We're short $CRWV. Report at https://t.co/sqX8rDCkf6. $CRWV has shot to a $75B valuation on technicals and hype, but strip away the noise and it’s just a levered GPU rental shop with lousy economics. 1/10

$CWRV is not the future of AI infrastructure. It’s other people’s chips, in other people’s data centers financed with other people’s money. Lenders may get paid, but equity holders get left holding the bag. Fair value: $10. 10/10

Meanwhile, the debt stack and cash burn is brutal: 11–15% delayed-draw loans, 9–10% vendor financing, $3B in junk bonds. As CEO admits: “Debt is the engine, it’s the fuel.” We see $19B of cash burn in 2025 and $40B through 2028. NGMI. 9/10

We rebuilt the same case with realistic assumptions. The truth: equity holders see no cash flow during the contract term, and only speculative crumbs afterward. No wonder the biz model has drawn red flags from day one. 8/10

80%+ margins, zero cost of equity, and residuals on 5-year-old GPUs. A former $CRWV engineer called it “utopian.” These numbers exist for one reason: to distract from unit economics that don’t work. 7/10

Management knows the economics of this model are weak. That’s why in June it handed analysts an “illustrative” contract with dazzling IRRs and rapid payback. Too bad the math only works in fantasyland. 6/10

$CRWV is boxed in with biz model that revolves around hyperscalers signing multi-year deals to backstop expensive debt. That keeps it out of the mid-market, kills flexibility, and leaves nothing beyond renting bare metal cheap to giants building their own. 5/10

$CRWV shareholders only benefit from whatever’s left when the chips are no longer under long term contract. Nvidia’s annual cadence means today’s GPUs are obsolete not too long after they are installed, leaving equity stuck with the scraps of dramatically depreciated hardware

Growth means nothing if returns don’t cover the cost of capital. For $CRWV, they don’t even come close. Every dollar borrowed to spend on GPUs and leasing datacenters just deepens the hole. This is value destruction dressed up as expansion. 3/10

Stock is up 30% since last week’s $MSFT–$NBIS deal and $ORCL’s all OpenAI backlog surprise. Bulls are missing the full story: $CRWV anchor customers are signing bigger deals elsewhere bc it has no moat amid rising competition. 2/10

Finally, Pure doesn’t deserve the premium SaaS valuation multiple that bulls want to ascribe it. Its “software” revenue largely comes from maintenance, support, and pre-paid hardware upgrades. Another former tech darling now legacy, we see 50%+ downside in $PSTG shares. (10/10)

Despite its perception as a disruptive high-performance pure play, $PSTG is poorly positioned to capitalize on cloud or AI infrastructure spend themes. NetApp is arguably better positioned for each, delivers superior profitability, and trades at half the price. (9/10)

$PSTG may have benefited from the adoption of all-flash arrays, but its AFA market share has been largely stagnant, and the migration of workloads to the cloud makes its core business a melting ice cube. (8/10)

$PSTG’s core enterprise business is vulnerable: big competitors like $DELL, $HPE, and $NTAP have closed the gap on storage-as-a-service offerings, non-disruptive upgrades, and management software while their superior product breadth makes them more strategic to customers. (7/10)

There’s a good reason why VAST is rumored to be raising money from Google and NVIDIA at a $30B valuation (10% higher than Pure’s market cap): they are a true AI enabler. Former $PSTG executives have noted they were surprised by VAST’s outperformance. (6/10)

$PSTG’s supposed architecture advantage has been nullified by the new generation of high-performance data infrastructure players such as DDN, WEKA, VAST Data, and Hammerspace who leverage commodity hardware and can achieve far greater scale. (5/10)

Pure’s deal with Meta has been overhyped. Far from buying Pure’s products, Meta is simply licensing its software for a narrow and replicable capability. $PSTG has failed to win subsequent projects and was guarded on new customer prospects on its recent earnings call. (4/10)

Our detailed TCO analysis confirms what we heard from hyperscaler storage architects: HDDs still maintain a 5-6x cost advantage over flash-based solutions for bulk storage applications and the HDD “access density” issue is a red herring. (3/10)

Pure insults investors by suggesting leading hyperscalers such as Google and Amazon would adopt its products if they only looked at total cost of ownership. These companies are arguably the most sophisticated IT buyers on the planet, and they aren’t buying $PSTG’s products (2/10)

We’re short $PSTG. Report available at https://t.co/w1sGp2IVZB. Pure Storage has gone from disruptor to disrupted. We believe its hyperscaler strategy will disappoint, while its core enterprise storage products face declining differentiation (1/10)