Google DeepMind has made significant strides in using AI for cybersecurity, formally verifying Signal's security code with about 6.5 million lines of proof generated through the Lean system. Human experts reviewed the specifications before automated proofs ran, covering compiled code including protection during and after device compromise. Google plans to release the proof artifacts for community use, marking a notable shift in how AI can validate critical security infrastructure.
The work was powered by Google's Gemini 4 Carbon model, which helped five researchers complete the Signal proof work in just four weeks — a task that traditionally takes years. Carbon also helped Google's Cloud Vulnerability Research team confirm two exploitable open-source bugs. Wiz used Carbon to find zero-days at two major SaaS providers in under two hours, and the model ranks first on key cybersecurity benchmarks. In Project Fortify, Carbon achieved 100% recall on scanning and 95% precision on triage, with pricing of $2 per million input tokens and $10 per million output tokens.
Meanwhile, the AI stock market is navigating turbulence. Anthony Pompliano, CEO of ProCap Financial Inc., suggested that controversy surrounding NVIDIA CEO Jensen Huang could mark a turning point for AI stocks. The iShares Semiconductor ETF fell 21% over the month but rebounded 7% on Monday. Nvidia shares closed 4.27% lower at $178.67 during Monday's regular trading session, though they were up 1.48% in overnight trading afterward. Polymarket prices in a 7% chance of an AI downturn by year-end.
Wells Fargo upgraded its outlook on US AI infrastructure stocks from negative to neutral after semiconductor shares fell about 38% from June highs. The brokerage noted that current prices already reflect much of the expected slowdown in AI-related earnings growth, but analysts warned that slowing AI investment and rising depreciation costs could pressure earnings in 2028.
In other industry developments, Credo Technology Group reported fiscal Q1 2027 revenue that rose 114.7% to $479 million, with the stock trading at $150.91 well below its 52-week high of $308.67. Analysts expect fiscal 2027 revenue of $2.50 billion and EPS of $6.3093, while management projects more than $600 million of optical revenue in fiscal 2027. France also launched the Paris Call for AI safety and security, bringing together governments, AI companies, and researchers around a shared framework for managing risks from advanced AI systems through common evaluation standards and accountability.
Key Takeaways
- Google DeepMind formally verified Signal's security code using 6.5 million lines of AI-generated proofs via the Lean system
- Gemini 4 Carbon model helped complete Signal verification in four weeks, a task that traditionally takes years
- Carbon found zero-days at two major SaaS providers in under two hours and ranks first on key cybersecurity benchmarks
- Carbon pricing is $2 per million input tokens and $10 per million output tokens
- Pompliano says Jensen Huang controversy could mark the bottom for AI stocks amid a 21% monthly semiconductor ETF decline
- Nvidia shares closed at $178.67, down 4.27% on Monday, with Polymarket pricing a 7% chance of AI downturn by year-end
- Wells Fargo upgrades AI infrastructure stocks to neutral after 38% decline from June highs
- Credo Technology Group posts 114.7% revenue growth to $479 million, with $2.50 billion expected for fiscal 2027 and over $600 million in optical revenue
- France launches the Paris Call for AI safety and security focused on frontier model risk evaluation and accountability
- Over one third of Mastercard's services already use AI, with AI revenue growing faster than other services
DeepMind proves Signal security code using 6 million lines
Google DeepMind used AI to formally verify Signal's security code. The team generated about 6.5 million lines of proof using Lean. Human experts reviewed specifications before the automated proofs ran. The verification covers compiled code, including protection during and after device compromise. Google plans to release the proof artifacts for community use.
Google expands Gemini 4 Carbon access after Signal code verification
Google expanded access to Gemini 4 Carbon for paying developers and businesses. Agents built on Carbon helped formally verify core security guarantees in the Signal protocol. The model also helped Google's Cloud Vulnerability Research team confirm two exploitable open-source bugs. Five researchers completed the Signal proof work in four weeks, a task that traditionally takes years. Google is publicly releasing the specifications and proofs.
Google's Carbon model finds zero-days in two hours
Google DeepMind's Carbon model ranks first on key cybersecurity benchmarks. Wiz used Carbon to find zero-days at two major SaaS providers in under two hours. Google's Cloud Vulnerability Research used Carbon to confirm two bugs were exploitable at PHP2Shell severity. In Project Fortify, Carbon achieved 100% recall on scanning and 95% precision on triage. Carbon costs $2 per million input tokens and $10 per million output tokens.
Pompliano says Jensen Huang affair may mark bottom for AI stocks
Anthony Pompliano, CEO of ProCap Financial Inc., suggested that controversy surrounding NVIDIA CEO Jensen Huang could mark a turning point for AI stocks. He proposed it would be objectively hilarious if the affair marked the bottom of the AI sell-off. The iShares Semiconductor ETF fell 21% over the month but rebounded 7% on Monday. Polymarket prices in a 7% chance of an AI downturn by year-end. Nvidia shares closed 4.27% lower at $178.67 during Monday's regular trading session.
Pompliano sees Jensen Huang affair as potential AI stock bottom
Anthony Pompliano, CEO of ProCap Financial Inc., believes the controversy around NVIDIA CEO Jensen Huang could signal a bottom for AI stocks. Most investors would focus on negative fallout, but Pompliano sees a contrarian possibility. The iShares Semiconductor ETF dropped 21% over the month but rebounded 7% on Monday. Polymarket currently prices in a 7% chance of an AI industry downturn by year-end. Nvidia shares were up 1.48% in overnight trading after closing 4.27% lower at $178.67 on Monday.
Wells Fargo upgrades AI infrastructure stocks to neutral
Wells Fargo changed its outlook on US AI infrastructure stocks from negative to neutral. The upgrade came after semiconductor shares fell about 38% from June highs. The brokerage noted that current prices already reflect much of the expected slowdown in AI-related earnings growth. However, analysts warned that slowing AI investment and rising depreciation costs could pressure earnings in 2028.
AI use in holiday shopping this year
The holiday season brings gift shopping with new AI technologies. Many retailers now use artificial intelligence and machine learning to enhance the shopping experience. AI is being used in various ways during holiday shopping, offering benefits and drawbacks for consumers. The article explores what shoppers can expect from this technology in coming years.
AI trading technology targets fintech stock opportunities
AI-powered trading systems combine market analysis and performance tracking to evaluate opportunities across technology-driven financial companies. Tickeron's technology uses algorithmic trading and AI-powered market analysis. The company offers services including Daily Buy/Sell Signals and AI Trading Robots. Financial Learning Models support the analysis of financial market behavior and trading automation development.
France launches Paris Call for AI safety and security
France launched the Paris Call for AI safety and security on 10 Nov 2026. The initiative brings governments, AI companies, researchers and civil society together around a shared framework. The declaration focuses on managing risks from advanced AI systems through common evaluation standards and accountability. Three main areas for cooperation were identified: evaluating frontier model risks, clearer responsibility across the AI ecosystem, and stronger information sharing.
Wall Street may underestimate this AI chip stock growth
Credo Technology Group fiscal Q1 2027 revenue rose 114.7% to $479 million. The stock trades at $150.91, well below the 52-week high of $308.67. The average analyst price target is $284.05, meaning 88% upside. For fiscal 2027, analysts expect revenue of $2.50 billion, up 87%, and EPS of $6.3093, up 82%. Management expects more than $600 million of optical revenue in fiscal 2027.
Wintermute CEO warns AI could wipe out mid-tier market makers
Wintermute CEO Evgeny Gaevoy spoke at the Digital Asset Summit in London. He said the middle tier of market making is failing because firms cannot afford the AI arms race. Wintermute has 160 employees and enough capital to grow. He believes new top-tier firms are unlikely to form. He advises new trading firms in 2026 to expect small teams and limited pay growth.
Mastercard says AI and security are driving its services growth
Mastercard's value-added services business makes up 40% of its revenue and is growing in the high teens. Chief Services Officer Linda Kirkpatrick said AI and security products are key growth drivers. Over one third of Mastercard's services already use AI, and AI revenue is growing faster than other services. The company also offers agentic commerce tools like Agent Pay and Agent Connect, though adoption is still small.
SafeCache framework improves safety of LLM response reuse
Semantic response caching helps LLMs run faster by reusing answers to similar questions. But it can be unsafe when safety conditions change between queries. Researchers proposed SafeCache, a framework that checks safety conditions before reusing a cached response. SafeCache scores risk by matching cached conditions to new requests and uses a contradiction guard to block conflicts. Experiments show it offers a better safety and utility trade-off than existing methods.
Sources
- DeepMind proved Signal code with 6M lines
- Google Widens Gemini 4 Carbon Access After AI Agents Verify Signal Code
- Google's Carbon hunts 0-days in two hours
- 'Objectively Hilarious': Anthony Pompliano Sees Jensen Huang Affair Potentially Marking the Bottom for AI Stocks — What Do Crypto Bettors Think?
- 'Objectively Hilarious': Anthony Pompliano Sees Jensen Huang Affair Potentially Marking the Bottom for AI Stocks — What Do Crypto Bettors Think?
- Wells Fargo upgrades US AI infrastructure stocks to neutral after chip rout
- Are you using AI for your holiday shopping this year? 🛍️
- AI Trading Technology Targets Fintech and Digital Infrastructure Stock Opportunities (SQ, HOOD, COIN, SOFI, AFRM)
- France launches Paris Call for AI safety and security
- Wall Street May Be Underestimating This AI Chip Stock's Next Growth Phase
- Trading firm CEO investing $1bn in AI suggests an entire class of market maker will be wiped out
- Mastercard Sees AI and Security Driving Growth in Its Fast-Rising Services Business
- Beyond Query Similarity: Safety-Conditioned Cache Routing for LLM Response Reuse
Comments
Please log in to post a comment.