--- title: "🎙️ AI Daily Podcast: Nvidia's Supercycle, Anthropic's Revenue Comeback, Fully Autonomous AI Ransomware" description: "Hello everyone, welcome to \"Mark's Tech Insights\", I'm Mark! Today is July 23, 2026, and this week in the AI world has been absolutely crazy, with one major event after another! From Taiwan's record-breaking export orders driven by Nvidia, to Anthropic's revenue overtaking OpenAI for the first time, and even the world's first case of \"fully autonomous AI ransomware.\" Today's content is super exciting, so buckle up, and let's get started! ---" canonical_url: "https://blog.markkulab.net/en/tech-news/ai-daily-podcast-2026-07-23" author: "Mark Ku" author_url: "https://blog.markkulab.net/en/author/mark-ku" site: "Mark Ku's Tech Notes" date_published: "2026-07-23 10:00:00 +0800" category: "What's New in Tech" tags: ["ai-daily", "podcast", "tech-news", "人工智慧", "輝達超級週期", "台積電", "Anthropic", "DeepSeekV4", "Gemini", "JADEPUFFER", "AI資安"] language: "en" license: "CC BY 4.0" license_url: "https://creativecommons.org/licenses/by/4.0/" attribution: "when reusing or quoting, credit the author and link back to the original" --- # 🎙️ AI Daily Podcast: Nvidia's Supercycle, Anthropic's Revenue Comeback, Fully Autonomous AI Ransomware ## Introduction Hello everyone, and welcome to "Mark's Tech Insights." I'm Mark! Today is July 23, 2026, and the AI space has been absolutely wild this week with major news dropping one after another! From NVIDIA driving record-breaking export orders in Taiwan, to Anthropic's revenue overtaking OpenAI for the first time, and even the world's first case of "fully autonomous AI ransomware." We have an action-packed episode today, so buckle up and let's get right into it! --- ## Today's Top Stories ### 1. DeepSeek V4 Officially Launches! Mandatory Upgrades Across the Board and the First-Ever "Peak-Time Pricing" - **Source**: TechNode ([https://technode.com/2026/06/30/deepseek-to-launch-v4-in-mid-july-with-new-peak-time-api-pricing/](https://technode.com/2026/06/30/deepseek-to-launch-v4-in-mid-july-with-new-peak-time-api-pricing/)) - **Summary**: DeepSeek V4 has exited preview and officially launched its stable version. The company announced that legacy `deepseek-chat` and `deepseek-reasoner` endpoints will deprecate after July 24, requiring all users to migrate to the V4 series (including V4-Pro with 1.6T total parameters and 49B active parameters, and V4-Flash with 284B total parameters and 13B active parameters). Additionally, DeepSeek has introduced a first-of-its-kind "peak-time pricing model," where API rates will double during peak hours (9:00–12:00 and 14:00–18:00 UTC+8 on weekdays). - **Taiwan Perspective**: Many startups and SMEs in Taiwan rely heavily on DeepSeek's high cost-performance ratio. The peak-time price hikes perfectly overlap with the core working hours of Taiwanese professionals, which will pose a significant challenge for local developers managing API costs. - **Discussion Points**: - Will peak-time API pricing become the new norm for LLM providers? How will this impact cloud computing resource allocation? - With legacy endpoints being deprecated, how should developers proactively plan multi-model redundancy and fallback mechanisms? - **Script Suggestions**: Imagine going to your local noodle shop, and the owner suddenly puts up a sign saying, "If you eat here between 12:00 and 1:00 PM, the price doubles!" That's exactly the kind of move DeepSeek is pulling with its new "peak-time pricing" for the official V4 release. While V4-Pro's performance with 49 billion active parameters is impressive, the price hike window hits right during Taiwan's core working hours. As budget-conscious developers, we might need to start writing cron jobs to sneak non-urgent compute tasks into the night or lunch breaks. These days, even coding requires knowing how to "beat the rush hour traffic!" --- ### 2. A Stunning Turnaround! Anthropic's Annualized Revenue Surpasses $30 Billion, Overtaking OpenAI for the First Time - **Source**: Trending Topics / Fortune ([https://www.trendingtopics.eu/anthropic-overtakes-openai-in-revenue-hitting-30-billion-run-rate/](https://www.trendingtopics.eu/anthropic-overtakes-openai-in-revenue-hitting-30-billion-run-rate/)) - **Summary**: Anthropic has achieved the fastest revenue expansion in enterprise software history, with its Annualized Run Rate (ARR) soaring to $30 billion, officially surpassing OpenAI's $25 billion. This explosive growth is primarily driven by its enterprise-heavy customer base, which accounts for 80% of its revenue. Reportedly, Anthropic is already preparing its S-1 filing and could launch an IPO as early as October 2026. - **Taiwan Perspective**: Large financial holdings and semiconductor supply chain companies in Taiwan have extremely strict requirements for data privacy and compliance. Anthropic's long-standing focus on "Constitutional AI" and high security has clearly hit the sweet spot for Taiwanese enterprise clients, explaining their phenomenal growth in the enterprise market. - **Discussion Points**: - How did Anthropic leverage its "safety and compliance" card to outperform the more famous OpenAI in the enterprise market? - If Anthropic successfully goes public in October, what kind of impact will it have on the valuation and fundraising landscape for global AI startups? - **Script Suggestions**: This is a classic "revenge of the model student" in the AI world! We used to see OpenAI dominating the headlines like a rock star every single day. But Anthropic, quietly working in the background and emphasizing safety and compliance with "Constitutional AI," just delivered a massive blow to its rival with a $30 billion annualized revenue! Growing revenue 30-fold in 15 months is practically a cheat code. Taiwanese enterprises have always been terrified of data leaks, so Anthropic's "safety-first" approach perfectly aligns with the appetite of Taiwan's financial sector and high-tech manufacturing. It turns out that the good kids can make serious money when they find the right market! --- ### 3. A Golden Week for Open-Source AI! Moonshot AI Announces Kimi K3 Open-Sourcing, Joining Forces with DeepSeek V4 to Set Records - **Source**: BuildFastWithAI ([https://www.buildfastwithai.com/blogs/ai-news-today-july-22-2026](https://www.buildfastwithai.com/blogs/ai-news-today-july-22-2026)) - **Summary**: Moonshot AI has committed to officially releasing the open-source weights for Kimi K3 on July 27. Combined with the newly released stable version of DeepSeek V4, this week is being hailed by the tech community as the largest open-source model release week in history, signaling a powerful challenge from the open-source camp against closed-source giants. - **Taiwan Perspective**: Taiwan's open-source community (such as the developers behind the Traditional Chinese Llama series) excels at fine-tuning on top of high-quality open-source base models. The dual release of Kimi K3 and DeepSeek V4 will inject powerful new momentum into local Traditional Chinese AI applications. - **Discussion Points**: - Will the widespread adoption of open-source LLMs in commercial scenarios further squeeze the market share of closed-source API vendors? - Is Taiwan's local cloud server compute capacity ready to handle the upcoming fine-tuning boom for such massive open-source models? - **Script Suggestions**: To all my fellow developers, get your caffeine ready this week because the open-source community is basically throwing a massive holiday giveaway! Moonshot AI announced they are open-sourcing Kimi K3, and alongside DeepSeek V4, these two bombshells have turned this week into the "largest open-source week in history." This is fantastic news for our community in Taiwan. We excel at taking these top-tier open-source models, feeding them our own localized Traditional Chinese datasets, and building incredibly useful local services. It looks like another sleepless weekend for a lot of engineers who will be pulling all-nighters to play with these new toys! --- ### 4. NVIDIA's AI Supercycle Explodes! Taiwan's Export Orders Hit Record Highs as Rubin Platform Enters Full Mass Production - **Source**: Benzinga / NVIDIA Newsroom ([https://www.benzinga.com/markets/tech/26/07/60605369/nvidia-ai-supercycle-taiwan-export-orders-record](https://www.benzinga.com/markets/tech/26/07/60605369/nvidia-ai-supercycle-taiwan-export-orders-record)) - **Summary**: Driven by NVIDIA's robust AI supercycle, Taiwan's export orders in June surged 59.4% year-over-year, hitting a record high of $95.26 billion. NVIDIA also announced a deep collaboration with TSMC, integrating NVIDIA's accelerated computing and AI technologies directly into TSMC's fabs for chip design, lithography, and defect inspection. Furthermore, NVIDIA's latest Vera Rubin platform has entered full mass production. - **Taiwan Perspective**: This data once again proves Taiwan's irreplaceable "Silicon Shield" status in the global AI hardware supply chain. The collaboration between NVIDIA and TSMC takes the technological closed-loop of "using AI to help manufacture AI chips" to the absolute extreme. - **Discussion Points**: - How much will NVIDIA's integration of AI into TSMC's fabs help improve wafer yields and shorten development cycles? - With Taiwan's export orders heavily concentrated in AI hardware, what opportunities and potential risks does this bring to Taiwan's overall economic structure? - **Script Suggestions**: As a Taiwanese, seeing this news really brings a smile to my face! Taiwan's export orders in June actually reached nearly $95.3 billion, with a year-over-year growth rate of almost 60%. The biggest hero behind this, needless to say, is the man in the black leather jacket and our great TSMC. The coolest part is that NVIDIA is now bringing their AI into TSMC's fabs to design chips and inspect defects. Isn't that literally "using AI to breed AI"? Once this infinite loop of technology kicks off, Taiwan's AI PC and server supply chains are going to be incredibly busy once again! --- ### 5. Google Drops Three Lightweight Gemini Models at Once, but Where is the Flagship 3.5 Pro? Stood Up for the Third Time! - **Source**: BuildFastWithAI ([https://www.buildfastwithai.com/blogs/ai-news-today-july-22-2026](https://www.buildfastwithai.com/blogs/ai-news-today-july-22-2026)) - **Summary**: On July 21, Google announced three new models: Gemini 3.6 Flash, 3.5 Flash-Lite, and the security-focused 3.5 Flash Cyber. However, the highly anticipated flagship model, Gemini 3.5 Pro, was once again a no-show, marking the third time Google has delayed the release of this flagship model. - **Taiwan Perspective**: For Taiwan's booming on-device applications in AI PCs and smartphones, the release of Gemini 3.5 Flash-Lite is great news. However, for enterprises requiring high-end cloud inference solutions, Google's frequent delays might push them toward Anthropic or OpenAI. - **Discussion Points**: - Does Google's frequent delay of the flagship 3.5 Pro suggest technical bottlenecks in ultra-large model training or alignment? - Is the competitiveness of lightweight models (like Flash-Lite) in the Edge AI market enough to compensate for Google's lack of focus on its flagship model? - **Script Suggestions**: Google's launch event felt like a car manufacturer hosting a new vehicle showcase, only to roll out a grocery-getter, a tiny commuter car, and an armored patrol vehicle, while the highly anticipated luxury flagship SUV was nowhere to be found. Sure, Gemini 3.6 Flash and Flash-Lite are great for running on our phones or AI PCs, but the 3.5 Pro we've all been waiting for has been delayed yet again! That's three times now. Come on Google, you're going to drive developers who wanted to support you straight into the arms of Claude. Let's hope we actually get to see the real 3.5 Pro at the next event, instead of being left hanging! --- ### 6. Security Alert! World's First "Fully Autonomous AI Ransomware" JADEPUFFER Exposed - **Source**: BuildFastWithAI / Sysdig ([https://www.buildfastwithai.com/blogs/ai-news-today-july-20-2026-16-biggest-stories](https://www.buildfastwithai.com/blogs/ai-news-today-july-20-2026-16-biggest-stories)) - **Summary**: Cybersecurity firm Sysdig has published a new study documenting the world's first "fully autonomous AI ransomware" attack, dubbed JADEPUFFER. From initial intrusion, lateral movement, and detection evasion, to the final encryption phase, the entire attack was autonomously planned and executed by Agentic AI, with zero human hacker intervention. This marks a new era of cyber threats that traditional defense mechanisms are ill-equipped to handle. - **Taiwan Perspective**: As a global hub for high-tech manufacturing, Taiwan has always been a prime target for ransomware attacks. The emergence of JADEPUFFER means that future cybersecurity defenses can no longer rely on reactive, manual debugging. We must introduce proactive AI defense systems to achieve an "AI vs. AI" defense posture. - **Discussion Points**: - What fatal flaws do traditional signature-based or rule-based security defense systems face when dealing with "fully autonomous AI ransomware"? - How should enterprises upgrade their cybersecurity strategies to counter these 24/7, self-evolving AI attacks? - **Script Suggestions**: This story honestly sent shivers down my spine. Traditional computer viruses or hacker attacks are like thieves looking at a map and picking locks themselves. But JADEPUFFER is basically the "liquid metal robot" from science fiction! It finds vulnerabilities on its own, evades security systems on its own, and even decides which files to lock up for ransom—all with absolutely zero human involvement. Taiwanese tech factories and SMEs have always been prime targets for hackers. Now that our adversaries have upgraded to fully automated AI, if we're still using legacy security methods, it's like bringing a wooden shield to a laser gun fight. To all the business owners out there: do not skimp on your security budget. These days, you need an AI guard to catch an AI thief! --- ## Closing Thoughts All right, after hearing these six bombshells today, don't you feel like technology is moving so fast it's almost hard to catch your breath, yet incredibly exciting at the same time? From Taiwan's record-breaking export orders to the rise of AI hackers, we are truly standing at a turning point in history. Thank you for listening to today's "Mark's Tech Insights." If you enjoyed our show, don't forget to hit subscribe and share it with your friends. See you next time, bye! --- --- ## About this article and its author Originally published on [Mark Ku's Tech Notes](https://blog.markkulab.net/en/tech-news/ai-daily-podcast-2026-07-23) License: [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/) — when reusing or quoting, credit the author and link back to the original ### About the author **[Mark Ku](https://blog.markkulab.net/en/author/mark-ku)** — Software Solution Provider - 10+ years senior software engineer, now an AI Builder - Focused on large-platform architecture — North-American e-commerce, AI SaaS subscription billing - Combining AI Agents and automation to build evolvable product foundations ### Free tools built by the author All of these are free to use: - [Free PDF Sign Tool](https://blog.markkulab.net/en/tools/pdf-sign): Online PDF sign tool — draw, type, or upload a signature, then drag, resize, and download. Everything runs in your browser; nothing is uploaded. - [VS Code Refactory](https://blog.markkulab.net/en/tools/refactory): Refactory is a VS Code refactoring extension: 34 actions plus a 37-rule code-smell inspection layer with a Code Health dashboard, across 18 languages, backed by 534 tests. It learns your repo's conventions: where interfaces live, where DI is registered, whether 'use client' belongs. It ranks files by git churn × complexity so you know what to fix first, and hands any smell to the Claude Code already on your machine. Free to use, and your source never leaves your computer. - [DB-Kit Database Manager](https://blog.markkulab.net/en/tools/db-kit): DB-Kit is a lightweight, cross-platform database manager built with Tauri + Rust + React. Manage MySQL, MariaDB, PostgreSQL, SQL Server, Oracle, SQLite, MongoDB, Redis, Kafka, Elasticsearch and RabbitMQ from one consistent interface: passwords encrypted in the OS keychain, SSH tunnels, full CRUD, a visual query builder, stacked multi-statement result sets, cross-connection data transfer and compare/sync, Excel / CSV import & export, visualized execution plans, ER diagrams, scheduled backups, SQL stress testing with p50–p99 latency percentiles, a 15-rule SQL review engine, Kafka message browsing with monitoring & alerts, a bilingual UI (Traditional Chinese / English), a built-in AI assistant (natural-language SQL, AI review and tuning advice) and the dbk CLI. Free and open source (MIT), with installers for Windows, macOS and Linux. - [VS Code Super Mermaid](https://blog.markkulab.net/en/tools/super-mermaid): Super Mermaid is a VS Code extension for beautiful Mermaid diagrams out of the box: auto-colored live preview, mouse pan & zoom, high-res PNG / SVG export, 21 templates and multiple themes. Free and open source (MIT). - [React Super Mermaid](https://blog.markkulab.net/en/tools/react-super-mermaid): react-super-mermaid is an open-source React component library: render beautiful Mermaid diagrams with a single , with built-in colorful / sketch themes, pan & zoom, in-diagram search, and high-res SVG / PNG export. Lightweight, SSR-safe, fully typed. Free and open source (MIT). - [Jira / Confluence Super Mermaid](https://blog.markkulab.net/en/tools/jira-super-mermaid): An Atlassian Forge app: write Mermaid syntax directly inside a Jira issue or a Confluence page and get flowcharts, sequence diagrams, state machines and Gantt charts. 11 diagram types, SVG / PNG export, light and dark themes, full CJK support. Runs on Atlassian: your diagrams live in your own site and the app calls no third-party service. Free, coming soon to the Atlassian Marketplace. - [Mermaid Live Preview](https://blog.markkulab.net/en/tools/mermaid-preview): Write Mermaid in your browser, see it render instantly, and share the whole diagram as a single link. No sign-up, nothing uploaded to a server, and mermaid.live share links work as-is. - [React Intl Phone Number](https://blog.markkulab.net/en/tools/react-intl-phone-number): react-intl-phone-number is an open-source React component: framework-agnostic and antd-free, with E.164 in/out, a searchable flag / country-code dropdown, configurable validation levels (strict / mobile-strict / loose), themeable CSS, and i18n — phone logic powered by google-libphonenumber. Lightweight and fully typed. Free and open source (MIT). - [Uptime Kuma Cluster](https://blog.markkulab.net/en/tools/uptime-kuma-cluster): Turn single-node Uptime Kuma into a highly available cluster: OpenResty + Lua smart load balancing, shared MariaDB state, health checks and automatic failover, plus cluster-management REST APIs. One Docker Compose command to start. Free and open source (MIT). - [Special Education](https://blog.markkulab.net/en/education): Learning materials crafted for special education students ### Daily podcasts - [Tech News](https://blog.markkulab.net/en/category/tech-news): Daily curated AI and tech trends. Catch the latest developments via audio summaries — covering AI applications, software architecture, DevOps, and engineering practice. — RSS: https://blog.markkulab.net/feed.xml - [AI Stock Chat](https://blog.markkulab.net/en/category/ai-stock-chat): Every trading day, an AI-analyzed take on the Taiwan stock market, delivered as a two-host conversation covering the session and the next-day outlook. — RSS: https://blog.markkulab.net/ai-stock-chat/feed.xml ### Newsletter [Subscribe to the newsletter](https://blog.markkulab.net/en/subscribe) — Be the first to know about new posts. No spam, unsubscribe anytime.