Ray Swank

[Elon Musk @elonmusk]

Built with Grok

[Jukan @jukan05]

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 @SpaceX]

SpaceXAI and @cursor_ai are now working closely together to create the world’s best coding and knowledge work AI.

The combination of Cursor’s leading product and distribution to expert software engineers with SpaceX’s million H100 equivalent Colossus training supercomputer will allow us to build the world’s most useful models.

Cursor has also given SpaceX the right to acquire Cursor later this year for $60 billion or pay $10 billion for our work together.

[eToro @eToro]

The market moves in real time... Now Tori does too! 🔥

Powered by @Grok, Tori brings live sentiment from X, remembers your context, and manages Agent Portfolios through conversation.

From insight to action, all in one place.

Tori - Your AI Edge 👇

https://t.co/tJDLqDWfrv

[Mario Nawfal @MarioNawfal]

Colossus 2. 1 gigawatt. Built in 122 days. Let that sit.

Seven models training at once. Not queued. Not waiting. Parallel.

Up to 10 trillion parameters. Real scale, not slides.

Others still stitching clusters across cloud partners. xAI just built the whole thing.

@xai

[Jon Shulkin @jon]

To all Sora Customers: xAI is open for business on image and video generation. Send X Chat or post a reply with contact info and we will get in touch.

[Shaun Maguire @shaunmmaguire]

https://t.co/KXLgdhFe1P

[Elon Musk @elonmusk]

The Grok 4.2 release candidate (public beta) is now available for use. You need to select it specifically.

Critical feedback is appreciated. Unlike prior versions of Grok, 4.2 is able to learn rapidly, so there will be improvements every week with release notes.

[Ray Swank @RaySwank]

@DKSportsbook @LAClippers = the new Skid Row, #SkidBros

[Ray Swank @RaySwank]

Awesome article by Carl Eschenbach and Kais Khimji on behalf of Team Sequoia Capital & Ben Horowitz on behalf of Andreessen Horowitz on the power of Cresta. #ExpertiseAI 🌊

"The ROI is immediate: higher conversion, better NPS,…https://t.co/WvcRCyaYMj https://t.co/BmXV7YoLwg

[Ray Swank @RaySwank]

Join Dynamic Signal's Chief People Officer, Regan Cain, and Chief Financial Officer, Bong Suh, on Friday, August 21 at 10 a.m. PT to hear how every executive at the leadership table has a role to play in improving employee wellness. #EmployeeExperience https://t.co/dErmoPEfUV

[Ray Swank @RaySwank]

Brightview Senior Living on the #DySi platform: “It means so much when our leadership is able to speak directly to all of our frontline associates during such a challenging time." #EmployeeComms #EmployeeExperience https://t.co/oZgSjPr8D4

[Matt 🅱️ @MatthiasBeezy]

Misleading headline. BR is totally twisting words to make a “story.” Listen to the entire quote from @DGreen_14 on his podcast & you’ll understand why this is inaccurate. https://t.co/RTxp02xGCI

[Ray Swank @RaySwank]

4 Reasons You’ll Miss Your Number Without Customer Success https://t.co/sQdmLZTokw

[Rand Paul @RandPaul]

Big Pharma manipulates the system to keep prices high. It is not capitalism, and it’s big government, and we’ve got to fix it.

[The Economist @TheEconomist]

Donald Trump is politically inept, morally barren and temperamentally unfit for office
https://t.co/Us8KyyYndq

[David Axelrod @davidaxelrod]

It is not going to change. It doesn't get better. It only seems to get worse.
https://t.co/FDQnuB1BNX

[Ray Swank @RaySwank]

[Register Now] Mapping the Customer Journey with Engagement Models https://t.co/Y9fO6uIMJB

[Ray Swank @RaySwank]

@MikeMcDanielACC Solid breakdown

[Mike McDaniel @MikeMcDanielSI]

Where the Spurs stand after the draft https://t.co/H8PzMRBp8A