Skip to main content
Back to News

OpenAI at $1.25 Trillion, AI Spending Doubts, and a Major Security Breach

Updated on November 07, 20265 minutes read

AI valuations just hit a number that would have sounded fictional five years ago, and yet cracks in the spending story are showing up in real corporate data. This week delivered a lot to unpack for anyone trying to understand where the industry is actually heading.

OpenAI's staggering new valuation

OpenAI has reportedly closed an additional $11 billion in funding, pushing its valuation to $1.25 trillion. To put that in perspective, the company was valued at $852 billion just months ago after a $122 billion raise. That kind of acceleration is almost without precedent for a private company.

For people learning to code or studying AI, the size of the number matters less than what it signals: investors still believe AI products will generate enormous revenue, even though that revenue hasn't fully materialised yet. Whether that confidence holds is a separate question entirely.

Read OpenAI's new funding details on The Verge or dig into the breakdown on TechCrunch.

AI spending is stalling, and that matters

Here's the tension: while OpenAI's valuation soars, data from corporate spending platform Ramp tells a quieter story. AI usage among businesses keeps growing, but actual spending has plateaued. Price cuts from AI providers and smarter cost controls among heavy users appear to be squeezing the revenue assumptions that underpin those giant valuations.

This is worth paying close attention to if you're entering the field. It suggests companies are becoming more selective about which AI tools they pay for, which means useful, well-scoped AI products may win out over broad, expensive ones. TechCrunch covers the Ramp data and what it could mean for Anthropic's upcoming IPO.

Chainalysis hit by a devastating breach

Blockchain analytics company Chainalysis confirmed it suffered what it described as a "multifront cyberattack." The attackers stole credentials and tokens for every single one of its customers, reportedly using AI agents as part of the operation alongside an on-path data interception technique.

This is a textbook example of a supply-chain style attack hitting a company whose customers depend on it for security intelligence. If you're studying cybersecurity, notice how the attack combined multiple vectors at once. A single layer of defence would not have been enough. Ars Technica has the full story on the Chainalysis breach.

Hundreds of AI hacking tools left wide open

Security researchers discovered hundreds of servers running Artex, a penetration testing tool powered by AI, left completely exposed on the public internet. Anyone who found them could access the tools with no authentication.

Penetration testing tools are designed to probe for weaknesses in systems, so having them unsecured is a significant problem. It's a reminder that misconfigured infrastructure is one of the most common real-world vulnerabilities, and something every developer and security learner should understand early. TechCrunch reports on the exposed Artex servers.

AI data centers: Google buys in, communities push back

Two related stories paint a fuller picture of where AI infrastructure stands. Google acquired a troubled Wyoming data center project from startup Crusoe, which had run into difficulties developing it at scale. The deal illustrates just how hard it is to build the physical foundations that AI models actually run on. TechCrunch covers the Crusoe-to-Google sale.

Meanwhile, voters in this week's midterm elections weighed in on 41 data-center-related ballot measures. In all but one, majorities backed tighter restrictions on new developments, citing concerns about power grids, local utility costs, and environmental impact. The Verge has an ongoing tracker on AI data center controversies. For anyone building a career in cloud or infrastructure, understanding community and regulatory resistance is now part of the job.

Meta's Muse gets a major new feature

Meta expanded its Muse AI assistant with a feature called Spaces, which lets users build customised, app-like environments directly inside the Muse interface. Think a personal shopping hub, a school dashboard, or a music tracker, all generated and managed by the AI without installing a separate app.

This kind of embedded, agentic experience is where several big tech companies are heading at once. For UX/UI designers and front-end developers, it raises real questions about how interfaces get designed when an AI is assembling them on the fly. The Verge explains how Meta Muse Spaces works.

Tesla ends Cybertruck production

Tesla confirmed it will wind down Cybertruck manufacturing at Gigafactory Texas in 2027, freeing up the facility to produce its self-driving Cybercab instead. The Cybertruck had a complicated commercial life, and this move suggests Tesla is placing a bigger bet on autonomous vehicles than on its polarising pickup.

For anyone interested in the EV and autonomous vehicle space, this is a signal about where Tesla sees the next wave of demand. CNET covers the end of Cybertruck production.

SpaceX announced it is pursuing a serious push to make Starlink Mobile a "major" carrier, with thousands of direct-to-cell satellites already in orbit and plans for a ground-based 5G network to complement them. The company has reportedly signed partnership deals as it works to compete directly with established carriers.

Satellite-to-cell connectivity is a genuinely different approach to mobile infrastructure, and the technical challenges of making it work at scale are substantial. For anyone studying networking or telecommunications, it's one of the more interesting engineering projects happening right now. The Verge has the full Starlink Mobile news hub.

The FAA is finally dropping floppy disks

The US Federal Aviation Administration announced a plan to remove floppy disks from its information display systems at 435 sites by the end of 2028. The agency called it a "historic" upgrade.

It sounds absurd, but legacy infrastructure in critical systems is genuinely common, and understanding why organisations hold onto it for so long is useful knowledge. Replacement cycles in regulated, safety-critical environments are slow, expensive, and risk-averse by design. Ars Technica has the story on the FAA's floppy disk phase-out.

The weeks ahead look busy. Anthropic's IPO will be a real test of whether investor appetite for AI companies holds when the numbers get scrutinised publicly, and the data center regulation story will keep developing as newly elected officials start making policy decisions. The gap between AI hype and AI economics is narrowing fast, and that's exactly the right moment to be paying attention.

Learn technical skills online with Code Labs Academy

Learn technical skills online with Code Labs Academy

Join our supportive community, unlock your potential, and embark on a rewarding career path.

Apply Now

Frequently asked questions

Why is OpenAI valued so much higher than its actual revenue?

Private company valuations reflect what investors expect the business to earn in the future, not what it earns today. OpenAI's valuation is based on the assumption that AI products will eventually generate enormous revenue. Whether that plays out depends on adoption, competition, and whether companies keep paying for AI tools as prices fall.

What is an on-path attack, like the one used against Chainalysis?

An on-path attack (sometimes called a man-in-the-middle attack) happens when an attacker positions themselves between two communicating parties and intercepts or alters the data passing between them. In Chainalysis's case, this was one part of a broader multi-vector attack that also used AI-driven techniques.

Why does it matter that AI spending is stalling even as usage grows?

If companies use more AI but spend the same amount, it means providers are charging less per unit of usage, often due to competition and price cuts. That squeezes revenue growth for AI companies and challenges the assumption that rising usage automatically translates into rising income, which matters a lot for valuations and IPO prospects.

What is a penetration testing tool and why is leaving one exposed so dangerous?

Penetration testing tools are used by security professionals to find weaknesses in computer systems before attackers do. If one is left exposed on the public internet with no access controls, anyone can use it to attack real targets, effectively turning a defensive tool into an offensive one.

Why does the FAA still use floppy disks in 2026?

Safety-critical systems like air traffic control infrastructure are updated extremely slowly because any change requires rigorous testing, regulatory approval, and careful rollout to avoid disrupting operations. Legacy hardware often remains in place for decades simply because replacing it safely is a massive logistical challenge, not because no one noticed it was outdated.

Career services

Personalized career support to help you launch your tech career. Get résumé reviews, mock interviews, and industry insights, so you can showcase your new skills with confidence.

Explore