2025 tech highlights: the calendar roundup

Concept artwork for tech highlights of 2025
generated by Google Gemini Flash 2.5.

Milestones for 2025 included:

January

Deepseek R1 was launched.

"The launch of LLM DeepSeek-V3 and DeepSeek-R1 marked a breakthrough in the Chinese generative AI sector, bringing much confidence to users, companies, and investors. In the past months, many Chinese brands have integrated DeepSeek-supported functions into their products and services," said Ipsos in a 2025 review.

LLM stands for large language model. 

February

Microsoft introduced Majorana 1, the first quantum chip built using topological qubits.

Microsoft’s Majorana 1 marked a major development toward more robust quantum systems, Jason Zander, Executive VP of Microsoft Discovery and Quantum said. 

Majorana 1 has a design that inherently makes fragile qubits more stable and reliable. It is also the only quantum solution engineered to catch and correct errors, Microsoft said. That architecture paves the way for machines with millions of qubits on a single chip, providing the processing power needed for complex scientific and industrial problems.

“Quantum advantage will drive breakthroughs in materials, medicine and more,” Zander said.

“The future of AI and science won’t just be faster, it will be fundamentally redefined.”

July

A survey conducted by the Cyber Security Agency of Singapore found that eight in 10 respondents are confident they can identify deepfakes, but just one in four can actually distinguish between real and fake video content.  

"As AI technology advances, the creation of realistic deepfakes is becoming easier and more widespread. We’re already seeing a rise in fake videos and audio recordings that can be used to deceive individuals.

"It’s critical that users stay ahead of threats to protect not only themselves, but the organisational infrastructures their actions impact. All it takes is one click to let attackers into your business’s four walls, so educate yourself and your employees on safe cyber practices,” said Scott Harrell, Infoblox's President & CEO.

August

MIT Media Lab/Project NANDA released a report that found 95% of investments in generative AI have produced zero results.

"The issue isn't that the models aren't capable, it's that there is a massive gulf between a prototype built in three days and a secure production system. In 2026, we may see high-profile failures where companies give models 'too much rope' without adequate guardrails, resulting in reputational damage or data loss," said Edward Funnekotter, Chief AI Officer at Solace.

"However, this isn't a sign of the technology failing, rather a signal that we need to apply what can be considered traditional, reliable engineering principles to these systems. The bottom-up adoption where workers find tangible, small-scale uses for AI remains highly successful, while massive top-down initiatives struggle. Here’s where a cohesive architecture designed specifically for the complexity of the enterprise comes into its own.

"We need to treat AI projects not as standalone science experiments, but as first-class citizens of the IT landscape, and this is precisely what an agent mesh does. An agent mesh provides a real-time data platform that connects AI agents to the nervous system of the enterprise. Supported by a sturdy event-driven platform, an agent mesh will fundamentally transform how agentic AI systems serve users, respond to business events and integrate with enterprise data, allowing any AI project – from simple single-agent to powerful multi-agent orchestrated solutions – to interact in real-time with enterprise applications and data."

OpenAI launches the GPT-5 AI model.

"On average across 30 countries, 54% say they trust their governments to regulate AI responsibly. But fewer (48%) say they trust companies that use AI to keep their data safe," noted Ipsos in a 2025 review.

September

The Monetary Authority of Singapore's deadline for two-factor authentication for online financial services is 12 September, 2025.  

"Biometric adoption accelerated in high-trust sectors such as government, enterprise and finance, especially in Singapore, where regulatory standards are tightening. With the Monetary Authority of Singapore now mandating two-factor authentication for all online financial services platforms, organisations are bolstering traditional systems with additional verification layers. In turn, hybrid authentication has emerged as the region’s default for secure, frictionless access – designed around how people move, work, and interact," said Prabhuraj Patil, Senior Director, Physical Access Control Solutions, Asean & India Subcontinent at HID.

"Without strong biometric authentication, organisations risk relying on legacy identity systems that were not designed to support today’s expectations around assurance and trust. As Singapore advances its smart city ambitions and builds a more connected digital ecosystem, stronger identity verification becomes essential to safeguard access and reinforce confidence across public- and private-sector services." 

October

The Singapore Police Force announced that as of 1 October 2025, the owners of approximately 1,100 out of the 2,655 lorries required to install speed limiters in line with a 1 January 2026 requirement had not yet done so.

“Singapore’s speed-limiter mandate is a catalyst for digital transformation across the heavy vehicle companies. As fleets prepare for compliance, businesses are realising that the transition isn’t just purely about installing a device - it’s about choosing technology partners that can support them through change,” said Carmen Calisto, Group CSMO of Cartrack Singapore. 

Key provisions detailed in the Singapore Cybersecurity (Amendment) Act - which was passed in Parliament in May 2024 - commenced from 31 October 2025.

"As adversaries continue to outpace traditional defences, fuelled by AI and increasingly sophisticated tactics, organisations will no longer be able to rely on periodic or reactive risk assessments. A recent amendment to Singapore’s cybersecurity regime now requires designated essential service providers that rely on third-party–owned critical information infrastructure (CII) to secure legally-binding commitments on security standards, timely incident notification, audits, risk assessments, and incident reporting—reinforcing accountability across the supply chain," said Shane Buckley, President & CEO, Gigamon.

"At the same time, cyber insurers will tie premiums and coverage to these practices, rewarding organisations that demonstrate continuous monitoring and penalising those that lack complete visibility. Real-time risk assessment powered by deep observability will become both a governance requirement and a financial lever, ensuring organisations detect and respond to threats before they escalate." 

December

Australia initiated a social media ban for those under 16.

"Australia focused on AI regulation and youth exposure to technology. Sixty-seven percent say AI products and services make them nervous, the highest across 30 countries. The upcoming social media ban for under-16s also influenced national discussion on digital wellbeing," noted Gillian Guerin, Regional Director, Marcom APAC from thinktank Ipsos in a 2025 review. 

According to Ipsos, 71% of people on average – and a majority in 30 countries surveyed – say they support banning children under 14 from using social media.

The Singapore Police Force shared in late December that since 1 November 2025, at least 223 cases of phishing scams involving fake buyers on online platforms had been reported, with total losses amounting to at least S$622,000.

"Police statistics show that e-commerce scams are still a huge problem, making up nearly one in five (16.5%) of all scams in Singapore. As AI technology advances, the creation of realistic deepfakes is also becoming easier and more widespread," said Infoblox.

Explore

Highlights were also highlighted monthly in January, February, March, April, May, June, July, August, September, October, November, and December.

Hashtag: #2025highlights