Anthropic releases Opus 5.5 with lower prices and Fable-level performance
9:30 AM PDT · September 22, 2026
Anthropic’s newest model, Opus 5.5, was released on Tuesday, setting a new state-of-the-art in coding and knowledge work performance, according to the company. Notably, the company says, the release outpaces the larger Fable model in many benchmarks, and succeeded in a number of informal tasks that Fable failed to complete.
The new model is also significantly cheaper than its predecessor. Output tokens will be charged at $20 per million tokens for Opus 5.5, compared to $25 for the previous model. Other metrics have similar price drops. The model is also faster to run, reflecting an overall drop in the compute required to serve it.
The new version also makes significant changes to how Opus communicates, with the Opus 5.5 less likely to use jargon and more likely to put important information at the start of its messages.
The launch comes just two months after the release of Opus 5 on July 24th. According to the announcement, Sonnet 5.5 and Haiku 5.5 will be released “in the coming weeks,” with similar performance improvements.
Anthropic says that Opus 5.5 is comparable to Mythos in its biology and cybersecurity capabilities, so its release is subject to the same safeguards as the company’s Fable model. Those safeguards limit how much the models can be used to discover exploits in compiled programs or developing recognizable biological weapons, among other tasks.
Opus 5.5 is Anthropic’s first model release since CEO Dario Amodei embraced calls to pace the frontier, deliberately slowing down progress on AI capabilities to match the rate of progress on alignment.
“I have become convinced that fully addressing the risks requires even more prudence,” Amodei wrote in a post earlier this month, “not just investing in risk prevention, but pacing the rate of capabilities advancement so that risk prevention has time to keep up.”
Opus 5.5’s safety training was broadly similar to its predecessors, with alignment testing and pre-release evaluation by outside organizations like METR and Frontier Design. But Anthropic emphasized that more advanced training and evaluation systems were already being prepared for future models, including improved security and monitoring systems.
“As AI becomes more capable, public policy should play a larger role in making sure the systems people rely on are safe. That capacity takes time to build, and we’ve started to put the infrastructure in place to support it,” the blog post reads. “We expect to share more details on these efforts soon.”
Nscale’s IPO will test Wall Street’s appetite for concentrated AI bets once again
5:23 AM PDT · September 22, 2026
When British neocloud Nscale goes public, it will test public investors’ appetite for a stock whose revenue is tied primarily to two customers.
Since it was spun out of Australian cryptocurrency mining company Arkon Energy two years ago, Nscale has amassed over $103 billion worth of contracts, according to its IPO filing. But there’s a catch: Most of that, about 85%, comes from a deal to supply Microsoft with $43.8 billion worth of compute through 2033, and another supply agreement worth $44.6 billion with Anthropic.
Moreover, Anthropic’s agreement is contingent on Nscale obtaining financing, and the AI lab retains the right to walk away from or cancel the deal if Nscale fails to hit milestones that the filing explicitly categorizes as “stringent.”
Nscale’s customer concentration is a reminder of just how interconnected the AI industry has become. A recent paper by credit hedge fund Sona Asset Management, featured in Financial Times, found that many AI infrastructure providers heavily depend on a limited number of customers. Nscale’s competitor CoreWeave, for example, generates 67% of its revenue from Microsoft, and data center builder Applied Digital derives 67% of its revenue from Oracle, and 30% from CoreWeave.
While Sona noted that such interconnectedness is not necessarily a bad thing, it pointed out that a single setback or strategic shift by a major player can easily affect the entire industry.
Nscale, which plans to list on the NYSE, expects to be valued at $35 billion, Financial Times reported, and is seeking to raise $3 billion in the offering, according to Bloomberg.
The company reported revenue of $140.6 million for the six months ended June 30, up significantly compared to $10.4 million a year earlier. Net losses jumped to $1.02 billion from $369 million in the same period.
Earlier this month, one of Nscale’s major investors, Nvidia, agreed to provide the company with $1 billion in convertible debt as part of a larger $3.1 billion financing deal. The startup was valued at $14.6 billion when it raised a $2 billion Series C led by Aker ASA and 8090 Industries.
Aside from CoreWeave, Nscale’s competitors include Nebius, Lambda, and Crusoe — the latter last week said it raised $3.9 billion at a $30.9 billion valuation.
Nscale operates data centers in Norway, Portugal, Texas, and West Virginia. Its board of directors includes former Meta executives Sheryl Sandberg and Nick Clegg, as well as former OpenAI executive Fidji Simo.
World model companies are keeping a lot of secrets
1:29 PM PDT · September 20, 2026
This week, I moderated a panel on world models at the All In conference (no relation to the podcast), and it gave me a chance to dig into one of the most mysterious corners of the AI world. The big players in the space are Yann LeCun’s AMI Labs and Fei-Fei Li’s World Labs — and while both have accumulated a lot of buzz and funding, they also rank pretty low on the trying-to-make-money scale.
At their core, world models are about automating spatial intelligence, so the field could head in lots of exciting and lucrative directions, from robotics to interactive video to more complex self-driving systems.
But when I started to press on where we would actually see the tech commercialized, things got foggy. The closest thing I found to an authority was Michael Rabbat, a co-founder of AMI Labs and the company’s VP of World Models, who joined me on the panel. But when I pressed him on exactly what the company was working on, he was cagey. “We’ll talk about it when we’re ready to talk about it.” Over email, he clarified, “We’re still in a research and building phase, so we’re not talking publicly about any product plans or timeline.”
To be fair, AMI is less than a year old, so it’s fair enough to keep quiet. But this sort of caginess extends to the whole world-modeling space. World Labs’ Marble is probably the most fully developed product in the space, and its demos range from straightforward media creation, building explorable environments for video games, or CGI effects. There are robotics use cases too, but the whole platform seems more designed to demonstrate capabilities.
That secrecy even extends to these companies’ suppliers. On the sidelines of the same conference, I spoke to Alex de Vigan, CEO of Physicl — a data supplier for the burgeoning world model business. He says he knows Physicl’s data has been useful for whatever they’re building, but he’s still in the dark about what exactly that is. “I wish they would tell us more. We could build more useful data if we knew what they were working on,” de Vigan told me.
Part of the mystery comes from how versatile world models are as an idea. The simplest version is a navigable map of the world, similar to the AI models that power self-driving cars. But the same modeling approach that helps a Waymo weave through traffic could also help a humanoid robot carry boxes, or turn a few minutes of video footage into an explorable environment. AMI has already dipped its toe in manufacturing, biomedicine, robotics, and even AI software for doctors through its Nabla partnership. Surely it won’t pursue all of those — but maybe one or two of them are standing out?
No one doubts that there are lots of viable businesses to be built on world model tech — and as long as it’s easy to fundraise, there’s no particular pressure to focus on one. In fact, there’s good reason not to. If AMI announced tomorrow that they had built a humanoid OpenClaw or a next-generation Hollywood rendering system, a lot of other labs would suddenly be very interested in the space. Soon, the lab would face potential competition from the other world model companies, the neolabs, and even OpenAI and Anthropic.
In some ways, it’s the flip side of all that easy fundraising. Your competitors can fundraise, too — and the same money that lets you build under the radar is also funding lots of potential rivals once the path to market becomes clear. But even if that competition is inevitable, it’s best if you delay it for as long as possible, which means keeping quiet about exactly what you’re building.
Cixin Liu fans will recognize this as a dark forest scenario: If you don’t know who else is in the woods, it’s best not to attract attention.
Five AI safety sessions every founder should have on their TechCrunch Disrupt 2026 agenda
8:00 AM PDT · September 22, 2026
Would you trust an AI agent with access to your company’s systems? Put your employees in an autonomous vehicle? Deploy a robot that has to make sense of an unpredictable physical world?
These aren’t hypothetical questions for AI founders anymore. As AI moves out of demos and into businesses, vehicles, robots, and autonomous agents, safety and security become part of the product.
At TechCrunch Disrupt 2026, five sessions across the AI Stage and Real World AI Stage put those challenges front and center. From securing agents to getting enterprise AI into production, here’s where founders can dig into what it takes to build artificial intelligence (AI) that people and enterprises will actually trust and use.
Building the next generation of AI? Secure your pass to Disrupt and add these five AI safety sessions to your agenda. Save up to $200 on your pass before September 25 at 11:59 p.m. PT. Get a second pass at 50% off.
1. What Anthropic Sees When Enterprises Actually Deploy Claude
Some enterprises are getting measurable value from AI. Others are still running pilots 18 months later. Why?
Anthropic Head of Applied AI Cat de Jong gets a close-up view. She works directly with enterprises putting Claude into critical workflows and sees where deployments succeed, where they stall, and what separates the two.
In “ What Anthropic Sees When Enterprises Actually Deploy Claude,” de Jong will take the AI Stage to explore what happens when companies move from experimenting with AI to putting it into production. For founders trying to sell AI into the enterprise, her experience offers a firsthand look at why some deployments deliver real value while others remain stuck in pilot mode.
Want to know what separates AI pilots from deployments that deliver real business value? Hear what Anthropic is seeing firsthand. Register for Disrupt now to save up to $200 before September 25.
2. The Agent Security Problem Nobody Is Talking About
An AI agent that can take action can also create an entirely new set of security problems. What should it have access to? What should it be allowed to do? And what happens when application-level permissions aren’t enough?
In the AI Stage session “ The Agent Security Problem Nobody Is Talking About,” Okta President of Products and Technology Ric Smith and NanoCo co-founder and CEO Gavriel Cohen will tackle agent security at the infrastructure level, including weaknesses in application-level permission models and architectural decisions founders need to consider when deploying agentic AI.
If agents are on your product roadmap, security belongs there, too. Get your ticket today to save up to $200 and add this session to your Disrupt agenda.
3. Securing the AI Enterprise: Why the Cloud Just Got a Lot More Complicated
Getting an AI product through the enterprise door takes more than an impressive demo.
Security, governance, and observability all become part of the conversation when companies consider putting AI into critical systems.
In the AI Stage session “ Securing the AI Enterprise: Why the Cloud Just Got a Lot More Complicated,” AWS VP of Security Services Rudy Mitra; Luta Security CEO Katie Moussouris; and cybersecurity veteran Wendy Nather will examine the infrastructure that enterprises need as AI takes on more autonomous roles.
For founders building enterprise AI, this is a chance to hear what happens when innovation meets the security requirements of the organizations you want as customers.
Building AI for enterprise customers? Get closer to the security requirements that can stand between your product and deployment. Secure your pass to Disrupt before September 25 to save up to $200.
4. Building AI Systems When Failure Is Not an Option
AI that operates in the physical world raises the stakes. A mistake doesn’t stay on a screen; it can affect vehicles, aircraft, industrial systems, and critical missions.
On the Real World AI Stage, Shield AI Chief Technology Officer Nathan Michael; General Motors Director of Robotics Strategy Mikell Taylor; and Waabi founder and CEO Raquel Urtasun will bring perspectives from defense and autonomous systems to one of the toughest questions hard tech founders face: How do you know when an autonomous system is safe enough to deploy?
The conversation in “ Building AI Systems When Failure Is Not an Option” will get into creating a safety culture, testing and validating AI, navigating regulatory hurdles, and building companies that can earn trust when the consequences of failure are physical.
When failure isn’t an option, testing can’t be an afterthought. See how leaders in autonomy are approaching safety at Disrupt. Explore your ticket options, and save up to $200 on your pass before September 25.
5. Robots Are Waiting for Their ChatGPT Moment. Here Is What Is Standing in the Way
Robots have a data problem. They don’t have access to the massive pools of training data that helped accelerate advances in language models and self-driving vehicles.
That gap is one of the biggest obstacles to scaling physical AI. In the Real World AI session “ Robots Are Waiting for Their ChatGPT Moment. Here Is What Is Standing in the Way,” Nvidia Inception Global Head of Physical AI Les Karpas will explore how data pipelines, simulation environments, and foundation models could help close it and what it will take to build, test, and deploy more capable and reliable robots.
For founders working in physical AI, the session tackles a big question: What will it take for robotics to reach its own ChatGPT moment and earn the trust needed to put those systems to work in the real world?
Get an inside look at what could help robotics clear its next big hurdle. Register for Disrupt today to save up to $200 before prices increase.
Five sessions. One question every AI founder needs to answer.
Can people trust what you’re building? These five conversations are part of 200+ sessions across six industry stages, roundtables, and breakouts at Disrupt, taking place October 13-15 at Moscone West in San Francisco. More than 10,000 founders, investors, operators, and tech leaders are expected, along with 250+ speakers, 300+ exhibiting startups, and countless networking meetings.
You can build a smarter model, a more capable agent, or a robot that does something no one has done before. But getting customers to trust it enough to deploy it may be one of the hardest parts of bringing it to market.
Secure your pass to Disrupt now and save up to $200 before September 25 at 11:59 p.m. PT. Save 50% on a second ticket. Hear how AI leaders are tackling the safety, security, and reliability challenges that can stand between breakthrough technology and real-world adoption.
With Tabby, a former accountant is using AI to make accountants obsolete
9:38 AM PDT · September 21, 2026
In May of 2025, Ahad Ali was a successful accountant, overseeing a team of 20 people and handling more than 2,000 returns a year. “We did bookkeeping, accounting, sales tax, taxes, everything,” Ali says. It was a stable business, and one that would have been easy to stay in.
But Ali started to notice something. His work was increasingly dictated by the frictions of accounting software, which was rarely designed with actual accountants in mind. He became frustrated with specific accounting software like QuickBooks, and soon he became frustrated with the entire idea of SaaS-based accounting.
“I realized that small businesses don’t need better accounting software,” Ali told TechCrunch. “They need less accounting software. They need something that just does it for them.”
Now, Ali is building just that. His product, Tabby, is designed to be a real-time bookkeeping interface, handling clients’ paperwork as it gives them up-to-the-minute data on their business’s profit and loss. Using Plaid to import live account data, Tabby uses AI to quickly build a real-time dashboard for a business’s financial health.
There’s also a version of Tabby for accounting firms that want a smarter alternative to QuickBooks, but Ali’s goal is to go direct to the end user as often as possible. As Ali sees it, products like Tabby have a chance to automate away the work of bookkeeping once and for all, turning it into a monthly charge on the order of internet access or payroll services. If successful, it could seriously change the calculus of operating a small business.
“You can say QuickBooks digitized bookkeeping,” Ali told TechCrunch, “but we want to make [that visible software layer] nonexistent.”
So far, it seems to be working. Fourteen months in, Tabby has 5,500 small businesses using the platform, with roughly $100,000 in annual recurring revenue. The team is still just seven people — Ali, his technical co-founder, four engineers and a go-to-market specialist — and in the process of raising $1 million in pre-seed funding. When I spoke to Ali, his team was just preparing to launch Tabby Talk, a natural language interface for the product.
There’s plenty of competition in the accounting space, particularly now that AI has shaken things up. Ali has his eyes focused on QuickBooks, which is still the biggest player — but he’ll also be facing well-funded startups like the Sequoia-backed Rillet, as well as the major labs’ own finance tools. It’s a daunting task, but Ali thinks that Tabby can carve out a niche by focusing on the same small- to medium-sized firms that helped his accounting business thrive.
“Everything we see is making us believe that there’s a path for other players, because the space is big,” Ali told TechCrunch. “There’s 30 million more businesses.”
Learn more about Tabby and the other innovative startups involved in our prestigious Startup Battlefield 200 competition at TechCrunch Disrupt, happening October 13-15 in downtown San Francisco.
