A quick peek at our NVIDIA GB300 NVL72 deployment process. NVIDIA #GB300# NVL72 brings next-generation rack-scale AI performance to the era of reasoning, with ultra-dense compute optimized for large-scale training and high-throughput inference. More to come. Stay tuned👐! #GPU# #AIInfrastructure# #neocloud#

## Earnings Wrap-Up ### **AMD (Buy) | TP: $511** * **Guidance slightly beat expectations ($11.2bn vs. $10.5bn Bloomberg consensus, Buy side $10.96b). ***Server CPU TAM** now projected at >35% CAGR. AMD’s leadership will be further strengthened by Venice and a diversified CPU portfolio. ***Progress on **MI455/Helios** remains on track. * **Outlook:** Some investors were a bit cautious before the print; additionally, the frequent demand upward revision misled analysts’ near-term estimates. That said, the pause before the print could further boost the share price outperformance. Overall, we remain bullish ### **Lumentum (LITE - Buy) | TP: $1,168** * **Slightly below** FY3Q revenue ($808m) slightly missed buy-side targets, and F4Q’s $985m below buy sides $1bn, though gross margins expanded to **47.9%**. * **Key Drivers:** Laser supply remains structurally tight as demand outpaces capacity. OCS is gaining momentum via multi-year agreements. * **Outlook:** While the FY4Q guidance led to a brief after-hours pull-back, earnings power is expected to scale dramatically into CY2027. ### **Supermicro (SMCI - Hold) | TP: $43** * **Shares jumped 18% after hours following a strong Gross Margin beat. * **Key Drivers:** Margin gains were driven by a better customer mix (enterprise) and fewer low-margin GB300s. * **Outlook:** While margins are expected to remain elevated, the rating remains **Hold** due to uncertain top-line growth and limited visibility on new GB300 bookings.

We are excited to share that Bitdeer AI Cloud is now supporting @nvidia #OpenShell#, an open source runtime designed to safely build and deploy autonomous, self-evolving agents 🤖. Powered by high-performance NVIDIA GB200 NVL72 systems in our AI DC, Bitdeer AI Cloud provides the infrastructure required to run long-running agent workloads at scale. Developers can build your own AI agent on NVIDIA OpenShell, and seamlessly deploy them on Bitdeer AI Cloud, enabling secure execution, sandboxed environments, and reliable compute for persistent agent systems. This collaboration aligns closely with key themes from the NVIDIA GTC 2026 keynote, where Jensen Huang highlighted the shift toward agentic AI, AI factories, and infrastructure designed to support reasoning, inference, and execution at scale. As AI systems move from simple inference toward autonomous workflows, secure agent runtimes combined with high-performance infrastructure will become a critical foundation for the next generation of AI applications. We’ve compiled more key highlights from the keynote. Read the full summary on our blog🔗： #neocloud# #NVIDIA# #NemoClaw# #Inference# #GTC26#

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)

SpaceX/xAI has just announced that it has signed an agreement with Anthropic to provide access to its Colossus 1 supercomputer. "Anthropic plans to use this additional compute to directly improve capacity for Claude Pro and Claude Max subscribers. As part of this agreement, Anthropic also expressed interest in partnering to develop multiple gigawatts of orbital AI compute capacity. The compute required to train and operate the next generation of these systems is outpacing what terrestrial power, land, and cooling can deliver on the timelines that matter. Built from the ground up in record time, Colossus delivers unprecedented scale for AI training, fine-tuning, inference, and high-performance computing workloads. Colossus 1 features over 220,000 NVIDIA GPUs, including dense deployments of H100, H200, and next-generation GB200 accelerators. The cluster delivers extreme parallel performance for large language models, multimodal systems, scientific simulations, and generative AI at frontier scale. SpaceX is the only organization with the launch cadence, mass-to-orbit economics, and constellation operations experience to make orbital compute a near-term engineering program rather than a research concept. If engineering challenges can be overcome, space-based compute offers near-limitless sustainable power with less impact on Earth." I think this is smart for SpaceX/xAI, as it seems to have some excess capacity as Colossus, so why not make a some money on the side by renting that compute capacity out. I think we'll see some more similar partnerships soon.

Part 4  Close Associates of Chen Zhi 1. Li Tian (also known as Thet Li,Figure1) Date of Birth: July 6, 1987 Place of Birth: Beijing, China Citizenship: Cambodia and Vanuatu Passport: RV091789 (Vanuatu) Li Tian obtained a Bachelor’s degree in E-commerce from Beijing Information Science and Technology University in 2009. As Chief Financial Officer (CFO) of the Prince Group, Li Tian was responsible for managing illicit financial flows within the organization, including the handling of laundered funds and the coordination of large-scale cash smuggling operations. In April 2020, Li Tian became the ultimate controlling shareholder of Hong Kong-listed company FSM Holdings Limited (Stock Code: 1721), holding approximately 60.23% of its issued shares. According to records from the Hong Kong Companies Registry, Li Tian also serves as a director of four additional companies registered in Hong Kong: Concept Planet (HK) Limited Silver Prosper International Limited Sky Talent Investment Limited STAR MAX DEVELOPMENT LIMITED 2. Zhou Yun (also known as Sandy Zhou,Figure2) Date of Birth: October 17, 1982 Place of Birth: Hubei Province, China Identification: Hong Kong Identity Card and Hong Kong SAR Passport Passport: G30024177 (China) Sandy Zhou  serves as Chen Zhi’s financial assistant and wealth manager. She has been designated under United States sanctions programs. Multiple insurance brokerage and securities firms of Mighty Divine Group in Hong Kong, which were beneficially owned or controlled by Chen Zhi through Sandy Zhou , have been subject to enforcement actions and regulatory intervention. 3. Wei Qianjiang （Figure3) Date of Birth: April 28, 1983 Wei Qianjiang graduated in January 2015 from the Open University of China (formerly the Central Radio and Television University), obtaining an Associate Degree in Business Administration. Through a Bitcoin mining enterprise in Laos operated by the Prince Group, namely Warp Data Technology Lao Sole Co., Ltd., large volumes of Bitcoin (BTC) were allegedly transferred to cryptocurrency wallets controlled by Chen Zhi. Wei Qianjiang served as the person in charge of this operation. Key business activities include: 2014: Established Chongqing Venture Capital and Longxun Technology in Chongqing, China  2017: Registered a Hong Kong entity under the same name, Loncent Technology, with plans to pursue an IPO on the HKEX August 2018: Appointed as a Non-Independent Executive Director of HKEX-listed company Geotech Holdings Ltd. (Stock Code: 1707), a company controlled by Chen Zhi July 2020: Loncent Technology was investigated by police for promoting gambling services and facilitating cyber-enabled criminal activities; multiple senior executives were detained or defected to authorities, while Wei Qianjiang, as the company’s legal representative and owner, was handled in a separate case 2020: Registered Merak Technology (Hong Kong) Limited in Hong Kong @BBCWorld @BBCBreaking @WSJ @business @nytimes @cnni @Reuters @Forbes @TIME @TheEconomist @UN @AP @washingtonpost @MarketWatch @WSJecon @FAANews @NTSB_Newsroom @FoxNews @FT @YahooFinance @SkyNews  @NBCNews