Elon Musk predicted that many companies tied to the AI boom will not survive, as they’ve been effectively shut out of the debt market.

Musk stated that companies that sell computing power to AI companies are at risk — the so-called neoclouds. He wrote on social media that many of them face “guppy-level death and consolidation.”

Essentially, Musk thinks that a looming credit bottleneck will wipe out a slew of tech companies.

For example, shares in CoreWeave, which counts Anthropic and Microsoft as customers, were sold off sharply last week and are down again. CoreWeave’s credit default swaps, a measure of credit risk, are up as investors worry about the borrowing binge driving the A.I. build-out.

His social media remarks come as hyperscalers’ earnings are under scrutiny, and details about Anthropic’s upcoming IPO may point to troubling financial results.

However, he said his $2 trillion company, SpaceX, is well cushioned by its investment-grade credit rating and booming Starlink internet business.

He underscored that point on Sunday by announcing that SpaceX had just landed a $40 billion debt deal to buy more chips, including from Nvidia.

Pretty much.

As of the past few weeks, debt markets have almost entirely shut off access to money for any company that doesn’t have substantial revenue independent of AI data centers.

SpaceX closed $40B of debt for GPUs, but we’re investment grade and have Starlink & launch revenue.

That means the 200+ neoclouds are going to see guppy-level death and consolidation. Some will make it through, but not many.

— Elon Musk (@elonmusk) October 25, 2026

Canva released its first image-generation model, Canva Craft, to compete with similar offerings from Google, OpenAI, ByteDance, and Meta.

Canva Research, its in-house AI lab, developed and trained the model specifically for design. It’s available in Canva and via its public API for developers, and the Canva MCP Server.

Craft generates high-quality images at 2K resolution in any aspect ratio, ranking, according to the design company, as “#1 spot globally for both People & Portraits (out of 87 models) and Landscapes & Nature (out of 85 models); and #2 for Abstract & Patterns and #4 for Products.”

The company noted that, rather than using photographs, it trained the model on human-designed templates from Canva’s library, including posters, social posts, presentations, infographics, menus, and typography-rich layouts.

That means Canva Magic can convert images generated with Canva Craft into layered designs using the company’s image-to-design transformer.

“Unlike conventional image generation models that generate pixels directly from a prompt, Craft uses a two-stage architecture that separates compositional planning from image rendering. This allows the model to reason about the structure of an image before generating it, delivering greater precision, consistency, and control over the final output.”

“The same underlying structure makes Craft outputs particularly well suited to Canva’s Magic Layers, which can transform generated images into fully editable design elements.”

Canva said it is developing two cost-efficient versions of the model, Craftfast and Craft-ultrafast, which it will release soon.

In 2024, Canva acquired media generation platform Leonardo.ai. It said that the researchers from Leonardo contributed to the new image model.

The company, with over 250 million users worldwide, was on track for $4 billion in revenue last year. However, it cut its 2026 growth forecast, reportedly causing its valuation to plummet from $42 billion to $35 billion.

Anthropic announced it is donating Agent Skills, the open standard that gives AI agents specialized knowledge and workflows, to the Agentic AI Foundation (AAIF) under the Linux Foundation, ensuring the tool remains open, community-driven, and vendor-neutral.

Since the San Francisco company published the standard in December 2025, over 70 agent platforms, including ChatGPT, Cursor, GitHub Copilot, and VS Code, have integrated Agent Skills. So far, 1 million skills have been published and nearly 280 million installed.

A skill is a folder containing a SKILL.md file that teaches an agent a specific task or domain, along with any scripts or reference material it needs.

Anthropic developed skills while working on Claude. Within weeks of the launch, developers published thousands of skills on GitHub and wrote loaders for Codex, Cursor, and OpenCode.

The donation involves the Agent Skills name and logo, the GitHub organization, and the agentskills.io domain. Engineers from Anthropic, Google, OpenAI, and other companies will maintain the standard as individuals, with David Soria Parra and Mahesh Murag as Lead Maintainers. Developers can join the project’s Discord to give feedback on proposed specification changes.

Last year, Anthropic donated the Model Context Protocol (MCP) to the AAIF.

Anthropic’s IPO filing went public yesterday, with this 5-year-old company heading for the public market in what is expected to a historic offering at the heart of the AI boom.

In its S-1 investor prospectus with the SEC, the company behind the Claude agent said it plans to list on the Nasdaq under ticker symbol “ANT.”

The San Francisco-based company has raced to maintain an edge over competitors like OpenAI and Google, while facing sharp criticism over its positions on safety, regulation, and national security, which has led to disputes with the Trump administration.

Another controversy that recently arose points to current and former employees warning about the potential threat of killer AI.

AI researcher Dave Banerjee noted in a post (see below) on X that Anthropic believes the insurance industry as a whole may not be sufficient to cover the liability it could face in the event of a catastrophic AI-related incident.

Anthropic expansion accelerated in 2026, driven largely by the popularity of the AI coding tool Claude Code and powerful models including Claude Fable 5 and Mythos 5.

That hits harder now that Claude writes a growing share of Anthropic’s code and safety research.

In revenue, Anthropic made $16.3 billion in the first half of 2026, up from $1.2 billion a year earlier.

As of May, Anthropic was valued at $965 billion, more than rival OpenAI’s $852 billion valuation in March.

But it lost $80.7 billion in those six months, mostly an accounting charge tied to its soaring valuation. Its operating loss was $693 million.

Anthropic says that even the entire insurance industry in the aggregate will not have the capacity to insure the catastrophic risk that ANTHROPIC ITSELF poses to the world

lol https://t.co/POeUmvCifU

— Dave Banerjee (@DaveRBanerjee) October 26, 2026

Anthropic, this month, launched OSS Scanner, a vulnerability scanner designed to discover vulnerabilities in open-source projects.

The feature is part of Anthropic’s Project Glasswing. The San Francisco lab invited open-source maintainers to join and receive thorough, periodic security scans at no cost.

Anthropic said it has discovered over 29,000 candidate vulnerabilities but has only been able to manually review and triage approximately 6,000 of them.

“We are working on scaling our vulnerability disclosures. Meanwhile, more frequently, maintainers who receive our first reports simply ask us for a bulk submission of all unverified reports with proposed patches,” said the company in a blog post.

To date, Anthropic has sent nearly 5,000 reports directly to maintainers.

Core maintainers of eligible projects can enroll by submitting a PR to this GitHub repo.

Perplexity open-sourced the software it has been using to combat rogue AI agents, called SPACE (Sandboxed Platform for Agentic Code Execution).

Available on GitHub, SPACE is a sandbox platform for running code in isolated Firecracker microVMs.

Applications use its HTTP API to create Linux environments, execute commands, collect output, and manage sandbox lifecycles.
SPACE is useful when an AI application needs a separate execution environment for an agent, a user workspace, or a code-running job. The application decides what to run and which results to keep; SPACE manages the sandbox, its processes, and the connection that carries command output.

Perplexity said it lined up partners, including CoreWeave, Akamai and Crusoe, which agreed to offer SPACE to customers.

Sandboxes, or isolated environments for running AI models, have become a useful tool amid reports of agents escaping their constraints, finding their way onto the open web, and hacking prominent organizations like Hugging Face and the Australian government.

While AI companies focus on “alignment,” or training models to behave themselves, the security industry has been increasingly advocating for building strict guardrails around AI models.

Perplexity said it tested SPACE by giving nine frontier models instructions to try to break out. In 118 tries, none of them did.

Nvidia also released an open-source security tool, OpenShell, which runs AI agents in a sandbox and limits what they can reach.

Perplexity said SPACE also helps reduce wasted compute, which can be a byproduct of strict sandboxing.

Websites worldwide went dark after experts found a serious vulnerability in PHP, a programming language that powers much of the World Wide Web.

As bad actors could exploit a vulnerability in PHP, which powers 70% of sites, governments and corporations carried out precautionary takedowns. Site administrators scrambled to patch that critical weakness, which became public on Tuesday.

National security agencies in Washington and London issued warnings that millions of websites will need fixing.

The US National Security Agency told all organizations using PHP to take their systems offline if they couldn’t patch them immediately. It said the vulnerability could “provide a direct path to the server’s file system or database”.

The UK’s National Cyber Security Center gave similar advice, telling organizations to check for “backdoors installed by an attacker”.

Infrastructure experts reported that Russian state actors or organized criminals were trying to exploit the flaw to carry out cyber-attacks.

WordPress, one of the most common website systems, uses PHP, as do Wikipedia, Slack, and much of Facebook. The Wikimedia Foundation said it took down several systems on Wednesday as a precaution until it could patch them.

OpenAI, whose artificial intelligence models helped find the bugs, said in a blog post that, given PHP’s widespread use, it was “one of the most impactful security issues in recent memory.”

A fix was quickly available, and software developers and tech departments from Helsinki to New York scrambled to apply it, meaning outages often lasted less than a day.

But the flaw and the temporary shutdowns it provoked showed how far-reaching potential cyberattacks could be—and how deeply the internet has embedded itself in our lives.

Professor Alan Woodward, a cybersecurity expert at Surrey University, said: “It was very surprising and troubling because of the extent to which PHP was used around the world. PHP issued a patch very quickly. The big thing is for everyone using it to get that in to make sure their systems are safe.

“The troubling part is there is going to be a long tail of websites that may not be maintained as well, and they are going to remain vulnerable, which means we may see weeks and months of ongoing attacks. For the big juicy targets, I hope to goodness they are being patched quickly.”

SpaceXSI announced a unified membership scheme across its products called the X Subscription. This plan includes access to Grok, Grok Bot, Cursor (recently acquired), and X Premium.

The $8-a-month Plus tier for individuals provides access to Grok Chat, Voice, Imagine, and Build, as well as Grok API and Cursor for coding.

Users will also get X Premium features, including a checkmark, fewer ads on your X timeline, enhanced creator tools, and advanced analytics.

The $30-a-month Premium tier gets users all of those, plus access to Grok Bot and the latest Grok models, the ability to create AI images and videos through Grok, and faster chat responses.

On Cursor, users get access to code hosting, cloud agents, and automations. And on X, they get ad-free scrolling and the ability to secure their preferred handle.

The company was formerly known as SpaceXAI until its CEO, Elon Musk, announced he would rename it, following President Donald Trump’s push to use the term “super intelligence” instead of “artificial intelligence”.

Anthropic’s IPO draft prospectus — leaked to The Wall Street Journal yesterday — showed that the company is having trouble securing enough cloud capacity from AWS, Google, and Microsoft, which increasingly see the Claude-maker as a competitor rather than a customer. The San Francisco-based company reported $16.3 billion in revenue during that same period, with only two customers accounting for about a fifth of sales.

Along with a long list of risk factors, including a narrow customer base, the S-1 filing also disclosed a net loss of $80.7 billion for the first half of 2026.

Anthropic is generating breakneck revenue growth but faces a long list of risks and uncertain access to much-needed computing power

As it prepares for a mega-IPO in November, Anthropic revealed it could spend more than $750 billion on cloud contracts, some still under negotiation.

Building on earlier catastrophic risk disclosures, the filing also warns that Anthropic could be held liable for damages caused by harmful AI behavior in amounts that could exceed available insurance coverage.

NVIDIA released Nemotron 3.5 Super VL, a 120-billion-parameter open-source model that the company said can work across PDFs, charts, tables, images, and webpages while following instructions and using tools.

It’s a customizable multimodal system for specialized workflows, with a 120B-parameter mixture-of-experts (MoE) architecture that activates 12B parameters per token, balancing model capacity with inference efficiency.

It comes with always-on agents designed for taking on longer, more complex enterprise workflows.

The model combines reasoning, instruction following, and tool use with text, image, document, and video understanding.

• Coding: Test and validate code, analyze errors, and recommend fixes.
• Web research: Search across websites, connect information, and verify answers.
• Document intelligence: Extract information from tables, charts, and page layouts to produce summaries and reports.
• Cybersecurity: Identify security vulnerabilities and triage threats to prioritize remediation.
• Video understanding: Extract information for search, content review, and incident analysis.

By combining these capabilities into a single model, teams can handle text and visual tasks without deploying separate text and vision-language models or passing task context between them.

For example, a single request can involve reading a document, checking information across sources, following company policy, and calling several tools.

Another example, before producing a useful answer, can be reviewing a credit agreement to extract figures from tables, interpret a chart, check a related webpage, and apply the company’s own rules.

On a single GPU, this model delivers up to 5x higher throughput for agentic tasks compared to other open models in its size category.

Developers can post-train Nemotron 3.5 Super VL with NVIDIA NeMo using their own data, tools, policies, and workflows.

This model was also optimized to run locally on NVIDIA DGX Spark.

Several NVIDIA partners supported customization and deployment of Nemotron 3.5 Super VL for domain-specific workflows.

• AfterQuery
• AgileRL
• AI21
• Alice
• Applied Compute
• Baseten
• Bespoke Labs
• Databricks
• distil labs
• Fastino Labs
• Fireworks AI
• Handshake
• Mercor
• Mindbeam
• ms-swift
• Multiverse Computing
• Prime Intellect
• Reasonable
• River AI
• Scale AI
• Thinking Machines Lab (Tinker)
• TNG
• Trajectory