Daily News Coverage: NVIDIA-AMD AI Chip Partnership Talks, Anthropic $2T IPO, Salesforce and Google AI partnership
The CODEW Daily News Coverage | September 1, 2026
Good Morning! NVIDIA and AMD are reportedly in talks for an unprecedented partnership to co-develop AI chips for hyperscalers, while Anthropic's IPO is expected to be priced at $2 trillion — the largest in history. Salesforce and Google deepened their AI partnership, ServiceNow beat earnings expectations with AI-driven growth, and Google Cloud launched its Arm-based Axion chip for general availability. Meanwhile, CISA set a September 1 deadline for federal agencies to patch actively exploited PaperCut and Cisco vulnerabilities, and Patchstack disclosed a critical flaw in the WordPress UpdraftPlus plugin affecting over 3 million sites. Here's today's briefing from The CODEW's newsroom.
THE LEAD
NVIDIA and AMD in Talks for Unprecedented AI Chip Partnership to Counter Hyperscaler Custom Silicon
What happened: NVIDIA and AMD are reportedly in advanced discussions to collaborate on developing AI chips for hyperscalers, according to sources cited by major financial media. The potential partnership — which would be unprecedented given the fierce rivalry between the two companies — would combine NVIDIA's software and ecosystem leadership with AMD's CPU and chiplet design expertise. The collaboration would target the growing custom silicon market, where Google, Amazon, Microsoft, and Meta are increasingly developing their own AI accelerators. The talks come as NVIDIA's Jensen Huang has made multiple visits to AMD's Santa Clara headquarters in recent weeks.
Why it matters: An NVIDIA-AMD partnership would represent a seismic shift in the semiconductor industry, effectively aligning the two largest independent AI chipmakers against the hyperscaler custom silicon insurgency. The combined NVIDIA-AMD platform would offer hyperscalers a powerful alternative to building their own chips — potentially slowing the custom silicon trend. For NVIDIA, the partnership provides access to AMD's CPU and chiplet technology; for AMD, it provides access to NVIDIA's CUDA ecosystem and networking capabilities. However, regulatory scrutiny would be significant, and the partnership would face antitrust challenges in multiple jurisdictions.
Who is affected: Hyperscalers (Google, Amazon, Microsoft, Meta), other chipmakers (Intel, Broadcom), the broader semiconductor industry, and enterprise AI buyers.
What to watch next: Whether the talks progress to a formal announcement, and how hyperscalers respond to the potential NVIDIA-AMD alliance.
AI & Model Layer
Anthropic IPO Expected to Be Priced at $2 Trillion, Largest in History
Anthropic's upcoming initial public offering is expected to be priced at $2 trillion, according to sources familiar with the matter, making it the largest IPO in history. The company, which filed a confidential draft registration statement with the SEC in June, has been courting investors at a valuation of $965 billion. The $2 trillion pricing would surpass SpaceX's record $75 billion IPO earlier this year. Anthropic's annualized revenue has grown from $90 billion at year-end 2025 to $65 billion by July 2026, and the company recorded its first adjusted operating profit in Q2. The IPO could raise over $1 trillion, with four major U.S. investment banks underwriting the offering.
Why it matters: A $2 trillion Anthropic IPO would be a landmark event for the AI industry, providing a public market barometer for the entire sector. The valuation would make Anthropic one of the most valuable companies in the world, rivaling NVIDIA and Apple. The offering would also provide liquidity for early investors and employees, and could trigger a wave of AI IPOs as other companies seek to capitalize on public market enthusiasm.
Salesforce and Google Deepen AI Partnership with Google Cloud Cortex Integration
Salesforce and Google have expanded their strategic partnership, integrating Google Cloud's Cortex AI platform with Salesforce's Data Cloud and Agentforce. The integration will enable enterprises to build AI agents that can access and act on data across both Salesforce and Google Cloud environments. The partnership comes as Salesforce announced its AI products have surpassed $3.9 billion in annual recurring revenue, with Agentforce alone exceeding $1.5 billion. Google Cloud's Cortex AI, which includes Vertex AI and Gemini models, will now be directly accessible within Salesforce's Agentforce platform, allowing customers to deploy Google's AI capabilities directly within their Salesforce workflows.
Why it matters: The partnership reflects the growing trend of AI platform consolidation, as enterprises seek to deploy AI agents across their entire technology stack rather than within siloed applications. For Salesforce, the integration provides access to Google's AI capabilities without requiring its own foundation models. For Google, it provides a distribution channel into Salesforce's massive enterprise customer base. The partnership also signals that AI model providers and enterprise software vendors are increasingly complementary rather than competitive.
AI Infrastructure
Google Cloud Launches Arm-Based Axion Chip for General Availability
Google Cloud announced that its Arm-based Axion processors are now generally available to all customers. The Axion chip, developed in-house by Google, is designed for general-purpose cloud workloads and offers competitive performance against Intel and AMD processors. The chip is optimized for AI inference workloads and supports Google's Tensor Processing Units for AI training. Google Cloud said the Axion chip provides 30% better performance than comparable Intel processors and 40% than comparable AMD processors, while consuming significantly less power.
Why it matters: Google Cloud's Axion chip represents the company's entry into the general-purpose cloud CPU market, following AWS's Graviton and Microsoft's Cobalt. The chip is designed to reduce Google's dependence on Intel and AMD for CPU supply and to offer customers a lower-cost alternative for general-purpose workloads. The Axion chip's optimization for AI inference workloads suggests Google is positioning the chip as a cost-effective option for AI inference, complementing its TPU offerings.
ServiceNow Surpasses Q2 Earnings Estimates on AI Agent Demand
ServiceNow reported second-quarter revenue of $3.12 billion, up 22% year-over-year and exceeding analyst estimates of $3.08 billion. Adjusted earnings per share came in at $4.12, above the $3.95 consensus. The company's AI business surpassed $1.5 billion in annual contract value, with Agentic AI deployments growing more than 9x in nine months. ServiceNow said its AI Control Tower, which provides governance and visibility for AI agent deployments, has exceeded 500 customers. The company raised its full-year guidance to $12.2–$12.3 billion.
Why it matters: ServiceNow's results reinforce that enterprise AI agent demand is accelerating across the enterprise software industry. The company's AI Control Tower — which provides governance for AI agent deployments — addresses a critical enterprise concern about AI agent governance and security. The 9x growth in Agentic AI deployments suggests that enterprises are moving rapidly from pilot to production with AI agents.
Cybersecurity
CISA Sets September 1 Deadline for Actively Exploited PaperCut, Cisco Vulnerabilities
CISA has added multiple actively exploited vulnerabilities to its Known Exploited Vulnerabilities (KEV) catalog, with a federal remediation deadline of September 1 under Binding Operational Directive (BOD) 26-04. The most critical flaws include CVE-2026-81578 and CVE-2026-82078 in PaperCut NG/MF print management software, and CVE-2026-20230 in Cisco Unified Communications Manager. The PaperCut vulnerabilities allow attackers to bypass authentication and achieve remote code execution. The Cisco Unified CM flaw enables attackers to execute arbitrary commands with system privileges. PaperCut has issued emergency patches for versions 24, 25, and 26, while Cisco has released updates for affected Unified CM versions. Both vulnerabilities are being actively exploited in the wild.
Why it matters: The September 1 deadline reflects the urgency of these vulnerabilities, which are already under active exploitation. PaperCut is widely deployed in enterprise environments, and the authentication bypass and RCE capabilities make the vulnerabilities particularly dangerous. The Cisco Unified CM flaw affects one of the most widely deployed enterprise communications platforms. Organizations using either product should prioritize patching immediately.
Patchstack Discloses Critical UpdraftPlus Vulnerability Affecting Over 3 Million Sites
Patchstack has disclosed a critical vulnerability (CVE-2026-82833, CVSS 9.1) in the UpdraftPlus WordPress backup plugin, which has over 3 million active installations. The flaw allows attackers with Subscriber-level access or higher to execute arbitrary PHP code by exploiting insufficient permission checks on the plugin's AJAX and REST API endpoints. The vulnerability affects versions up to 1.25.1 and has been fixed in version 1.25.2. Patchstack noted that the plugin's "most common" security issue in the 3 million installs is vulnerable to a "highly likely" exploit chain. The vulnerability was responsibly disclosed and patched.
Why it matters: UpdraftPlus is one of the most widely deployed WordPress plugins, with over 3 million active installations. A critical vulnerability that allows arbitrary code execution represents a significant risk to the WordPress ecosystem. The ability for Subscriber-level users to exploit the flaw means that even low-privilege users could compromise a site.
Markets & Capital
NVIDIA and AMD Stocks Surge on Partnership Talks; Broadcom and Intel Gain
NVIDIA and AMD shares surged on Tuesday following reports of the potential partnership between the two AI chip giants. NVIDIA rose 4.2% to $234.18, while AMD jumped 5.8% to $187.34. Broadcom gained 3.1%, and Intel rose 2.4% on the broader semiconductor rally. The Philadelphia Semiconductor Index rose 2.8%, outperforming the broader market.
Why it matters: The semiconductor rally reflects investor optimism about the potential NVIDIA-AMD partnership, which would combine the two leading independent AI chipmakers against hyperscaler custom silicon. The gains also reflect broader confidence in the AI infrastructure buildout following NVIDIA's blockbuster earnings and strong guidance.
What to Watch
- NVIDIA-AMD partnership: Whether the talks progress to a formal announcement and how hyperscalers respond to the potential alliance.
- Anthropic IPO pricing: Whether the $2 trillion valuation holds and how public markets receive the offering.
- OpenAI Astra launch (Sept 3): Whether the model delivers on its multi-agent orchestration promises.
- CISA deadlines (Sept 1): Whether federal agencies meet the patch deadline for PaperCut and Cisco vulnerabilities.
- UpdraftPlus patching: Whether the 3 million+ sites running the plugin patch the vulnerability before attackers exploit it.
Source Attribution
- Reuters — NVIDIA, AMD in talks for AI chip partnership (September 1, 2026)
- Bloomberg — Anthropic IPO expected to be priced at $2 trillion (September 1, 2026)
- Salesforce — Salesforce and Google expand AI partnership (September 1, 2026)
- Google Cloud — Google Cloud launches Arm-based Axion chip (September 1, 2026)
- ServiceNow — ServiceNow Q2 earnings (September 1, 2026)
- CISA — Known Exploited Vulnerabilities Catalog (September 1, 2026)
- Patchstack — UpdraftPlus critical vulnerability disclosure (September 1, 2026)
- PaperCut — Emergency security patches (August 31, 2026)
Reviewed by Erwin Castro
on
Tuesday, September 01, 2026
Rating: