When Regulation Meets Reality and Siri Stays Silent: Your February 12th Tech Roundup

 Apologies for getting this out a bit later than usual today. Sometimes life happens even when tech news doesn't slow down. And trust me, it hasn't. If anything, the pace is accelerating in ways that would have felt impossible just months ago. Today's update brings us political money flooding into AI regulation debates, Apple struggling to deliver on promises it made nearly two years ago, a startup doubling its valuation in less than five months, and Microsoft declaring independence from its biggest AI partner. Let's get into it.

Anthropic Drops 20 Million Dollars Into the Fight Over AI Regulation

The gloves are officially off in the battle over AI regulation. Anthropic announced today that it's donating 20 million dollars to Public First Action, a political group backing congressional candidates who support AI safety regulations. This isn't just corporate posturing. It's a direct counter to Leading the Future, a pro-AI PAC that's already raised 125 million dollars and is backed by heavy hitters like Andreessen Horowitz, OpenAI co-founder Greg Brockman, and investor Ron Conway.

Public First Action has already started running six-figure ad campaigns supporting candidates like Republican Marsha Blackburn in Tennessee and Nebraska's Pete Ricketts, both of whom favor AI regulations over a hands-off approach. Anthropic's move puts real money behind a position that's been politically risky in Silicon Valley. The company has faced criticism from the White House, specifically from President Trump's AI advisor David Sacks, who accused Anthropic of regulatory capture and fear-mongering back in October.

What makes this particularly interesting is the underlying philosophy. Anthropic has positioned itself as the safety-first alternative in the AI race, emphasizing governance and transparency. That pitch helped it raise hundreds of billions in valuation, as I covered in yesterday's post about Blackstone's investment. Now it's taking that same pitch to voters and politicians, betting that public sentiment favors regulation even if it slows innovation.

For you, this matters because it signals the beginning of AI becoming a genuine election issue. A Gallup poll from September showed 80 percent of respondents favor regulations for AI safety and data protection, even if it slows development. That's not a niche concern. That's mainstream anxiety about a technology moving faster than most people can track. The outcome of these political battles will shape what AI tools you have access to, how your data gets used, and whether companies face real consequences for unsafe deployments.

This isn't theoretical anymore. As I mentioned in previous posts about California's SB 53 and OpenAI's compliance issues, regulation is moving from debate to enforcement. Now it's moving into campaign ads and midterm races. The AI industry is no longer just competing on technology. It's competing on policy, and the side with more money and better messaging will shape the rules for the next decade.

Apple's New Siri Delayed Again Because It Still Doesn't Work

If you've been waiting for Apple to finally deliver on its promise of an AI-powered Siri, you'll be waiting longer. Bloomberg reported today that Apple is considering delaying the new Siri yet again, pushing features originally planned for March into May, September, or potentially beyond. The problems are the same ones Apple has been dealing with for months. Slow processing times, accuracy issues, bugs that cut off users who speak too quickly, and integration problems with both its own AI models and third-party services like Google's Gemini and OpenAI's ChatGPT.

Apple first teased the new Siri at its Worldwide Developers Conference in June 2024. That's nearly two years ago. Since then, the company has repeatedly pushed back timelines, and now it's looking at a piecemeal rollout that spreads features across multiple iOS updates throughout 2026. Some capabilities, like analyzing personal data to find an old text message or performing multi-step tasks within apps, have been scaled back or shelved entirely due to privacy concerns and technical limitations.

What's striking is the contrast. While Anthropic, OpenAI, and others are deploying increasingly capable AI models at breakneck speed, Apple can't get its voice assistant to reliably handle basic requests without lagging or failing. The company recently announced it would use Google's Gemini for future Siri versions, but internal testing shows the system still accidentally reverts to OpenAI's technology from an earlier partnership. It's a mess, and it's getting harder to see how Apple closes the gap.

For you, this delay is more than just an inconvenience. It's a signal that even the world's most valuable company is struggling with the same integration challenges that are slowing AI adoption across the industry. I've talked in previous posts about how AI capabilities are advancing faster than the infrastructure to deploy them safely and reliably. Apple is the poster child for that problem. The company has the resources, the talent, and the ecosystem, but it still can't ship a product that meets its own standards.

If you're an iPhone user, expect to keep waiting. If you're following the AI space more broadly, this is a reminder that capability and deployment are two very different things. The models exist. Making them work seamlessly in consumer products is still the hard part.

Modal Labs Talks New Funding at 2.5 Billion Dollar Valuation

While Apple struggles with deployment, the companies building AI infrastructure are seeing their valuations explode. Modal Labs, a startup specializing in AI inference, is reportedly in talks to raise a new funding round at a 2.5 billion dollar valuation. That's more than double the 1.1 billion dollar valuation it achieved just five months ago in September.

General Catalyst is reportedly leading the round, and Modal's annualized revenue is around 50 million dollars. For context, AI inference is the process of running trained AI models to generate responses, and it's where costs add up quickly for companies deploying AI at scale. Modal's competitors, Baseten and Fireworks AI, recently raised at valuations of 5 billion and 4 billion dollars respectively, so there's clearly investor appetite for companies solving this bottleneck.

What this signals is a shift in where the value is concentrating. Training AI models gets the headlines, but inference is where the money gets spent long-term. Every time you use ChatGPT, Claude, or any other AI tool, inference is happening in the background. Companies that can make that process faster, cheaper, and more efficient are becoming essential infrastructure.

For you, this means the AI tools you use daily are backed by an entire ecosystem of companies you've probably never heard of. These companies aren't building chatbots or image generators. They're building the plumbing that makes those tools work at scale. As AI becomes embedded in more products, the infrastructure layer becomes more valuable, and the cost of that infrastructure gets passed down to consumers in the form of subscription prices and usage limits.

I've talked before about how AI is transitioning from experimental to essential. Modal's valuation surge is another data point. The market is betting that inference infrastructure will be a multi-billion dollar industry, and the companies that control it will have significant leverage over everyone building on top of it.

Microsoft Pushes Toward AI Self-Sufficiency and Reduces OpenAI Dependence

Speaking of infrastructure, Microsoft just made its intentions clear. The company's AI chief, Mustafa Suleyman, told the Financial Times that Microsoft is pursuing true self-sufficiency in AI by building its own powerful models and reducing its reliance on OpenAI. This comes after the two companies renegotiated their partnership earlier this year, giving Microsoft more flexibility to develop its own technology.

Suleyman said Microsoft is building models at significant scale using the same hardware resources as competitors, and the company's new Maia 200 chip is designed specifically to reduce dependence on Nvidia's dominance in AI processing. He also hinted that new Microsoft AI models could debut sometime in 2026, and the company is positioning these as professional-grade tools capable of handling routine tasks for knowledge workers.

This is a big shift. Microsoft has been deeply tied to OpenAI for years, investing billions and integrating OpenAI's models across its product lineup. Now it's publicly signaling that it wants to stand on its own. That doesn't mean the partnership is over, but it does mean Microsoft is hedging its bets and building alternatives in case OpenAI's trajectory doesn't align with its needs.

For you, this could mean more choice in AI tools, or it could mean more fragmentation as every major company builds its own models and ecosystems. Microsoft's push for self-sufficiency reflects the same dynamic I covered in yesterday's post about cofounder exits at xAI. The pressure to compete is so intense that even established partnerships are getting renegotiated. Companies that were collaborating are now competing, and the pace of development is only going to increase.

If you use Microsoft products, expect to see more AI features powered by Microsoft's own models rather than OpenAI's. If you're watching the industry, this is another sign that the AI landscape is consolidating and fragmenting at the same time. The biggest players are building everything in-house, and smaller companies are fighting for whatever niche they can carve out.

Quick Hits: Instagram, Lenovo, and the Broader Picture

A few other stories worth noting. Instagram's head, Adam Mosseri, testified in court this week that he doesn't believe users can be clinically addicted to the app, though he admitted problematic use is possible. The case is the first of over 1,500 lawsuits accusing Meta of designing features that harm young users' mental health. Mosseri compared heavy Instagram use to watching too much TV, but given the scale of litigation, it's clear the courts aren't buying that framing.

Meanwhile, Lenovo reported record revenue of 22 billion dollars for its third quarter, driven by a 72 percent year-over-year increase in AI-related sales. AI now represents nearly a third of Lenovo's total revenue, with strong demand for AI devices, servers, and infrastructure. The company is betting big on AI moving from experimentation to production, and so far the numbers back that up.

Wrapping It All Up

Today's stories fit into a pattern that's been building for weeks. The AI industry is splitting into two tracks. One track is moving incredibly fast, with startups doubling valuations in months and major companies pouring billions into infrastructure and political influence. The other track is struggling with the basics, unable to ship reliable products or meet timelines that were set years ago.

Regulation is no longer a distant threat. It's happening now, and companies are spending real money to shape it. Deployment remains the hardest problem, even for companies with unlimited resources. And the market is rewarding infrastructure over innovation, betting that the companies controlling the plumbing will outlast the companies building on top of it.

For you, the takeaway is the same as it's been for weeks. The tools you use are going to keep changing, the companies behind them are locked in existential competition, and the rules governing all of it are being written in real time. None of this is settled. All of it matters. And the decisions being made right now will shape what AI looks like for the rest of the decade.

That's it for today. I'll be back tomorrow with more updates as this transformation continues.

Sources:

https://www.anthropic.com/news/donate-public-first-action

https://www.cnbc.com/2026/02/12/anthropic-gives-20-million-to-group-pushing-for-ai-regulations-.html

https://www.reuters.com/legal/government/anthropic-donate-20-million-us-political-group-backing-ai-regulation-2026-02-12/

https://thehill.com/policy/technology/5735472-anthropic-donates-20m-public-first/

https://www.wsj.com/tech/ai/anthropic-enters-midterm-election-showdown-over-ai-regulation-1541556e

https://www.nytimes.com/2026/02/12/technology/anthropic-super-pac-openai.html

https://www.bloomberg.com/news/articles/2026-02-11/apple-s-ios-26-4-siri-update-runs-into-snags-in-internal-testing-ios-26-5-27

https://www.investors.com/news/technology/apple-stock-aapl-siri-release-delayed-again/

https://techcrunch.com/2026/02/11/apples-siri-revamp-reportedly-delayed-again/

https://www.cnet.com/tech/services-and-software/stay-patient-apple-fans-siri-ai-delayed-again-to-late-2026-or-beyond/

https://www.macrumors.com/2026/02/11/siri-features-delayed-ios-26-4/

https://gizmodo.com/long-delayed-siri-functions-are-reportedly-being-delayed-once-again-because-theyre-slow-and-inaccurate-2000720246

https://techcrunch.com/2026/02/11/ai-inference-startup-modal-labs-in-talks-to-raise-at-2-5b-valuation-sources-say/

https://finance.yahoo.com/news/ai-inference-startup-modal-labs-224835401.html

https://techfundingnews.com/can-modal-labs-hit-2-5b-as-ai-inference-becomes-the-new-infrastructure-gold-rush/

https://seekingalpha.com/news/4550903-microsoft-eyes-ai-self-sufficiency-after-openai-deal-rejig

https://finance.yahoo.com/news/microsoft-building-own-ai-model-153309747.html

https://www.indiatoday.in/technology/news/story/microsoft-ai-boss-mustafa-suleyman-says-ai-will-replace-most-white-collar-jobs-in-the-next-2-3-years-2675849-2026-02-12

https://www.cnn.com/2026/02/11/tech/instagram-chief-trial-social-media-addiction

https://apnews.com/article/instagram-social-media-addiction-trial-mosseri-los-angeles-a49f64e21682e8bc1b98fbdf3d061579

https://www.wsj.com/business/earnings/lenovo-posts-earnings-beat-on-ai-driven-demand-293e2211

https://www.crn.com/news/data-center/2026/lenovo-restructures-data-center-unit-as-ai-drives-double-digit-growth-to-record-sales

https://finance.yahoo.com/news/lenovo-group-ltd-lnvgf-q3-130111555.html