AI Threatens Cybersecurity: WeWorm Malware & Amazon-Qualcomm Partnership

Key Takeaways

• AI enables attackers to automate the entire cyber kill chain, creating continuous attack loops that bypass traditional defense models.
• Researchers used AI to create WeWorm, a zero-click WeChat malware that spreads via calls without user interaction.
• Data exposure is the second most common AI security risk in Australia and New Zealand, following data sent to AI tools.
• AI agent connections to data sources in the ANZ region have increased by 89 percent recently.
• Amazon and Qualcomm partnered on September 8 to build custom silicon and optical connectivity for AI inference at speeds up to 1.6 terabits per second.
• Qualcomm stock rose 2.5% in pre-market trading after the Amazon deal announcement.
• Verily secured a strategic investment from NVIDIA to power its precision health platform using B200 GPUs and the NeMo framework.
• Cloudera and Mistral AI collaborated to allow enterprises to train models on private data within secure hybrid environments.
• Senator Bernie Sanders is organizing a bipartisan briefing to discuss AI risks for lawmakers.
• Popular tools like Anthropic Claude and ChatGPT are widely used, yet many employees still use personal accounts for work tasks.

AI Turns Cyber Attacks Into One Continuous Loop

Artificial intelligence is changing how hackers attack by combining reconnaissance, weaponization, and exploitation into one continuous process. This shift breaks the traditional Cyber Kill Chain model that security teams have used for years to understand attacks. Attackers can now automate these steps at machine speed, making it harder for defenders to keep up. Organizations in the Asia Pacific region face new challenges as their digital ecosystems expand and become more interconnected. Security experts warn that defenders must learn to anticipate threats before they reach critical assets.

CYFIRMA Webinar Explains AI Impact on Cyber Security

CYFIRMA is hosting an exclusive webinar on September 22, 2026, at 3:30 PM AEST to discuss how AI is changing cyber operations in Australia and New Zealand. The event will examine how attackers use AI tools to automate every stage of cyber campaigns in near real time. Experts will share insights on using proactive threat intelligence and continuous exposure management to detect risks early. The session is designed for CISOs, CIOs, security executives, and government leaders who need to strengthen their cyber resilience. Attendees will learn how to move from reactive defense to anticipating and mitigating cyber risks before they cause business impact.

Amazon and Qualcomm Partner on AI Inference Hardware

Amazon and Qualcomm announced a deal on September 8 to develop custom silicon and high-speed optical connectivity for AI inference. This partnership aims to help data centers manage rising power and infrastructure costs while supporting large-scale AI workloads. The companies will work on multiple generations of hardware that can reach speeds up to 1.6 terabits per second. Qualcomm will use its SerDes and optical digital signal processor technologies, including capabilities gained from its recent acquisition of Nuvia. This agreement adds another option for AWS users and data center operators looking for specialized AI infrastructure.

Qualcomm Stock Rises After Amazon AI Deal Announcement

Qualcomm stock rose 2.5% in pre-market trading after Amazon announced a deal to use Qualcomm's custom AI silicon in its cloud infrastructure. Analysts believe this partnership changes the growth outlook for Qualcomm by opening new revenue streams beyond smartphones. The deal provides a boost to the chipmaker, which has faced pressure from ongoing trade tensions with China. Experts note that hyperscalers rarely commit to multiple generations of custom parts, suggesting Amazon plans for long-term hardware architecture. This agreement highlights Qualcomm's expanding role in AI infrastructure alongside its existing data center products.

AI Helps Researchers Create WeChat Worm That Spreads via Calls

Researchers at Calif used artificial intelligence to discover and exploit a memory-corruption flaw in WeChat to create a zero-click worm called WeWorm. This malware can spread from phone to phone without victims answering calls or taking any action. The AI found the vulnerability in July, produced an exploit in two days, and helped build the worm in another week. The tool can take over WeChat accounts on both Android and iOS devices to read messages and make calls. Tencent has patched the flaw, but the incident shows how AI lowers the barrier for creating sophisticated cyberattacks.

Korea benefits from AI boom but faces growing risks

South Korea is currently benefiting greatly from the global artificial intelligence boom as US companies invest heavily in data centers. This demand helps local chipmakers Samsung Electronics and SK hynix sell more memory chips. However, experts warn that this success depends on capital markets continuing to fund these expensive projects. Rising interest rates and competition from Chinese chipmakers could threaten this growth in the future.

Cloudera partners with Mistral AI for secure data tools

Cloudera announced a new partnership with Mistral AI to bring advanced models into secure hybrid data environments. This collaboration helps regulated enterprises keep their sensitive data safe while still using powerful artificial intelligence. Companies can now train models on their own private data using Mistral Forge without sending information to external servers. This approach improves security and reduces costs for organizations that need strict control over their data.

Verily receives NVIDIA investment for health AI platform

Verily has secured a strategic investment from NVIDIA to accelerate its precision health artificial intelligence platform. The funding will support the Verily Pre Platform which uses NVIDIA B200 GPUs and the NeMo framework to process complex medical data. This partnership also helps advance Forecast 1.0, a model that predicts chronic disease risks using genomic and health record data. Additionally, the investment will expand Verily Me and its AI companion Violet to improve patient care and preventive health.

Senator Sanders holds briefing on AI risks for lawmakers

Senator Bernie Sanders is organizing a bipartisan briefing for senators to discuss growing concerns about artificial intelligence. The meeting is scheduled for next week to address the mounting alarm regarding AI technology. This event aims to bring lawmakers together to review the current state and potential dangers of AI development.

Data exposure is second most common AI security risk in ANZ

A new report from Netskope Threat Labs shows that AI systems revealing sensitive data is the second most common security incident in Australia and New Zealand. While sending data to AI tools remains the top risk, the number of AI agents connecting to data sources has increased by 89 percent recently. The region also sees twice the global rate of prompt injection attacks and a high number of workers clicking malicious links found in AI responses. Popular tools like Anthropic Claude and ChatGPT are widely used, but many employees still use personal AI accounts at work.

Sources

• When AI Turns Reconnaissance, Weaponization and Exploitation into One Continuous Machine.
• When AI Turns Reconnaissance, Weaponization and Exploitation into One Continuous Machine - AU & NZ
• Amazon's $60B Qualcomm deal expands AI inference hardware
• QCOM Stock Extends Gains Overnight: Analyst Says Amazon AI Deal 'Changes Calculus For Qualcomm'
• Using AI, Calif Creates Demo WeChat Exploit that Spreads Through Phone Calls
• Korea still riding AI investment supercycle, but risks are building
• Cloudera brings Mistral AI's frontier models into its secure hybrid data environments
• Verily Secures Strategic Investment from NVIDIA to Accelerate Precision Health AI Platform
• Scoop: Sanders convenes senators for AI briefing amid mounting alarm
• AI data exposure ranks second among AI security incidents in Australia and New Zealand