24h volume against a market cap implied in the tens of billions, so slippage and funding can both move violently. Two: settlement risk is asymmetric — if Cerebras prices the IPO inside the underwriter range and the perp is sitting at a 1.6x premium, longs eat the gap; if pricing surprises to the upside (the reports of $30B-plus implied valuations are not idle), shorts get squeezed. Three: TWAP settlement is the tail risk — if the listing slips past May 30, the contract settles against its own price history, which means whoever controls the tape during the settlement window controls the outcome. None of that is reason not to trade it. It is reason to know exactly which side of the basis you are on.