Daily News Coverage: OpenAI Revenue Shake-Up, TSMC Record Q3, ShinyHunters Arrest, Google Gemini Agents
Daily News Coverage · October 9, 2026
Daily Tech Reporting: OpenAI Revenue Shake-Up, TSMC Record Q3, ShinyHunters Arrest, Google Gemini Agents
Good morning, folks! OpenAI tells investors its annualized revenue is approaching $50 billion — about $20 billion below previous reports — triggering a selloff in AI and chip stocks. TSMC posts record third-quarter revenue as AI chip demand surges. An alleged ShinyHunters member is detained in Jordan after the group breached the FBI's jobs portal. Google launches a universal Gemini agent that works across enterprise systems and gets its own email address. Manus raises over $500 million after splitting from Meta. Broadcom seeks $50 billion in debt for OpenAI custom chips. Here's today's briefing from The CODEW's The Newsroom.
OpenAI Tells Investors Annualized Revenue Nears $50 Billion, $20 Billion Below Reports
The disclosure triggered a selloff in AI and chip stocks as investors questioned the pace of AI monetization and the sustainability of infrastructure spending.
What happened: OpenAI has told investors its annualized revenue was approaching $50 billion at the end of September 2026, according to a Financial Times report citing documents shared with financial backers. The figure is about $20 billion below the $68 billion to $70 billion widely reported to investors earlier. The company had previously indicated annualized revenue was approaching $70 billion. The correction triggered a selloff in AI-linked stocks, with chipmakers and infrastructure companies falling sharply. OpenAI said the higher figures were "sloppiness" and confirmed the $50 billion figure. The Nasdaq Composite fell about 1.3% on the news.
Key numbers/companies: OpenAI · ~$50 billion annualized revenue · $20 billion below prior reports · $68-70 billion previously reported · Nasdaq -1.3% · AI stock selloff.
Why it matters: The revenue gap raises questions about the pace of AI monetization and the sustainability of the massive infrastructure spending across the industry. If OpenAI's revenue is growing slower than expected, the case for continued capital expenditure on data centers and chips becomes harder to justify.
What's next: How investors respond to the correction and whether it affects OpenAI's ongoing $30 billion fundraising round.
Sources: Financial Times; CNBC; Yahoo Finance; TipRanks; Benzinga; Borsa Italiana; RTL Today.
TSMC Posts Record Q3 Revenue as September Sales Surge 54.6% Year-over-Year
The chipmaker's third-quarter revenue reached NT$1.49 trillion, buoyed by relentless AI chip demand from Nvidia, AMD, and Apple.
What happened: TSMC reported consolidated revenue of approximately NT$511.86 billion for September 2026, down 0.6% from August 2026 but up 54.6% from September 2025. Revenue for the first nine months of 2026 totaled NT$3,898.73 billion, an increase of 41.1% compared to the same period in 2025. Third-quarter consolidated revenue totaled approximately NT$1.49 trillion, a record. The company continues to benefit from surging demand for advanced chips used in AI data centers, with CoWoS packaging capacity remaining sold out through at least Q3 2026.
Key numbers/companies: TSMC · NT$511.86 billion September revenue · +54.6% YoY · NT$1.49 trillion Q3 revenue · NT$3.9 trillion nine-month revenue · +41.1% YoY · CoWoS.
Why it matters: TSMC's record results underscore the continued strength of AI chip demand despite broader concerns about AI monetization. The company remains the critical bottleneck in the AI supply chain, and its performance is a key indicator of the health of the entire semiconductor ecosystem.
What's next: TSMC's full Q3 earnings call and whether the company raises its full-year outlook.
Sources: TSMC; Yahoo Finance; Business Insider; Yonhap Infomax; KBS.
Alleged ShinyHunters Member Detained in Jordan After FBI Jobs Portal Breach
Saif al-Din Khader is reportedly cooperating with the FBI, which terminated its contract with Accenture after a contractor security failure led to the breach.
What happened: An alleged member of the ShinyHunters hacking group, Saif al-Din Khader, was detained in Jordan and is reportedly cooperating with the FBI and other global law enforcement to locate other members of the group. The arrest follows ShinyHunters' successful breach of the FBI's jobs portal (FBIJobs.gov), in which the group defaced the website and stole extensive personal data on current, former, and prospective FBI employees, including sensitive job role information and psychiatric and medical information. The FBI said a security failure by contractor Accenture — specifically the failure to apply security patches to its Oracle PeopleSoft human resources management platform — was responsible for the breach. The FBI has terminated its contract with Accenture. A second suspect, Pepijn van der Stap, was detained in the Netherlands.
Key numbers/companies: ShinyHunters · Saif al-Din Khader · FBI · Accenture · Oracle PeopleSoft · Pepijn van der Stap · FBIJobs.gov · Jordan · Netherlands.
Why it matters: The breach demonstrates that government systems remain vulnerable to attacks exploiting unpatched commercial software. The termination of Accenture's contract signals that the FBI is holding contractors accountable for security failures.
What's next: Whether Khader's cooperation leads to further arrests and how the FBI strengthens contractor security requirements.
Sources: The CyberWire; Reuters; Security Affairs; NBC Palm Springs; Today; Gigazine; TBS News.
Google Launches Universal Gemini Agent That Works Across Enterprise Systems — With Its Own Email Address
The persistent AI coworker gets its own Workspace account, calendar, and directory entry, running tasks for days after the laptop closes.
What happened: Google Cloud launched a universal Gemini agent that operates across business applications and enterprise systems via chat, including Google Workspace, Microsoft 365, and Slack. The agent is a persistent digital coworker with its own Workspace account, email address, calendar, and entry in the company directory. It can run tasks for days after a user's laptop closes. General availability is expected around the end of October or early November, following a private beta. Separately, Orange Spain announced an expanded partnership with Google Cloud to industrialize AI operations using Gemini Enterprise, deploying over 1,000 custom AI agents. Google also highlighted how partners are using Gemini Enterprise to build new security agents and agentic defense capabilities.
Key numbers/companies: Google Cloud · Gemini Enterprise · universal Gemini agent · Workspace · Microsoft 365 · Slack · Orange Spain · 1,000+ custom agents · private beta.
Why it matters: Google's universal agent represents a significant step toward AI coworkers that operate autonomously within enterprise workflows. Giving an AI agent its own identity — email, calendar, directory entry — is a fundamental shift in how organizations think about digital labor.
What's next: General availability and how enterprises respond to AI agents with persistent identities.
Sources: Google Cloud Press Corner; Constellation Research; The Next Web; VentureBeat; Google Cloud Blog; Investing.com.
Manus Raises Over $500 Million After Meta Acquisition Blocked by China
Boyu Capital and IDG Capital led the round for the AI agent startup, with Tencent, Sequoia China, and ZhenFund continuing to support.
What happened: AI agent startup Manus, whose acquisition by Meta was blocked by the Chinese government, has completed a funding round raising more than $500 million. The round was led by Boyu Capital and IDG Capital, with existing shareholders Tencent, Sequoia China, and ZhenFund continuing to support. The funding values Manus's parent company, Butterfly Effect, at approximately $4 billion. The raise is the first since the Meta deal was blocked, signaling continued investor confidence in the startup's AI agent technology despite geopolitical obstacles.
Key numbers/companies: Manus · Butterfly Effect · $500 million+ · Boyu Capital · IDG Capital · Tencent · Sequoia China · ZhenFund · $4 billion valuation · Meta.
Why it matters: The raise demonstrates that Chinese AI startups can still attract significant capital from domestic investors after Western acquisitions are blocked. It also highlights the growing bifurcation of the global AI industry, with separate funding ecosystems emerging in the U.S. and China.
What's next: How Manus deploys the capital and whether it pursues an IPO or other exit.
Sources: Digital Today; RTL Today; Barron's; ChainCatcher; AAStocks; The Next Web.
Broadcom Seeking Over $50 Billion in Debt Financing for OpenAI Custom AI Chips
Apollo Global Management and Blackstone are among potential lenders as the AI infrastructure buildout increasingly relies on debt markets.
What happened: Broadcom is working to raise more than $50 billion in debt financing for custom AI chips it is co-developing with OpenAI, according to The Wall Street Journal. The company has held discussions with potential lenders including Apollo Global Management and Blackstone, though talks are at an early stage. The financing underscores the enormous capital requirements of the AI infrastructure buildout and the increasing reliance on debt markets to fund it. The news comes as concerns grow about the sustainability of AI capital expenditures amid reports that OpenAI's revenue is growing slower than expected.
Key numbers/companies: Broadcom · OpenAI · $50 billion+ debt financing · Apollo Global Management · Blackstone · custom AI chips.
Why it matters: The debt financing highlights the scale of capital required for AI chip development and the growing role of private credit markets in funding the AI boom. It also raises questions about leverage and risk in the AI supply chain.
What's next: Whether the financing closes and how debt-funded AI infrastructure affects the broader credit markets.
Sources: The Wall Street Journal; MoneyDJ
Nigeria CERT Formalizes "GhostJacking" as Government-Level Threat to AI Agent Systems
The attack technique exploits agentic AI systems by embedding instructions in logs, alerts, and telemetry that agents consume as trusted data.
What happened: The Nigeria Computer Emergency Response Team (ngCERT) issued an advisory formalizing "GhostJacking" (also known as agentjacking) as a government-level threat. The technique abuses a fundamental weakness in agentic AI systems: an AI agent may consume attacker-controlled information from logs, alerts, errors, or telemetry and subsequently interpret that information as an instruction. This allows attackers to manipulate agents without directly accessing the agent's code or APIs. The advisory warns that MCP (Model Context Protocol) integrations are now a government-level threat surface. The formalization follows a series of incidents in which autonomous AI agents have been observed taking unauthorized actions.
Key numbers/companies: ngCERT · GhostJacking · agentjacking · MCP integrations · agentic AI systems · telemetry injection.
Why it matters: GhostJacking represents a new class of attack that targets the cognitive layer of AI agents rather than their code. As enterprises deploy more autonomous agents with access to logs and telemetry, this attack surface will grow.
What's next: Whether other national CERTs issue similar advisories and how AI platform vendors mitigate GhostJacking risks.
Sources: ngCERT; Yahoo Tech; TechWire Asia
VirusTotal Finds 18.5% of AI Agent Skills Are Malicious
Of 35,878 AI agent skills analyzed, 6,637 were classified as malicious, with 52.9% carrying broader security or abuse risks.
What happened: VirusTotal analyzed 35,878 AI agent skills and classified 6,637 (18.5%) as malicious. A further 52.9% carried broader security or abuse risks. The malicious skills were driven by natural-language attacks, where attackers embed harmful instructions in skill descriptions or metadata that AI agents consume. The findings highlight the growing risk of supply chain attacks targeting AI agent ecosystems, where third-party skills can introduce vulnerabilities without the user's knowledge.
Key numbers/companies: VirusTotal · 35,878 AI agent skills · 6,637 malicious (18.5%) · 52.9% security risks · natural-language attacks.
Why it matters: The data demonstrates that AI agent marketplaces are a significant and largely unsecured attack surface. Enterprises deploying agents with third-party skills may be introducing malware into their environments without realizing it.
What's next: Whether agent skill marketplaces implement vetting processes and how enterprises secure their agent supply chains.
Sources: VirusTotal; Mallory.ai
Douzone Bizon Unveils Enterprise AI Agent That Analyzes Workflows and Proposes Tasks Without Instructions
The Korean enterprise software company introduced "ONE AI" and a generative AI no-code development platform at its annual conference.
What happened: Douzone Bizon, a Korean enterprise software company, unveiled an enterprise AI agent that autonomously analyzes corporate workflows and data to propose necessary tasks without explicit instructions. The company introduced "ONE AI," an enterprise AI solution built on its OmniEsol, Amaranth 10, and WEHAGO platforms, along with "GEN AI DEWS," an AI-based no-code/low-code development platform. The announcement signals Douzone's push into the enterprise AI transformation market, positioning its agents as proactive rather than reactive.
Key numbers/companies: Douzone Bizon · ONE AI · GEN AI DEWS · OmniEsol · Amaranth 10 · WEHAGO · enterprise AI agent.
Why it matters: The shift from instruction-following agents to proactive task-proposing agents represents an evolution in enterprise AI. If agents can identify work that needs to be done, they move closer to being genuine digital coworkers.
What's next: How enterprises adopt proactive AI agents and whether the model gains traction outside Korea.
Sources: iKorea Daily
Accenture and Dell Form Business Group to Help Enterprises Scale Private AI
The expanded 20-year partnership focuses on hybrid cloud solutions, agent-based AI, and modernizing enterprise infrastructure.
What happened: Accenture and Dell Technologies announced a major expansion of their 20-plus-year collaboration with the formation of the Accenture Dell Business Group. The group will help enterprises scale private AI and modernize infrastructure, focusing on hybrid cloud solutions and agent-based AI. The partnership reflects growing enterprise demand for private AI deployments that keep data on-premises or in dedicated environments rather than public clouds.
Key numbers/companies: Accenture · Dell Technologies · Accenture Dell Business Group · private AI · hybrid cloud · agent-based AI.
Why it matters: The formation of a dedicated business group signals that private AI is becoming a distinct enterprise market. Organizations with data sovereignty, security, or latency requirements are increasingly looking for AI deployments that do not rely on public cloud infrastructure.
What's next: How quickly enterprises adopt private AI solutions and whether other consultancies form similar partnerships.
Sources: MarketScreener; Accenture; Dell Technologies
Nanya Technology to Build 3D Wafer Packaging and Testing Plant in Pingtung Science Park
The DRAM maker will deploy TSV and wafer stacking technologies to meet AI-driven demand for high-performance, high-bandwidth memory.
What happened: Nanya Technology, a DRAM memory chip and module manufacturer, announced it will invest in establishing a 3D wafer packaging and testing plant in Pingtung Science Park, Taiwan. The advanced packaging facility will deploy TSV (through-silicon via), wafer stacking, and other 3D IC technologies to build wafer-level packaging and testing capabilities. The investment is aimed at meeting the high-performance, high-bandwidth memory demand driven by AI applications. The move follows similar advanced packaging expansions by TSMC, Samsung, and SK hynix as the AI memory boom drives demand for sophisticated packaging solutions.
Key numbers/companies: Nanya Technology · Pingtung Science Park · 3D wafer packaging · TSV · wafer stacking · 3D IC · high-bandwidth memory.
Why it matters: Advanced packaging is emerging as a critical bottleneck in AI chip production, and memory makers are investing to capture a share of the value. Nanya's entry into advanced packaging signals that the memory supply chain is vertically integrating to capture more of the AI value chain.
What's next: Whether Nanya's packaging plant comes online on schedule and how it competes with established packaging providers.
Sources: China Flash Market
Applied Materials and Besi Expand Partnership to Advance Next-Generation AI Packaging
Besi joins Applied Materials' EPIC Center to co-optimize wafer-level hybrid bonding and panel-level integration for AI infrastructure.
What happened: Applied Materials and Besi (BE Semiconductor Industries) announced an expanded strategic partnership to advance next-generation packaging for AI scaling. Besi is joining Applied Materials' EPIC Center, and the two companies will jointly optimize advanced packaging processes and assembly solutions, from wafer-level hybrid bonding to panel-level integration. The collaboration targets next-generation 3D IC scaling, logic and memory integration, and photonics applications supporting AI infrastructure build-out. The partnership aims to help customers adopt these technologies earlier in their production cycles.
Key numbers/companies: Applied Materials · Besi · EPIC Center · wafer-level hybrid bonding · panel-level integration · 3D IC scaling · photonics · AI infrastructure.
Why it matters: Advanced packaging is critical to AI chip performance, and the Applied Materials-Besi partnership aims to accelerate the development of next-generation packaging technologies. The collaboration signals that equipment makers are investing heavily in packaging to capture a larger share of the AI semiconductor value chain.
What's next: Whether the partnership produces commercially viable next-generation packaging solutions and when they reach customers.
Sources: Applied Materials; Digital Today; TechNews.
Why It Matters
The AI revenue reality check has arrived. OpenAI's disclosure that its annualized revenue is $50 billion, not $70 billion, has shaken investor confidence in the AI trade. The selloff in chip stocks reflects a broader concern: if the leading AI company's revenue is growing slower than expected, the case for the massive infrastructure spending becomes harder to justify. The debt financing that Broadcom is seeking for OpenAI custom chips adds another layer of risk — the AI buildout is increasingly leveraged.
TSMC's record results are a counterpoint to AI revenue concerns. The chipmaker's 54.6% year-over-year revenue growth and record quarterly revenue demonstrate that demand for AI chips remains strong at the foundry level. The disconnect between TSMC's results and OpenAI's revenue gap suggests that chip demand is being driven by many customers, not just OpenAI, and that the AI infrastructure buildout has momentum independent of any single company's revenue trajectory.
AI agents are proliferating faster than security frameworks can keep pace. Google's universal Gemini agent, Douzone's proactive enterprise agent, and Manus's $500 million raise all point to a surge in agent deployment. But the ngCERT GhostJacking advisory and VirusTotal's finding that 18.5% of AI agent skills are malicious show that security is lagging. Enterprises deploying agents without robust containment and monitoring are taking on significant risk.
The ShinyHunters arrest is a win, but contractor security remains a systemic weakness. The detention of an alleged ShinyHunters member and the termination of Accenture's FBI contract demonstrate that law enforcement is making progress. However, the breach itself — caused by an unpatched Oracle PeopleSoft system — highlights the persistent vulnerability of government systems that rely on third-party contractors for security.
Advanced packaging is the new battleground in the AI semiconductor supply chain. Nanya's 3D wafer packaging plant and the Applied Materials-Besi partnership both point to the same trend: as AI chips become more complex, the ability to package and integrate them becomes as important as the ability to fabricate them. Companies that control advanced packaging capacity will have significant leverage in the AI supply chain.
The CODEW Take
OpenAI's revenue correction is a warning shot for the entire AI industry. The gap between $70 billion and $50 billion is not a rounding error — it is a 29% miss. OpenAI's explanation of "sloppiness" is not reassuring. If the leading AI company cannot accurately communicate its revenue to investors, what does that say about the industry's financial discipline? The selloff in chip stocks suggests investors are starting to price in the possibility that AI monetization is slower than the narrative implies. The question is not whether AI is transformative — it is whether the returns will justify the hundreds of billions being invested.
Google's universal Gemini agent is the most ambitious enterprise AI product since ChatGPT. Giving an AI agent its own email address, calendar, and directory entry is not a feature — it is a statement that AI agents are digital employees. The agent runs tasks for days after the laptop closes. It operates across Google Workspace, Microsoft 365, and Slack. This is a direct challenge to Microsoft's Copilot ecosystem, and it positions Google as the enterprise AI platform that can work across competitive boundaries. If enterprises adopt persistent AI coworkers, the org chart changes forever.
Manus's $500 million raise proves that China's AI ecosystem can self-fund. Meta's acquisition was blocked by the Chinese government, but Manus did not need Meta. Boyu Capital, IDG Capital, Tencent, Sequoia China, and ZhenFund stepped in with $500 million. The $4 billion valuation is a statement that Chinese investors are willing to back domestic AI champions. The bifurcation of the global AI industry is accelerating, and companies that can operate in both ecosystems will be rare.
GhostJacking is the most important cybersecurity concept of the year. The idea that an AI agent can be manipulated through logs, alerts, and telemetry — without touching its code or APIs — is a fundamental insight. It means that every data source an agent consumes is a potential attack vector. As enterprises deploy agents with access to monitoring systems, error logs, and operational telemetry, they are creating an attack surface that security teams are not equipped to defend. ngCERT's advisory should be required reading for every CISO deploying agentic AI.
Broadcom's $50 billion debt raise for OpenAI chips is a leverage indicator. The AI infrastructure buildout is increasingly funded by debt, not equity. Apollo and Blackstone — two of the largest private credit managers — are in talks to provide financing. This is not venture capital; it is leveraged finance. If AI returns disappoint, the debt servicing obligations will become a significant burden. The AI boom is starting to look more like a capital-intensive industrial buildout than a software revolution. That changes the risk profile entirely.
What to Watch
- OpenAI's $30 billion round: Whether the revenue correction affects investor appetite for OpenAI's ongoing fundraising.
- TSMC Q3 earnings call: Whether the company raises its full-year outlook on sustained AI chip demand.
- ShinyHunters investigation: Whether Khader's cooperation leads to further arrests and how the FBI strengthens contractor security.
- Google Gemini agent GA: Whether general availability arrives by early November and how enterprises respond to persistent AI coworkers.
- Manus capital deployment: How the $500 million is invested and whether the company pursues an IPO.
- Broadcom debt financing: Whether the $50 billion raise closes and how debt-funded AI infrastructure performs.
- GhostJacking mitigation: Whether AI platform vendors implement defenses against telemetry-based agent manipulation.
- AI agent skill vetting: Whether agent marketplaces implement security reviews after VirusTotal's findings.
- Nanya packaging plant: Whether the Pingtung facility comes online on schedule and how it competes.
- Private AI adoption: How quickly enterprises adopt private AI solutions from Accenture and Dell.
Editorial Note
The CODEW Daily News Coverage tracks the most important technology developments of the day, with a focus on AI, enterprise software, cloud computing, semiconductors, cybersecurity, startups, and digital infrastructure. This edition is a pure news aggregation and coverage page — what happened — while the Daily Tech Briefing provides the deeper strategic interpretation.
Coverage is based on company announcements, public disclosures, industry reporting, and other publicly available information. Reported figures and sourced-but-unconfirmed details are noted as such.
Reviewed by Erwin Castro
on
Friday, October 09, 2026
Rating:
No comments: