Google's TPUs as a Growing Challenge to Nvidia's AI Chip Dominance

 

Google's custom Tensor Processing Units (TPUs) are increasingly positioning themselves as a formidable rival to Nvidia's longstanding leadership in the AI chip market. While Nvidia's high-performance GPUs remain the cornerstone of most AI models and data centers worldwide, powering everything from training to inference for generative AI systems, Google's TPUs are carving out a significant niche, especially among users of Google Cloud's AI infrastructure.

Designed specifically for machine learning workloads, TPUs excel in tasks like processing large language models (LLMs) central to generative AI. The latest iteration, the TPU v5e, stands out for its cost-effectiveness, delivering superior performance-per-dollar and enhanced scalability within Google's ecosystem. This allows Google to compete not only as a cloud provider but also as a vertically integrated player in AI hardware, bundling its chips with proprietary software such as TensorFlow and Vertex AI for seamless integration.

Market dynamics are shifting as enterprises and AI startups increasingly adopt TPUs to sidestep Nvidia's GPU shortages and escalating prices. This trend is accelerating Google's momentum, particularly in sectors reliant on efficient, scalable AI compute. Although Nvidia still holds the largest global share of AI infrastructure, Google's advancements signal a credible long-term threat. As demand for generative AI surges, the competition in AI chips is evolving from Nvidia's near-monopoly into a more contested landscape, with Google's innovations potentially reshaping the industry. Experts note that this rivalry could drive broader advancements in AI hardware, benefiting developers and enterprises seeking more affordable, optimized solutions. Overall, Google's strategic push underscores a pivotal shift, challenging Nvidia to innovate faster to maintain its edge.

BY:- Nirosha Gupta:)

Microsoft will abandon AI systems that threaten human control, says AI chief Suleyman

 

Microsoft's AI chief Mustafa Suleyman announced that the company will halt development of any AI systems risking operation beyond human control, emphasizing alignment with human values through "humanist superintelligence," where AI serves people rather than acting independently. This follows a revised OpenAI agreement granting Microsoft greater autonomy to pursue superintelligent AI alone or with partners, while prioritizing safety, containment, and oversight. Suleyman advocates for stronger regulations, including audits, transparency, and government involvement, as AI grows more autonomous. He highlights healthcare as a near-term application and warns of potential job and economic disruptions over decades, suggesting measures like universal basic income and wealth redistribution to mitigate impacts.

BY:- Nirosha Gupta

AI promises growth

Transforming AI Awareness into Strategic Action: A CIO’s Guide

AI promises growth, yet many programs struggle to turn goals into results. In manufacturing, AI matters because customers expect shorter lead times and higher quality, plants need automation to unlock capacity, and regulators require stronger traceability. Among AI capabilities, computer vision is production-ready today for inspection, safety, and throughput. The opportunity is real, but when budgets are limited, value must be demonstrated on the line first.

Organizations can invest in a top-tier platform, but workforce adoption depends on enablement, workflow fit, and trust. In a 2024 Gartner survey, 87% of employees needed ongoing training and 60% had not used GenAI for critical work; top adoption blockers included limited GenAI training (30%) and poor-quality GenAI responses (29%).

This post covers why AI matters for enterprise growth in 2025, the real adoption challenges, a clear build-or-buy path, and four strategies to deliver measurable results.

What is an AI strategy?

An AI strategy is the cornerstone of effectively integrating AI into a business. It acts as a roadmap, guiding you on how to leverage AI to achieve specific business goals. Whether it’s extracting deeper insights from data, enhancing operational efficiency, or building better customer experiences, your AI strategy defines the how and why of achieving success with AI.

This plan also encompasses crucial considerations such as:

Tech infrastructure: Ensuring companies have the hardware, software, and resources needed for AI implementation.

Adaptability: Remaining flexible to embrace evolving technologies and industry shifts.

Ethical responsibility: Addressing concerns like bias, transparency, and regulations for responsible AI use.

BY:- Nirosha Gupta;)

Acronis researchers found that most Makop attacks begin with breaking into unsecured Remote Desktop Protocol (RDP) systems.

A recent report by Acronis Research revealed that India accounts for 55% of the victims in recent Makop ransomware operations, making it the most targeted country for this threat. Attackers are exploiting weaker cybersecurity practices and common local antivirus solutions in the region.

Key Findings of the Acronis Report

Primary Entry Point: Most Makop attacks begin by compromising unsecured Remote Desktop Protocol (RDP) systems, often using automated tools to guess weak passwords.

Evolving Delivery Method: The ransomware is now being distributed through Guloader, a type of malware downloader, which helps the attackers better hide the ransomware from security software.

Bypassing Security: The threat actors use a mix of off-the-shelf tools, including network scanners, credential stealers (like Mimikatz), and utilities designed to disable or uninstall security products, including specific Indian antivirus software like Quick Heal.

Targeting SMBs: The Acronis findings suggest a significant risk for India's Small and Medium Businesses (SMBs) and critical sectors, highlighting a need for improved cybersecurity hygiene.

Acronis researchers found that most Makop attacks begin with breaking into unsecured Remote Desktop Protocol (RDP) systems. Attackers use automated tools to guess weak passwords and gain access. After entering, they follow a simple but effective playbook: scanning the network, stealing login credentials, moving deeper into systems, disabling security products, and then finally encrypting data. In many cases, they use known tools like Mimikatz for credential theft and network scanners to map the environment.

Ilia Dafchev, Senior Security Researcher, Acronis, said, “Makop is not a brand-new family of ransomware, but it is changing in ways that are impossible for defenses to ignore. Makop is being deployed using Guloader for the first time, which is a significant change from its typical manual, RDP-based distribution. This modification makes the ransomware more difficult to identify and indicates that even low-complexity attackers are using increasingly complex methods. The regional targeting pattern, 55% of the victims we saw were in India, where attackers even created tools to remove popular local security products, is particularly alarming. These results demonstrate a straightforward reality: businesses that have inadequate security measures or exposed RDP services continue to be highly vulnerable. Improving fundamental cyber hygiene is now essential to staying ahead of fast-evolving threats like these.”

Acronis warns that this combination of old vulnerabilities, weak passwords, and exposed remote access systems continues to put organizations at high risk. The Makop campaign reflects a broader pattern across ransomware groups: attackers often rely on basic security gaps that are easy to fix but widely ignored.

The company recommends that businesses immediately secure all remote access with Multi-Factor Authentication (MFA), apply regular patches, limit public RDP access, and deploy strong endpoint protection capable of detecting loaders like Guloader. Better password practices and regular security audits can also significantly reduce risk.

By - Aaradhay Sharma