A Small Manifesto Against the Current @Zcash Bandwagon Zcash is a remarkable piece of cryptography, but cryptography isn't the bottleneck for crypto in 2026. Distribution, liquidity, and developer adoption are. And those are the exact dimensions on which Zcash is structurally disadvantaged. 1. Network effects work against single-purpose privacy chains Privacy is a network-effect product: the larger the anonymity set, the stronger the privacy guarantee. Zcash currently has ~30% of supply in shielded pools, and most activity moves in and out of the shielded layer rather than staying within it. A shielded pool with ~5M ZEC and a few thousand daily active users provides meaningfully less privacy than the same cryptographic primitives running on an L1 with 10M+ daily addresses. The math is brutal. If 100 people hide in a room, finding any individual is hard. If 100 million people hide in a room, it's impossible. Privacy coins concentrate users. Privacy features on general-purpose chains recruit them. 2. Liquidity and acceptance are non-negotiable A privacy coin that gets delisted from major exchanges, as Zcash repeatedly has across Japan, Korea, the UK, and parts of the EU, becomes harder to acquire, harder to exit, and harder to use at scale. Privacy tools built on Ethereum, Solana, or Base inherit the liquidity of the underlying chain. You don't have to choose between privacy and the ability to transact with the rest of the financial system. Zcash forces that choice. Nobody wants to make it. 3. People don't want private money. They want private applications Most people don't need to hide a $50 ZEC transfer. They need confidential business payments, private payroll, undisclosed treasury operations, sealed-bid auctions, private voting and confidential DeFi positions that don't leak through transaction graphs. None of these run on a privacy coin. They run on smart contract platforms with privacy primitives like @aztecnetwork on Ethereum, @AleoHQ as its own L1, @solana 's confidential transfers, @penumbrazone in the Cosmos ecosystem, FHE-based chains like @fhenix and ZK-rollups in general The future of privacy is programmable, not denominational. 4. The technology has been completely commoditized zk-SNARKs were Zcash's moat in 2016. By 2026, they're the foundation of every major L2, dozens of privacy systems, and most rollup architectures. The Zcash team did the foundational research, and then watched the IP escape. The chains that benefited most aren't paying rent to Zcash, and they never will. It's one of the cleanest examples in crypto of pioneering a technology and capturing none of the value. 5. Regulatory exposure cuts the wrong way Privacy coins occupy a uniquely vulnerable regulatory category. Privacy tools on general-purpose chains can be designed with selective disclosure, view keys for auditors, compliance hooks and they live inside chains regulators have already accepted as legitimate financial infrastructure. Zcash has built the same compliance tooling (view keys, selective disclosure protocols) but still carries the "privacy coin" label that triggers automatic delisting regardless of actual functionality. The technology isn't the problem. The category is. 6. The unit-of-account problem For privacy to matter for real economic activity, it has to be denominated in money people actually use, this is the biggest lesson in crypto over the past 5 years. Nobody pays salaries, settles invoices, or runs treasuries in ZEC. They use USD, EUR, USDC, USDT. Privacy that requires switching unit-of-account is privacy that won't be used at scale. The winning model is private stablecoins and private transfers of mainstream assets, which requires programmability Zcash structurally doesn't have and isn't on a path to building. 7. The "private Bitcoin" comparison is just stupid At the end of the day, Zcash only really competes with Bitcoin, except it doesn't, because the "private Bitcoin" framing falls apart on contact with reality. You don't get to slap "private" on as a feature and call yourself Bitcoin's successor when you don't have the liquidity, the decentralized robustness, the regulatory acceptance, the size, or the history. Bitcoin's hashrate is distributed across hundreds of pools and tens of thousands of independent miners globally. Zcash's hashrate is functionally controlled by a handful of pools running ASICs from a few Chinese manufacturers. Zcash inherited Bitcoin's consensus model with a fraction of Bitcoin's decentralization. And decentralization isn't a sliding scale where "more" earns you partial credit. It's binary. You're either close enough to Bitcoin to inherit the monetary properties that come with extreme decentralization, as Ethereum genuinely is, or you're not, and the "moneyness" argument doesn't apply to you at all. Ethereum and even Solana have an order of magnitude better chances of reaching Bitcoin's market cap than Zcash does. That's not a controversial claim. It's just looking at the data.

Why did xAI hand over a 220,000-GPU cluster to Anthropic? The technical backdrop to xAI's decision to hand Colossus 1 over to Anthropic in its entirety is more interesting than it appears. xAI deployed more than 220,000 NVIDIA GPUs at its Colossus 1 data center in Memphis. Of these, roughly 150,000 are estimated to be H100s, 50,000 H200s, and 20,000 GB200s. In other words, three different generations of silicon are mixed together inside a single cluster — a "heterogeneous architecture." For distributed training, however, this configuration is close to a disaster, according to engineers familiar with the setup. In distributed training, 100,000 GPUs must finish a single step simultaneously before the cluster can advance to the next one. Even if the GB200s finish their computation first, the remaining 99,999 chips have to wait for the slower H100s — or for any GPU that has hit a stack-related snag — to catch up. This is known as the straggler effect. The 11% GPU utilization rate (MFU: the share of theoretical FLOPs actually realized) at xAI recently reported by The Information can be read as the numerical fallout of this problem. It stands in stark contrast to the 40%-plus MFU figures achieved by Meta and Google. The problem runs deeper still. As discussed earlier, NVIDIA's NCCL has traditionally been optimized for a ring topology. It works beautifully at the 1,000–10,000 GPU scale, but once you push into the 100,000-unit range, the latency of data traversing the ring once around becomes punishingly long. GPUs need to churn through computations rapidly to keep MFU high, but while they sit waiting endlessly for data to arrive over the network fabric, more than half of the silicon falls into idle. Google sidestepped this bottleneck with its own custom topology (Google's OCS: Apollo/Palomar), but xAI, by my read, has not yet reached that stage. Layer Blackwell's (GB200) "power smoothing" issue on top, and the picture comes into focus. According to Zeeshan Patel, formerly in charge of multimodal pre-training at xAI, Blackwell GPUs draw power so aggressively that the chip itself includes a hardware feature for smoothing power delivery. xAI's existing software stack, however, was optimized for Hopper and does not understand the characteristics of the new hardware; when it imposes irregular loads on the chip, the silicon physically destructs — literally melts. That means the modeling stack must be rewritten from scratch, which in turn means scaling is far harder than most of us imagine. Pulling all of this together points to a single conclusion. xAI judged that training frontier models on Colossus 1 simply was not efficient enough to be worthwhile. It therefore moved its own training workloads wholesale onto Colossus 2, built as a 100% Blackwell homogeneous cluster. Colossus 1, on the other hand — whose mixed architecture is far less crippling for inference, which parallelizes more forgivingly — was leased in its entirety to an Anthropic that desperately needed inference capacity. Many observers point to what looks like a contradiction: Elon Musk poured enormous capital into building Colossus, only to hand the core asset over to a direct competitor in Anthropic. Others read it as xAI capitulating because it is a "middling frontier lab." But these are surface-level reads. Look at the numbers and a different picture emerges. xAI today holds roughly 550,000+ GPUs in total (on an H100-equivalent performance basis), and Colossus 1 (220,000 units) accounts for only about 40% of the total available capacity. Colossus 2 — built entirely on Blackwell — is already operational and continuing to expand. Elon kept the all-Blackwell homogeneous cluster (Colossus 2) for himself and leased out the older, mixed-generation Colossus 1. In other words, he handed the pain of rewriting the stack — the MFU-11% debacle — to Anthropic, while keeping his own focus on training the next generation of models. The real point, then, is this. Elon's objective appears to be positioning ahead of the SpaceXAI IPO at a $1.75 trillion valuation, currently floated for as early as June. The narrative SpaceXAI now needs is that xAI — long the "sore finger" — is not merely a research lab burning cash, but a business with a "neo-cloud" model in the mold of AWS, capable of leasing surplus assets at high yields. From a cost-of-capital perspective, an "AGI cash incinerator" is far less attractive to investors than a "data-center landlord generating cash." As noted above, the most important detail of the Colossus 1 lease is that it is for inference, not training. Unlike training, inference requires far less tightly synchronized inter-GPU communication. Even when the chips are heterogeneous, the workload parcels out cleanly across them in parallel. The straggler effect — the chief weakness of a mixed cluster — is essentially neutralized for inference workloads. Furthermore, with Anthropic occupying all 220,000 GPUs as a single tenant, the network-switch jitter (unanticipated latency) that arises under multi-tenancy disappears. The two sides' technical weaknesses end up complementing each other almost exactly. One insight follows. As a training cluster mixing H100/H200/GB200, Colossus 1 was an asset that could only deliver an MFU of 11%. The moment it was handed over to a single inference customer, however, that asset transformed into a cash-flow asset rented out at roughly $2.60 per GPU-hour (a weighted average of the lease rates across GPU types). For xAI, what was a "cluster from hell" for training has become a "golden goose" minting $5–6 billion in annual revenue when redeployed for inference. Elon's genius, I would argue, lies not in the model but in this asset-rotation structure. The weight of that $6 billion becomes clearer when set against xAI's income statement. Annualizing xAI's 1Q26 net loss yields roughly $6 billion in losses per year. The $5–6 billion in annual revenue generated by leasing Colossus 1 to Anthropic, in other words, almost perfectly hedges xAI's loss figure. This single deal effectively pulls xAI to break-even. Heading into the SpaceXAI IPO, this functions as a core line of financial defense. From a cost-of-capital standpoint, if the image shifts from "research lab burning cash" to "infrastructure tollgate stably printing $6 billion a year," the entire tone of the offering can change. (May 8, 2026, Mirae Asset Securities)

Dubai Just Dropped the AED 750K Investor Visa Floor to ZERO 👀🇦🇪 Major update for property investors: The AED 750,000 minimum property value for Dubai’s 2-year investor visa is gone. New rule, effective this week: If you’re the sole owner of any property in Dubai, you can now apply for investor residency. No floor, no threshold. What changed for joint ownership: Each investor needs an AED 400,000 share. Two people buying an AED 800,000 apartment can both qualify now — which wasn’t possible before. Who wins from this 👇 First-time buyers 🏠 You can enter the market below AED 750K and still get residency. Studio and 1BR buyers are now in the game. Couples & partners 💑 Joint purchases now give both investors residency rights. One property, two visas. Smaller investors 💼 Priced out before? The barrier just disappeared. Dubai opened the door to a wider pool of global talent and capital. The bigger signal 📡 While many countries make residency and investment more complex, Dubai keeps simplifying. Faster process. Lower barriers. More attractive to live, not just invest. This isn’t just about real estate. It’s about attracting people who want to build a life here, not just park capital. Smart policy creates smart growth 📈

It isn't unexpected that the focus of the Bun Rust rewrite is on the anti-Zig side more than anything, since the internet loves to hate. What is unexpected and unfortunate is that leadership within Bun hasn't tried to steer the conversation away from that at all. There are so many positive and interesting takeaways from this and I'm not really seeing any of them pushed as the primary message. A positive thing that hasn't been talked about at all is how far Bun came thanks to Zig. And even if you dump it now, its meaningful for how good Zig was to even build a product to this point and impact by any metric. I would've loved to see anyone in leadership say this. On the interesting side is how fungible programming languages are nowadays. Programming languages used to be LOCK IN, and they're increasingly not so. You think the Bun rewrite in Rust is good for Rust? Bun has shown they can be in probably any language they want in roughly a week or two. Rust is expendable. Its useful until its not then it can be thrown out. That's interesting! There's been a lot of talk about memory safety and no doubt Rust provides more guarantees than Zig. But I'd love to see a better analysis of why Bun in particular suffered so much rather than take the language-blame path. How could engineering as a practice been more rigorous to prevent this? What were the largest sources of crashes other programs should watch out for? How does Rust prevent them? How could Zig theoretically prevent them? That's interesting. I know the official blog post hasn't come out yet from Bun. But they're smart enough to know that that PR would stir up controversy the moment it opened, or they should've been. And plenty in the company have been tweeting and writing about it. Its somewhat telling to me in various dimensions what they chose to talk about first. I tend to think I'm pretty good at corporate PR/comms (especially when it comes to developer audiences) and I think appealing to the negative is never the right long term strategy; it does work to get short term eyes though.

Jensen Huang is angry and the numbers explain exactly why (Save this). This clip is his response to two specific arguments that have been used to justify restricting Nvidia's access to global markets and he is not being diplomatic about either of them. The first argument he calls stupid is that GPUs are comparable to atomic bombs. His rebuttal is that there are a billion people using Nvidia GPUs right now, they are inside every medical imaging system on the planet, they were used in the CT scan he got the day before this interview, and he advocates them to his own family. You cannot make that statement about atomic bombs, and the moment you accept the analogy, Huang says, you cannot finish any coherent thought that follows from it. The second argument he calls completely ridiculous: that American companies should not compete internationally because they will lose anyway. His response is the one that landed, "If you guys all apply that same philosophy, why wake up in the morning?" It sounds like a motivational line, but it is actually a specific policy argument that maps onto a catastrophic real-world outcome that has already partially occurred. Nvidia's share of the AI accelerator market in China went from 95% to zero. Because US export controls made it impossible for Nvidia to sell, and by the time Washington partially reversed course and approved H200 exports, Beijing had already launched security investigations into those chips, DeepSeek had announced a pivot to Huawei hardware, and Chinese customers had begun rebuilding their pipelines around domestic alternatives. China represents what Huang called a $50 billion market opportunity. Nvidia has now walked away from $15 billion in lost sales, took a $5.5 billion inventory write-down in a single quarter when the H20 ban hit, and projected an $8 billion revenue loss in the subsequent quarter, all while the market it was excluded from accelerated its own chip ecosystem in direct response to being cut off. Huang's point is not that national security does not matter. His point is that the defeatist logic, "you're going to lose it anyway, so why compete" is self-fulfilling in a way that pure restriction advocates never account for. Every restriction that pushes China toward technological self-reliance is a restriction that permanently reduces American leverage, permanently shrinks American market opportunity, and permanently accelerates the development of the exact domestic alternatives the policy was meant to prevent. And Nvidia, with a $5.4 trillion market cap and zero percent China revenue, is the most visible proof of what that trade-off actually costs.

Trading for the following perpetual futures will be suspended on Coinbase on May 21 on or around 13:00 UTC. KAITO-PERP SENT-PERP SAHARA-PERP CAKE-PERP TOSHI-PERP AKT-PERP VET-PERP ANIME-PERP THETA-PERP ZK-PERP KERNEL-PERP BARD-PERP Any remaining open positions will be settled automatically at the time of suspension. The final settlement price will be calculated as the average index price over the 60 minutes prior to the trading suspension. The funding rate will be set to zero for the final funding period before final settlement. Coinbase reserves the right to suspend trading at any point and adjust the final settlement price to a reasonable level. These suspensions reflect our ongoing effort to maintain high-quality derivatives markets by focusing on products that consistently meet our liquidity and market-quality standards. By maintaining these standards, we ensure our listings maintain price integrity, and provide users with deeper liquidity and better trading experiences.