Samsung officially unveiled the much anticipated Galaxy S26 and S26 Plus on February 26, 2026 with some fresh upgrades over the Galaxy S25 and S25 Plus.

While it’s fair to say the Galaxy S26 Ultra has stolen many of the headlines once again this year, there’s still much to love with the rest of the Galaxy S26 range. The base model Galaxy S26, for example, sports a slightly larger 6.3-inch screen over its predecessor (6.2 inches), and its storage options now start at 256GB, double from the 128GB that the S25 started with. The battery also gets a boost to 4,300mAh, up from 4,000mAh.

The larger-screened Galaxy S26 Plus, meanwhile, retains the 6.7-inch display and 4,900mAh battery from its predecessor, and gets Samsung’s new Exynos 2600 chipset, with the Snapdragon 8 Elite Gen 5 chip reserved for the top-tier S26 Ultra. While the battery capacity is the same, the S26 Plus can be charged wirelessly at 20W, compared to the S25 Plus’ slower 15W wireless charging.

If you’re looking to buy outright, the range is available at retailers like Amazon, JB Hi-Fi and The Good Guys, with the first two offering trade-in options. Buying direct from Samsung will also let you pay in instalments of up to 24 months, and trade-ins are also available (including non-Samsung products). When paired with one of our best SIM-only plans, buying outright will be the best bang-for-buck option.

These are flagship phones, so the base model Galaxy S26 and S26 Plus won’t fit in our best cheap phones list. It also doesn’t help that both handsets are more expensive than the S25 lineup, with the base model S26 starting from AU$1,549 (up from the S25’s AU$1,399) and the S26 Plus from AU$1,849 (vs the S25 Plus’s starting price of AU$1,699). This would make paying in monthly instalments an attractive option for some.

While the retailers have finished their pre-order specials, Australia’s big three telcos still have some active deals of up to AU$500 off the handset price for the Samsung Galaxy S26 and S26 Plus (as well as the Galaxy S26 Ultra), with some also coinciding with existing promotions.

With so many options available to score a brand new upgrade, fInding the best plan for these new handsets may not be the most straightforward process, so we’ve done the hard work for you. Take a look at our picks for the best phone plans for the Samsung Galaxy S26 and S26 Plus below:

The best Samsung Galaxy S26 plan

The best Samsung Galaxy S26 Plus plan

The best Samsung Galaxy S26 plan with a smartwatch

Compare Samsung Galaxy S26 and S26 Plus plans

Outright Samsung Galaxy S26 and S26 Plus deals

• Samsung: pay in instalments of up to 24 months through Samsung financing; also save up to AU$865 when you trade in your old device
• JB Hi-Fi: trade in your old tech for a JB Hi-Fi gift card to be used on a Galaxy S26 series handset
• The Good Guys: Galaxy S26 and S26 Plus available in 256GB and 512GB
• Amazon: Same day delivery with the world’s biggest retailer

Samsung Galaxy S26 and S26 Plus: Key Information

Unfortunately, both the headline-grabbing Privacy Display and the gimbal-like horizontal lock video mode are only exclusive to the Galaxy S26 Ultra, so if you’re specifically looking for those features, you can check out the best Galaxy S26 Ultra plans.

Thanks to all our beloved AI companions, RAM prices have more than doubled in the last 6 months. As you may have guessed, the smartphone industry has been having some tough times with budget devices. To help solve this problem, HMD (remember the people who resurrected Nokia?) has announced a strategic partnership with Flipkart to bring its upcoming 2026 smartphone lineup to Indian consumers.

HMD Expanding Its Presence in India

HMD’s upcoming smartphone lineup will target multiple price segments between ₹10,000 and ₹20,000, catering to users looking for reliable devices with modern features at affordable prices. The first smartphone under the partnership is expected to launch in the coming months, followed by a phased rollout of several additional devices over the next two to four months. Specifics about the phones are not yet known, but we are due to get our hands on them. So, stay tuned.

The partnership will kick off with HMD’s first smartphone launch of 2026 in India, which will debut on Flipkart before reaching other online and retail channels. The goal is to leverage Flipkart’s extensive reach, logistics network, and consumer insights to make its smartphones more accessible nationwide.

Ravi Kunwar, CEO and VP of HMD India and APAC, said, “We are excited to collaborate with Flipkart as one of our key e-commerce partners as we gear up to launch the first HMD smartphone of 2026 in India. Flipkart’s extensive reach and strong consumer connect will play an important role in bringing our latest innovation to customers across the country.”

Commenting on the same, Ajay Veer Yadav, Senior Vice President at Flipkart, said, “Our strategic collaboration with HMD brings their upcoming smartphone portfolio to millions of consumers across the country. With our expansive distribution network and flexible affordability offerings, we are well-positioned to make cutting-edge devices more accessible and inclusive.”

New Research from WiCyS and FourOne Insights explores how skills-based cyber practices can positively impact employees and their organisations.

Women in CyberSecurity (WiCyS), the nonprofit organisation dedicated to the recruitment, retention and advancement of women in cybersecurity and FourOne Insights, a research and advisory firm, have released a new report. 

The ‘ROI of Resilience: How Cybersecurity Talent Management Best Practices Improve the Bottom Line’ study explores the financial impact of skills-based talent strategies in cybersecurity. To gather information WiCyS and FourOne Insights leveraged data from an original survey, job posting data from labour market analytics provider Lightcast, and professional social profile data, also from Lightcast.

What was discovered is that skills-based, talent-friendly practices often generate the highest returns for an organisation and its workforce. Data from the report indicated that mentorship opportunities and skills-based development increased retention by up to 18pc, with skills-based promotions improving the representation of women in cyber leadership by upwards of 10pc and in some cases, as much as 20pc. 

The report said: “These practices benefit the entire workforce and are especially valuable for women. Panels for promotion decisions, internal skills profiles and formal mentorship programmes all correlate with significantly higher representation of women in cybersecurity management roles.

“Organisations using these practices see 10pc to 20pc higher representation of women in cybersecurity leadership than firms that do not. Skills-based promotion criteria and linking incentives to demonstrated skill growth further strengthen both equity and financial performance.”

Addressing challenges

The research indicates that an awareness of skills-based and talent-conscious practices can be mutually beneficial for those operating within an organisation. In fact, they have the potential to deliver more than $125,000 in savings per employee, according to the report. But despite the merits of this system, the data also indicated that the adoption of these practices is uneven. 

The report said: “Despite the mutual benefit to employers and employees, many high-impact practices are among the least utilised. None of the highest-value practices are leveraged by more than 55pc of firms. 

“When companies do implement these practices, they often base them on unreliable, subjective data. This threatens worse talent outcomes for organisations, while limiting career development opportunities for individuals.”

But, third-party partnerships could potentially expand capacity and ease the adoption of talent and skills-focused practices. Almost 80pc of contributing respondents explained that they find access to supportive, career-based organisations such as WiCyS to be valuable, with many of the opinion that they create stronger professional networks than those created by an employer. 

According to the report, the firms that offer this kind of access tend to fill roles 16pc faster, retain workers longer and avoid significant productivity losses, when compared to those who don’t. “These partnerships provide capabilities such as peer learning, industry context, and trusted communities that are difficult for employers to build internally.”

Future ready

WiCyS’ and FourOne Insight’s research suggests that the companies attracting the strongest talent and meeting business objectives have a common approach, in that they ground their strategies in skills data, leadership actions and clear employee development opportunities.

“High-ROI practices, such as transparent promotion processes, executive sponsorship, access to upskilling and mentorship and engagement with trusted third-party partners, can consistently reduce hiring friction and support retention,” said the report. “Over time, they open advancement pathways that have historically been narrow, especially for women.”

The framework to ensure effective practices should include the assessment of workforce pain points, the planning of targeted interventions, execution with stakeholder buy-in and continuously evaluating outcomes. This, the report states, will create a “durable, self-correcting system that strengthens workforce resilience and ensures that opportunities are genuinely accessible to all talent, not simply expanded in name only”.

As for further research, the report suggests that those collecting data should explore how these practices influence broader indicators of organisational performance, including profitability and long-term resilience. 

“What remains clear is that in a tightening labour market, workforce resilience is a strategic imperative. Skills-based, talent-friendly practices, reinforced by strong third-party partnerships, offer a path to building that resilience at scale.”

Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.

An SQL injection vulnerability in Ally, a WordPress plugin from Elementor for web accessibility and usability with more than 400,000 installations, could be exploited to steal sensitive data without authentication.

The security issue, tracked as CVE-2026-2313, received a high severity score. It was discovered by Drew Webber (mcdruid), an offensive security engineer at Acquia, a software-as-a-service company that provides an enterprise-level Digital Experience Platform (DXP).

SQL injection flaws have been around for more than 25 years and continue to be a threat today, despite being well understood and technically easy to fix and avoid. This type of security issue occurs when user input is directly inserted into an SQL database query without proper sanitization or parameterization.

This allows an attacker to inject SQL commands that alter the query’s behavior to read, modify, or delete information in the database.

CVE-2026-2313 affects all Ally versions up to 4.0.3 and lets an unauthenticated attacker to inject SQL queries via the URL path due to improper handling of a user-supplied URL parameter in a critical function.

“This is due to insufficient escaping on the user-supplied URL parameter in the `get_global_remediations()` method, where it is directly concatenated into an SQL JOIN clause without proper sanitization for SQL context,” reads a technical analysis from WordFence.

“While `esc_url_raw()` is applied for URL safety, it does not prevent SQL metacharacters (single quotes, parentheses) from being injected.

“This makes it possible for unauthenticated attackers to append additional SQL queries into already existing queries that can be used to extract sensitive information from the database via time-based blind SQL injection techniques,” the researchers explain.

Wordfence notes that exploiting the vulnerability is possible only if the plugin is connected to an Elementor account and its Remediation module is active.

The security firm validated the flaw and disclosed it to the vendor on February 13. Elementor fixed the flaw in version 4.1.0 (latest), released on February 23, and an $800 bug bounty was awarded to the researcher.

Data from WordPress.org shows that only about 36% of websites using the Ally plugin have upgraded to version 4.1.0, leaving more than 250,000 sites vulnerable to CVE-2026-2313.

In addition to upgrading Ally to version 4.1.0, site owners/administrators are also recommended to install the latest security update for WordPress, released yesterday.

WordPress 6.9.2, addresses 10 vulnerabilities, including cross-site request (XSS), authorization bypass, and server-side request forgery (SSRF) flaws. The new version of the platform is recommended to be installed “immediately.”

Malware is getting smarter. The Red Report 2026 reveals how new threats use math to detect sandboxes and hide in plain sight.

Download our analysis of 1.1 million malicious samples to uncover the top 10 techniques and see if your security stack is blinded.

Download The Report

Researchers say they have uncovered a takedown-resistant botnet of 14,000 routers and other network devices—primarily made by Asus—that have been conscripted into a proxy network that anonymously carries traffic used for cybercrime.

The malware—dubbed KadNap—takes hold by exploiting vulnerabilities that have gone unpatched by their owners, Chris Formosa, a researcher at security firm Lumen’s Black Lotus Labs, told Ars. The high concentration of Asus routers is likely due to botnet operators acquiring a reliable exploit for vulnerabilities affecting those models. He said it’s unlikely that the attackers are using any zero-days in the operation.

A botnet that stands out among others

The number of infected routers averages about 14,000 per day, up from 10,000 last August, when Black Lotus discovered the botnet. Compromised devices are overwhelmingly located in the US, with smaller populations in Taiwan, Hong Kong, and Russia. One of the most salient features of KadNap is a sophisticated peer-to-peer design based on Kademlia, a network structure that uses distributed hash tables to conceal the IP addresses of command-and-control servers. The design makes the botnet resistant to detection and takedowns through traditional methods.

“The KadNap botnet stands out among others that support anonymous proxies in its use of a peer-to-peer network for decentralized control,” Formosa and fellow Black Lotus researcher Steve Rudd wrote Wednesday. “Their intention is clear: avoid detection and make it difficult for defenders to protect against.”

Distributed hash tables have long been used to create hardened peer-to-peer networks, most notably BitTorrent and the Inter-Planetary File System. Rather than having one or more centralized servers that directly control nodes and provide them with the IP addresses of other nodes, DHTs allow any node to poll other nodes for the device or server it’s looking for. The decentralized structure and the substitution of IP addresses with hashes give the network resilience against takedowns or denial of service attacks.

If you’ve ever scrolled TikTok, caught a snippet of a tune, and thought, “I wish I could play this song all the way through,” this is for you. TikTok and Apple Music announced on Wednesday that they have partnered on two new features, Play Full Song and Listening Party. The goal is to offer listeners a seamless music listening experience without ever leaving the social media app.

Apple Music subscribers who discover a song on their TikTok For You Page or on the Sound Detail Page will be able to click Play Full Song to open the Apple Music player and listen to the track in its entirety. From there, subscribers to the music streaming service will be able to save the song as a favorite, add it to a playlist on Apple Music and listen to a customized stream of recommended songs.

When a full-length song is played, the stream will pay artists through Apple Music. 

“Tapping into the music you love should feel effortless,” Ole Obermann, co-head of Apple Music, said in a statement. “With Play Full Song, Apple Music subscribers can move easily from discovering a track on TikTok to listening to it in full instantly, without breaking the flow. This integration not only makes it easier for fans to discover, listen to, and engage with the artists they love, but also creates a powerful new pathway for artists — turning moments of discovery into deeper connection and sustained engagement in one simple, seamless experience.”

Listening Party sounds somewhat like Spotify‘s feature of the same name. Fans join a shared, real-time session where they listen to the same tracks together and interact live, with the songs streamed through Apple Music inside TikTok. Musicians can also join and chat with their fans.

“TikTok is where music discovery and culture move at the speed of the community,” Tracy Gardner, global head of music business development at TikTok, said in a statement. “Thanks to Apple Music, Play Full Song gives fans a seamless way to go from discovery to full-length listening, and Listening Party provides a shared place to experience music together in real time. It’s all about bringing artists and fans closer, and turning shared moments into lasting connections.”

Play Full Song and Listening Party will launch globally on TikTok over the next few weeks.

Meta buying Moltbook, the developer of a social media platform designed for AI agents to talk to each other, sounds a little like a joke someone might make about how there are too many bots on Facebook and other Meta platforms. But it looks like Meta hopes to use Moltbook to fill the internet with even more digital voices.

Meta has spent two decades building platforms that connect billions of people. Facebook, Instagram, and Threads all promise some version of the same basic idea: a digital place where humans share thoughts, photos, jokes, and complaints about social media.

Multi-agent systems, designed to handle long-horizon tasks like software engineering or cybersecurity triaging, can generate up to 15 times the token volume of standard chats — threatening their cost-effectiveness in handling enterprise tasks.

But today, Nvidia sought to help solve this problem with the release of Nemotron 3 Super, a 120-billion-parameter hybrid model, with weights posted on Hugging Face.

By merging disparate architectural philosophies—state-space models, transformers, and a novel “Latent” mixture-of-experts design—Nvidia is attempting to provide the specialized depth required for agentic workflows without the bloat typical of dense reasoning models, and all available for commercial usage under mostly open weights.

Triple hybrid architecture

At the core of Nemotron 3 Super is a sophisticated architectural triad that balances memory efficiency with precision reasoning. The model utilizes a Hybrid Mamba-Transformer backbone, which interleaves Mamba-2 layers with strategic Transformer attention layers.

To understand the implications for enterprise production, consider the “needle in a haystack” problem. Mamba-2 layers act like a “fast-travel” highway system, handling the vast majority of sequence processing with linear-time complexity. This allows the model to maintain a massive 1-million-token context window without the memory footprint of the KV cache exploding. However, pure state-space models often struggle with associative recall. 

To fix this, Nvidia strategically inserts Transformer attention layers as “global anchors,” ensuring the model can precisely retrieve specific facts buried deep within a codebase or a stack of financial reports.

Beyond the backbone, the model introduces Latent Mixture-of-Experts (LatentMoE). Traditional Mixture-of-Experts (MoE) designs route tokens to experts in their full hidden dimension, which creates a computational bottleneck as models scale. LatentMoE solves this by projecting tokens into a compressed space before routing them to specialists. 

This “expert compression” allows the model to consult four times as many specialists for the exact same computational cost. This granularity is vital for agents that must switch between Python syntax, SQL logic, and conversational reasoning within a single turn.

Further accelerating the model is Multi-Token Prediction (MTP). While standard models predict a single next token, MTP predicts several future tokens simultaneously. This serves as a “built-in draft model,” enabling native speculative decoding that can deliver up to 3x wall-clock speedups for structured generation tasks like code or tool calls.

The Blackwell advantage

For enterprises, the most significant technical leap in Nemotron 3 Super is its optimization for the Nvidia Blackwell GPU platform. By pre-training natively in NVFP4 (4-bit floating point), Nvidia has achieved a breakthrough in production efficiency.

On Blackwell, the model delivers 4x faster inference than 8-bit models running on the previous Hopper architecture, with no loss in accuracy.

In practical performance, Nemotron 3 Super is a specialized tool for agentic reasoning.

It currently holds the No. 1 position on the DeepResearch Bench, a benchmark measuring an AI’s ability to conduct thorough, multi-step research across large document sets.

It also demonstrates significant throughput advantages, achieving up to 2.2x higher throughput than gpt-oss-120B and 7.5x higher than Qwen3.5-122B in high-volume settings.

Custom ‘open’ license — commercial usage but with important caveats 

The release of Nemotron 3 Super under the Nvidia Open Model License Agreement (updated October 2025) provides a permissive framework for enterprise adoption, though it carries distinct “safeguard” clauses that differentiate it from pure open-source licenses like MIT or Apache 2.0.

Key Provisions for Enterprise Users:

• Commercial Usability: The license explicitly states that models are “commercially usable” and grants a perpetual, worldwide, royalty-free license to sell and distribute products built on the model.

• Ownership of Output: Nvidia makes no claim to the outputs generated by the model; the responsibility for those outputs—and the ownership of them—rests entirely with the user.

• Derivative Works: Enterprises are free to create and own “Derivative Models” (fine-tuned versions), provided they include the required attribution notice: “Licensed by Nvidia Corporation under the Nvidia Open Model License.”

The “Red Lines”:

The license includes two critical termination triggers that production teams must monitor:

1. Safety Guardrails: The license automatically terminates if a user bypasses or circumvents the model’s “Guardrails” (technical limitations or safety hyperparameters) without implementing a “substantially similar” replacement appropriate for the use case.

2. Litigation Trigger: If a user institutes copyright or patent litigation against Nvidia alleging that the model infringes on their IP, their license to use the model terminates immediately.

This structure allows Nvidia to foster a commercial ecosystem while protecting itself from “IP trolling” and ensuring that the model isn’t stripped of its safety features for malicious use.

‘The team really cooked’

The release has generated significant buzz within the developer community. Chris Alexiuk, a Senior Product Research Enginner at Nvidia, heralded the launch on X under his handle @llm_wizard as a “SUPER DAY,” emphasizing the model’s speed and transparency. “Model is: FAST. Model is: SMART. Model is: THE MOST OPEN MODEL WE’VE DONE YET,” Chris posted, highlighting the release of not just weights, but 10 trillion tokens of training data and recipes.

The industry adoption reflects this enthusiasm:

• Cloud and Hardware: The model is being deployed as an Nvidia NIM microservice, allowing it to run on-premises via the Dell AI Factory or HPE, as well as across Google Cloud, Oracle, and shortly, AWS and Azure.

• Production Agents: Companies like CodeRabbit (software development) and Greptile are integrating the model to handle large-scale codebase analysis, while industrial leaders like Siemens and Palantir are deploying it to automate complex workflows in manufacturing and cybersecurity.

As Kari Briski, Nvidia VP of AI Software, noted: “As companies move beyond chatbots and into multi-agent applications, they encounter… context explosion.”

Nemotron 3 Super is Nvidia’s answer to that explosion—a model that provides the “brainpower” of a 120B parameter system with the operational efficiency of a much smaller specialist. For the enterprise, the message is clear: the “thinking tax” is finally coming down.

In 2024, Microsoft caused a lot of head-scratching and general bemusement with the launch of its “This is an Xbox” marketing campaign. Now, though, it appears the quandary over what is and isn’t an Xbox has been resolved. Game Developer noticed that the original blog post on Xbox Wire that kicked off the whole affair has been removed. It seems Xbox will be going a new direction with its future promotions.

Maybe since the new Project Helix hardware it has in the works is more definite attempt to blur console and PC gaming, “This is an Xbox” might have been truly confusing as a tagline. Maybe with the recent changing of the guard at the company, the top brass decided that it was the right time to start fresh with a less meme-able marketing plan. Whatever the reason, we have enjoyed this opportunity to learn about the existential philosophy behind being an Xbox. And fortunately, although the blog post may be gone, the video trailer still exists whenever we need to remind ourselves of the many things that can be Xbox-ified.