Stay Ahead, Stay ONMINE

Qwen-Image is a powerful, open source new AI image generator with support for embedded text in English & Chinese

After seizing the summer with a blitz of powerful, freely available new open source language and coding focused AI models that matched or in some cases bested closed-source/proprietary U.S. rivals, Alibaba’s crack “Qwen Team” of AI researchers is back again today with the release of a highly ranked new AI image generator model — also open source.Qwen-Image stands out in a crowded field of generative image models due to its emphasis on rendering text accurately within visuals — an area where many rivals still struggle. Supporting both alphabetic and logographic scripts, the model is particularly adept at managing complex typography, multi-line layouts, paragraph-level semantics, and bilingual content (e.g., English-Chinese).In practice, this allows users to generate content like movie posters, presentation slides, storefront scenes, handwritten poetry, and stylized infographics — with crisp text that aligns with their prompts.

After seizing the summer with a blitz of powerful, freely available new open source language and coding focused AI models that matched or in some cases bested closed-source/proprietary U.S. rivals, Alibaba’s crack “Qwen Team” of AI researchers is back again today with the release of a highly ranked new AI image generator model — also open source.

Qwen-Image stands out in a crowded field of generative image models due to its emphasis on rendering text accurately within visuals — an area where many rivals still struggle.

Supporting both alphabetic and logographic scripts, the model is particularly adept at managing complex typography, multi-line layouts, paragraph-level semantics, and bilingual content (e.g., English-Chinese).

In practice, this allows users to generate content like movie posters, presentation slides, storefront scenes, handwritten poetry, and stylized infographics — with crisp text that aligns with their prompts.


The AI Impact Series Returns to San Francisco – August 5

The next phase of AI is here – are you ready? Join leaders from Block, GSK, and SAP for an exclusive look at how autonomous agents are reshaping enterprise workflows – from real-time decision-making to end-to-end automation.

Secure your spot now – space is limited: https://bit.ly/3GuuPLF


Qwen-Image’s output examples include a wide variety of real-world use cases:

  • Marketing & Branding: Bilingual posters with brand logos, stylistic calligraphy, and consistent design motifs
  • Presentation Design: Layout-aware slide decks with title hierarchies and theme-appropriate visuals
  • Education: Generation of classroom materials featuring diagrams and precisely rendered instructional text
  • Retail & E-commerce: Storefront scenes where product labels, signage, and environmental context must all be readable
  • Creative Content: Handwritten poetry, scene narratives, anime-style illustration with embedded story text

Users can interact with the model on the Qwen Chat website by selecting “Image Generation” mode from the buttons below the prompt entry field.

However, my brief initial tests revealed the text and prompt adherence was not noticeably better than Midjourney, the popular proprietary AI image generator from the U.S. company of the same name. My session through Qwen chat produced multiple errors in prompt comprehension and text fidelity, much to my disappointment, even after repeated attempts and prompt rewording:

Yet Midjourney only offers a limited number of free generations and requires subscriptions for any more, compared to Qwen Image, which, thanks to its open source licensing and weights posted on Hugging Face, can be adopted by any enterprise or third-party provider free-of-charge.

Licensing and availability

Qwen-Image is distributed under the Apache 2.0 license, allowing commercial and non-commercial use, redistribution, and modification — though attribution and inclusion of the license text are required for derivative works.

This may make it attractive to enterprises looking for an open source image generation tool to use for making internal or external-facing collateral like flyers, ads, notices, newsletters, and other digital communications.

But the fact that the model’s training data remains a tightly guarded secret — like with most other leading AI image generators — may sour some enterprises on the idea of using it.

Qwen, unlike Adobe Firefly or OpenAI’s GPT-4o native image generation, for example, does not offer indemnification for commercial uses of its product (i.e., if a user gets sued for copyright infringement, Adobe and OpenAI will help support them in court).

The model and associated assets — including demo notebooks, evaluation tools, and fine-tuning scripts — are available through multiple repositories:

In addition, a live evaluation portal called AI Arena allows users to compare image generations in pairwise rounds, contributing to a public Elo-style leaderboard.

Training and development

Behind Qwen-Image’s performance is an extensive training process grounded in progressive learning, multi-modal task alignment, and aggressive data curation, according to the technical paper the research team released today.

The training corpus includes billions of image-text pairs sourced from four domains: natural imagery, human portraits, artistic and design content (such as posters and UI layouts), and synthetic text-focused data. The Qwen Team did not specify the size of the training data corpus, aside from “billions of image-text pairs.” They did provide a breakdown of the rough percentage of each category of content it included:

  • Nature: ~55%
  • Design (UI, posters, art): ~27%
  • People (portraits, human activity): ~13%
  • Synthetic text rendering data: ~5%

Notably, Qwen emphasizes that all synthetic data was generated in-house, and no images created by other AI models were used. Despite the detailed curation and filtering stages described, the documentation does not clarify whether any of the data was licensed or drawn from public or proprietary datasets.

Unlike many generative models that exclude synthetic text due to noise risks, Qwen-Image uses tightly controlled synthetic rendering pipelines to improve character coverage — especially for low-frequency characters in Chinese.

A curriculum-style strategy is employed: the model starts with simple captioned images and non-text content, then advances to layout-sensitive text scenarios, mixed-language rendering, and dense paragraphs. This gradual exposure is shown to help the model generalize across scripts and formatting types.

Qwen-Image integrates three key modules:

  • Qwen2.5-VL, the multimodal language model, extracts contextual meaning and guides generation through system prompts.
  • VAE Encoder/Decoder, trained on high-resolution documents and real-world layouts, handles detailed visual representations, especially small or dense text.
  • MMDiT, the diffusion model backbone, coordinates joint learning across image and text modalities. A novel MSRoPE (Multimodal Scalable Rotary Positional Encoding) system improves spatial alignment between tokens.

Together, these components allow Qwen-Image to operate effectively in tasks that involve image understanding, generation, and precise editing.

Performance benchmarks

Qwen-Image was evaluated against several public benchmarks:

  • GenEval and DPG for prompt-following and object attribute consistency
  • OneIG-Bench and TIIF for compositional reasoning and layout fidelity
  • CVTG-2K, ChineseWord, and LongText-Bench for text rendering, especially in multilingual contexts

In nearly every case, Qwen-Image either matches or surpasses existing closed-source models like GPT Image 1 [High], Seedream 3.0, and FLUX.1 Kontext [Pro]. Notably, its performance on Chinese text rendering was significantly better than all compared systems.

On the public AI Arena leaderboard — based on 10,000+ human pairwise comparisons — Qwen-Image ranks third overall and is the top open-source model.

Implications for enterprise technical decision-makers

For enterprise AI teams managing complex multimodal workflows, Qwen-Image introduces several functional advantages that align with the operational needs of different roles.

Those managing the lifecycle of vision-language models — from training to deployment — will find value in Qwen-Image’s consistent output quality and its integration-ready components. The open-source nature reduces licensing costs, while the modular architecture (Qwen2.5-VL + VAE + MMDiT) facilitates adaptation to custom datasets or fine-tuning for domain-specific outputs.

The curriculum-style training data and clear benchmark results help teams evaluate fitness for purpose. Whether deploying marketing visuals, document renderings, or e-commerce product graphics, Qwen-Image allows rapid experimentation without proprietary constraints.

Engineers tasked with building AI pipelines or deploying models across distributed systems will appreciate the detailed infrastructure documentation. The model has been trained using a Producer-Consumer architecture, supports scalable multi-resolution processing (256p to 1328p), and is built to run with Megatron-LM and tensor parallelism. This makes Qwen-Image a candidate for deployment in hybrid cloud environments where reliability and throughput matter.

Moreover, support for image-to-image editing workflows (TI2I) and task-specific prompts enables its use in real-time or interactive applications.

Professionals focused on data ingestion, validation, and transformation can use Qwen-Image as a tool to generate synthetic datasets for training or augmenting computer vision models. Its ability to generate high-resolution images with embedded, multilingual annotations can improve performance in downstream OCR, object detection, or layout parsing tasks.

Since Qwen-Image was also trained to avoid artifacts like QR codes, distorted text, and watermarks, it offers higher-quality synthetic input than many public models — helping enterprise teams preserve training set integrity.

Looking for feedback and opportunities to collaborate

The Qwen Team emphasizes openness and community collaboration in the model’s release.

Developers are encouraged to test and fine-tune Qwen-Image, offer pull requests, and participate in the evaluation leaderboard. Feedback on text rendering, editing fidelity, and multilingual use cases will shape future iterations.

With a stated goal to “lower the technical barriers to visual content creation,” the team hopes Qwen-Image will serve not just as a model, but as a foundation for further research and practical deployment across industries.

Shape
Shape
Stay Ahead

Explore More Insights

Stay ahead with more perspectives on cutting-edge power, infrastructure, energy,  bitcoin and AI solutions. Explore these articles to uncover strategies and insights shaping the future of industries.

Shape

The US will install these country-specific tariffs Aug. 7

The U.S. plans to lift its pause on country-specific tariffs while implementing a range of new rates for specific trading partners on Aug. 7, per an executive order President Donald Trump signed Thursday.  The order lists rates for over 60 trading partners, ranging from 10% to 41%. The list includes

Read More »

Inclusive by design: A new model for expanding energy engagement

Across the country, far too many households continue to carry the burden of high energy costs. These challenges are compounded by complex program requirements, limited outreach, and gaps in program awareness—making it harder for many customers to participate in offerings designed to help them. Often labeled “hard to reach,” these customers are not disinterested—they’re underserved by outdated systems and disconnected touchpoints. To address this, it’s essential to reimagine energy engagement through a more equitable lens—one rooted in data, accessibility, and local relevance. It’s time to rethink how we connect people to energy savings. One of today’s most effective—and underutilized—tools? The online marketplace. Ready to learn more? Download our latest research-backed report, Optimizing Modern Energy Programs for Financially Constrained Households and explore how to turn energy equity goals into measurable outcomes. Marketplace: More Than a Storefront Today’s energy marketplaces go far beyond online product catalogs. They serve as dynamic entry points for energy engagement—places where customers can explore incentives, enroll in programs, and take action with confidence. A well-designed marketplace reduces friction by applying rebates at the checkout, providing pre-qualified product suggestions, and creating a familiar, online shopping experience that makes participation approachable. When paired with community-specific features like language options and ZIP code-specific offers they become even more powerful. They transform into bridges for equity: approachable, actionable, and built for real people. Data-Driven Design. Human-Centered Delivery. What truly unlocks the potential of a well-designed marketplace is the integration of data-driven targeting. Advanced customer propensity models use a wide range of data—demographics, building specs, past purchases, energy usage patterns, and geographic trends—to predict which customers are most likely to act. Whether it’s buying a smart thermostat or joining a demand response program, these insights allow energy providers to focus resources where they’ll have the biggest impact. This isn’t just about segmentation,

Read More »

Energy Department Announces First Pilot Project for Advanced Nuclear Fuel Lines

WASHINGTON— The U.S. Department of Energy (DOE) today announced the next steps in the Trump administration’s efforts to strengthen domestic supply chains for nuclear fuel. DOE has conditionally selected Standard Nuclear as the first U.S. company accepted into the recently announced fuel line pilot program, announced in July 2025. The initiative, issued in accordance with President Trump’s Executive Order Deploying Advanced Nuclear Reactors for National Security helps to eliminate America’s reliance on foreign sources of enriched uranium and critical materials while opening the door for private sector investment in America’s nuclear renaissance. “With President Trump’s leadership, the Energy Department is moving at a rapid pace to unleash innovation and maintain American leadership in nuclear energy development,” U.S. Secretary of Energy Chris Wright said. “Advanced nuclear reactors will be a game-changer for the United States, and with that comes the need to fabricate the fuel these reactors. The Department of Energy is partnering private sector innovation with DOE expertise to assure stronger U.S. nuclear supply lines.” Standard Nuclear (Oak Ridge, Tennessee) is the first conditional selection under DOE’s new pilot program and will leverage the Department’s authorization process to ensure a robust supply of nuclear fuel in both Tennessee and Idaho. The advanced fuel is in high demand as reactor developers get ready to test their designs that utilize TRISO fuel. Standard Nuclear will be responsible for all costs associated with the construction operation, and decommissioning of the facility. Reactor project developers will manage the sourcing of nuclear material feedstock for fuel fabrication, which could be acquired through DOE’s high-assay low-enriched uranium allocation program. The fuel line pilot program supports DOE’s new reactor pilot program that aims to have at least three advanced reactor designs achieve criticality by July 4, 2026. Both pilot programs directly support President Trump’s executive orders to

Read More »

Dominion’s 2.6 GW offshore wind farm sees slight price hike from tariffs, CEO says

Dive Brief: Dominion Energy’s 2.6-GW Coastal Virginia Offshore Wind project is still on schedule for completion by the end of 2026, and should qualify for Inflation Reduction Act tax credits under the new safe harbor deadlines, the utility’s leadership said in a Friday earnings call. However, the company expects a slightly higher price tag for the project due to President Trump’s new tariff policies. Dominion estimates that the new tariffs will make the project more expensive by $506 million, increasing the total cost of the project to $10.9 billion. The added expenses will increase customer bills by an average of three cents a month over the entire life of the project, said Bob Blue, Dominion’s president, CEO and chairman. The utility has slightly lowered its expectations for tariff-related price increases since last quarter “despite a doubling of the steel tariff,” Blue said, “due to both working with vendors to identify cost mitigation strategies as well as completing our analysis of the final trade regulations and appendices.” Dive Insight: Blue said that the proposed tariff increases for Mexico and the European Union would add an additional $134 million to the project’s costs. “The project fabrication and installation are going very well, and CVOW continues to be one of the most affordable sources of energy for our customers,” he said. “We will install our first turbine in September, which is in line with our original schedule … In fact, we’re well ahead of plan, installing monopiles at a pace that exceeds any other U.S. offshore wind project to date.” Blue said the project has installed 134 of its 176 monopiles, or 76%, and all of its pin piles. The utility’s anticipated completion of Charybdis, the first Jones Act-compliant offshore wind turbine installation vessel in the U.S., was delayed by work on the vessel’s internal communication technology. “We

Read More »

5 court cases that could derail Trump’s tariffs

President Donald Trump has employed several powers of the executive branch to enact a wide range of tariffs, including country-specific levies. However, Trump’s future use of tariffs as a negotiating tool is likely to be determined in the courts, where several states and businesses plan to argue that he lacks the statutory authority to impose sweeping tariffs on all countries that trade with the U.S. One of the critical proceedings challenging Trump’s tariff use began Thursday. The administration is facing off against lawyers from five companies and 12 states before 11 judges of the U.S. Court of Appeals for the Federal Circuit. The court will decide whether to lift its stay of a U.S. Court of International Trade ruling that blocked Trump’s broad use of duties, including a global 10% baseline tariff and various other country-specific levies, without congressional approval. The case, V.O.S. Selections, Inc. v Trump, is one of several lawsuits with plaintiffs fighting tariffs imposed by the Trump administration. Here’s a look at five such cases and where they currently stand: 1. V.O.S. Selections, Inc. v. Trump On April 14, V.O.S. Selections and four other import companies sued the Trump administration in the USCIT, claiming the president exceeded his statutory authority under the International Emergency Economic Powers Act (IEEPA) by imposing tariffs without congressional authorization.  The plaintiffs also claimed that the tariffs hurt them financially by increasing costs and disrupting their operations. On April 23, 12 states filed a parallel lawsuit, mainly making the same arguments. So, the USCIT consolidated the two cases, and a three-judge panel ruled on May 28 that the president had no authority to impose across-the-board tariffs under the IEEPA. As a result of the findings, the USCIT issued a permanent injunction against future tariffs. The Trump administration immediately appealed the ruling to the U.S.

Read More »

PG&E sees ‘Goldilocks’ growth and path to lower bills in 2027

Dive Brief: PG&E has lined up 10 GW of data center projects — with about 1.3 GW added in the second quarter — and expects that load growth to begin reducing customer rates as soon as 2027, company executives said Thursday during a second quarter earnings call. With more than 50 projects in its queue, PG&E continues to attract mostly smaller data centers, CEO Patti Poppe said. But the company is also pursuing contracts with larger customers, including a 90-MW deal with Microsoft. The company does not plan to issue additional equity to fund its five-year, $63 billion capital plan, regardless of pending state legislation on wildfire mitigation and bill affordability, Poppe said. Dive Insight: PG&E finds itself in a “Goldilocks” zone of growth — it’s on track for enough steady growth that customers should benefit within the next few years, Poppe said Thursday. “This is because our demand is differentiated by having a diverse set of projects, ensuring that our development pipeline remains robust and not reliant on a single location, counterparty or approval,” she said, later adding that customers would likely begin to see rate benefits in 2027. Poppe estimated, as she has in the past, that every 1 GW of additional load growth brought to the PG&E area will decrease customer bills by 1% to 2%. But Poppe said PG&E has found other potential areas of cost savings as well. Currently, the average customer pays $20 a month for vegetation management, she said, but just $1 per month for undergrounding power lines. “Undergrounding is the right permanent solution in our highest-risk areas. Not everywhere, but our highest-risk areas,” she said. “So we’re going to continue to advocate for undergrounding as a key part of the wildfire mitigation solutions here in California.” The company has been piloting new

Read More »

Move fast, adapt faster: A distributed generation playbook for turbulent times

Goksenin “Goksi” Ozturkeri is a senior advisor at GS Power Partners, an independent power producer. The One Big Beautiful Bill Act (OBBBA) made its way to President Donald Trump’s desk and was signed into law on Independence Day. Fortunately for those of us in the clean energy industry — though also unfortunately — this is not our first rodeo. We have dealt with tariffs in the past, along with nearly two decades of dead stops, step-downs and extensions. Tax credits expire, then get resurrected last-minute. Projects move forward, then stall, then move forward. While these policy pendulum swings can make long-term planning feel impossible, we shouldn’t confuse complexity with catastrophe. Our sector’s secret weapon is agility. We move fast, adapt faster and often turn policy hiccups into competitive advantage. If solar and wind are indeed the fastest and cheapest sources of energy, we can solve the rest. In the distributed generation space, our real edge is speed: going from development, construction to operation in months rather than years, even under OBBBA’s compressed timelines. Companies in our space are typically nimble and able to pivot quickly, more so than many of our utility-scale counterparts. Early in my career, I identified universal truths that all renewables stakeholders, big and small, live by, and that now underpin my best practices. Migraines per megawatt One hard truth is this: big projects, big problems; small projects, big problems. Distributed generation project development suffers many of the same headaches that our grid-scale brethren do. Uniquely for distributed generation is that every project has a ZIP code, and every ZIP code has an opinion. It is just as debilitating when we’re dealing with multiple small assets, and sometimes even more, than one large project.  I call this the “migraines per megawatt” ratio — the number of local

Read More »

Colo space crunch could cripple IT expansion projects

Batson paints a pretty dire picture. “Vacant and immediately available data center space is incredibly limited. Across North America there are very few blocks available larger than 5 MW. Any second-generation space that becomes available is re-leased within weeks. Nearly 6.5 GW is under construction, of which 72% is preleased. Tenants looking to lease any sizable amount of data center capacity must wait 24 months on average.” According to a commercial real estate services firm CBRE, “Demand continues to outpace new supply across both core and emerging hubs.” Inventory across the four largest U.S. data center markets—Northern Virginia, Chicago, Atlanta and Phoenix—increased 43% year-over-year in Q1 2025. But that increase in inventory was overwhelmed by skyrocketing demand. Northern Virginia remained the tightest market, with a vacancy rate at 0.76%. Phoenix was at 1.7%, Chicago at 3.1% and Atlanta’s vacancy rate was 3.6%. What’s driving the colo crunch? Demand has outstripped supply due to multiple factors, according to Pat Lynch, executive managing director at CBRE Data Center Solutions. “AI is definitely part of the demand scenario that we see in the market, but we also see growing demand from enterprise clients for raw compute power that companies are using in all aspects of their business.” Batson agrees. “AI is driving demand, but it’s not the sole driver. We estimate AI workloads are about 20% of all data center workloads.” The big wild card contributing to the colo space shortage is that the hyperscalers are snapping up colo space as fast as it comes on the market, as they try to stay ahead of the surge in demand for AI processing from their big customers.

Read More »

DOE announces site selection for AI data centers

“The DOE is positioned to lead on advanced AI infrastructure due to its historical mandate and decades of expertise in extreme-scale computing for mission-critical science and national security challenges,” he said. “National labs are central hubs for advancing AI by providing researchers with unparalleled access to exascale supercomputers and a vast, interdisciplinary technical workforce.” “The Department of Energy is actually a very logical choice to lead on advanced AI data centers in my opinion,” said Wyatt Mayham, lead consultant at Northwest AI, which specializes in enterprise AI integration. “They already operate the country’s most powerful supercomputers. Frontier at Oak Ridge and Sierra at Lawrence Livermore are not experimental machines, they are active systems that the DOE built and continues to manage.” These labs have the physical and technical capacity to handle the demands of modern AI. Running large AI data centers takes enormous electrical capacity, sophisticated cooling systems, and the ability to manage high and variable power loads. DOE labs have been handling that kind of infrastructure for decades, says Mayham. “DOE has already built much of the surrounding ecosystem,” he says. “These national labs don’t just house big machines. They also maintain the software, data pipelines, and research partnerships that keep those machines useful. NSF and Commerce play important roles in the innovation system, but they don’t have the hands-on operational footprint the DOE has.” And Tanmay Patange, founder of AI R&D firm Fourslash, says the DOE’s longstanding expertise in high-performance computing and energy infrastructure directly overlap with the demands we have seen from AI development in places. “And the foundation the DOE has built is essentially the precursor to modern AI workloads that often require gigawatts of reliable energy,” he said. “I think it’s a strategic play, and I won’t be surprised to see the DOE pair their

Read More »

Data center survey: AI gains ground but trust concerns persist

Cost issues: 76% Forecasting future data center capacity requirements: 71% Improving energy performance for facilities equipment: 67% Power availability: 63% Supply chain disruptions: 65% A lack of qualified staff: 67% With respect to capacity planning, there’s been a notable increase in the number of operators who describe themselves as “very concerned” about forecasting future data center capacity requirements. Andy Lawrence, Uptime’s executive director of research, said two factors are contributing to this concern: ongoing strong growth for IT demand, and the often-unpredictable demand that AI workloads are creating. “There’s great uncertainty about … what the impact of AI is going to be, where it’s going to be located, how much of the power is going to be required, and even for things like space and cooling, how much of the infrastructure is going to be sucked up to support AI, whether it’s in a colocation, whether it’s in an enterprise or even in a hyperscale facility,” Lawrence said during a webinar sharing the survey results. The survey found that roughly one-third of data center owners and operators currently perform some AI training or inference, with significantly more planning to do so in the future. As the number of AI-based software deployments increases, information about the capabilities and limitations of AI in the workplace is becoming available. The awareness is also revealing AI’s suitability for certain tasks. According to the report, “the data center industry is entering a period of careful adoption, testing, and validation. Data centers are slow and careful in adopting new technologies, and AI will not be an exception.”

Read More »

Micron unveils PCIe Gen6 SSD to power AI data center workloads

Competitive positioning With the launch of the 9650 SSD PCIe Gen 6, Micron competes with Samsung and SK Hynix enterprise SSD offerings, which are the dominant players in the SSD market. In December last year, SK Hynix announced the development of PS1012 U.2 Gen5 PCIe SSD, for massive high-capacity storage for AI data centers.  The PM1743 is Samsung’s PCIe Gen5 offering in the market, with 14,000 MBps sequential read, designed for high-performance enterprise workloads. According to Faruqui, PCIe Gen6 data center SSDs are best suited for AI inference performance enhancement. However, we’re still months away from large-scale adoption as no current CPU platforms are available with PCIe 6.0 support. Only Nvidia’s Blackwell-based GPUs have native PCIe 6.0 x16 support with interoperability tests in progress. He added that PCIe Gen 6 SSDs will see very delayed adoption in the PC segment and imminent 2025 2H adoption in AI, data centers, high-performance computing (HPC), and enterprise storage solutions. Micron has also introduced two additional SSDs alongside the 9650. The 6600 ION SSD delivers 122TB in an E3.S form factor and is targeted at hyperscale and enterprise data centers looking to consolidate server infrastructure and build large AI data lakes. A 245TB variant is on the roadmap. The 7600 PCIe Gen5 SSD, meanwhile, is aimed at mixed workloads that require lower latency.

Read More »

AI Deployments are Reshaping Intra-Data Center Fiber and Communications

Artificial Intelligence is fundamentally changing the way data centers are architected, with a particular focus on the demands placed on internal fiber and communications infrastructure. While much attention is paid to the fiber connections between data centers or to end-users, the real transformation is happening inside the data center itself, where AI workloads are driving unprecedented requirements for bandwidth, low latency, and scalable networking. Network Segmentation and Specialization Inside the modern AI data center, the once-uniform network is giving way to a carefully divided architecture that reflects the growing divergence between conventional cloud services and the voracious needs of AI. Where a single, all-purpose network once sufficed, operators now deploy two distinct fabrics, each engineered for its own unique mission. The front-end network remains the familiar backbone for external user interactions and traditional cloud applications. Here, Ethernet still reigns, with server-to-leaf links running at 25 to 50 gigabits per second and spine connections scaling to 100 Gbps. Traffic is primarily north-south, moving data between users and the servers that power web services, storage, and enterprise applications. This is the network most people still imagine when they think of a data center: robust, versatile, and built for the demands of the internet age. But behind this familiar façade, a new, far more specialized network has emerged, dedicated entirely to the demands of GPU-driven AI workloads. In this backend, the rules are rewritten. Port speeds soar to 400 or even 800 gigabits per second per GPU, and latency is measured in sub-microseconds. The traffic pattern shifts decisively east-west, as servers and GPUs communicate in parallel, exchanging vast datasets at blistering speeds to train and run sophisticated AI models. The design of this network is anything but conventional: fat-tree or hypercube topologies ensure that no single link becomes a bottleneck, allowing thousands of

Read More »

ABB and Applied Digital Build a Template for AI-Ready Data Centers

Toward the Future of AI Factories The ABB–Applied Digital partnership signals a shift in the fundamentals of data center development, where electrification strategy, hyperscale design and readiness, and long-term financial structuring are no longer separate tracks but part of a unified build philosophy. As Applied Digital pushes toward REIT status, the Ellendale campus becomes not just a development milestone but a cornerstone asset: a long-term, revenue-generating, AI-optimized property underpinned by industrial-grade power architecture. The 250 MW CoreWeave lease, with the option to expand to 400 MW, establishes a robust revenue base and validates the site’s design as AI-first, not cloud-retrofitted. At the same time, ABB is positioning itself as a leader in AI data center power architecture, setting a new benchmark for scalable, high-density infrastructure. Its HiPerGuard Medium Voltage UPS, backed by deep global manufacturing and engineering capabilities, reimagines power delivery for the AI era, bypassing the limitations of legacy low-voltage systems. More than a component provider, ABB is now architecting full-stack electrification strategies at the campus level, aiming to make this medium-voltage model the global standard for AI factories. What’s unfolding in North Dakota is a preview of what’s coming elsewhere: AI-ready campuses that marry investment-grade real estate with next-generation power infrastructure, built for a future measured in megawatts per rack, not just racks per row. As AI continues to reshape what data centers are and how they’re built, Ellendale may prove to be one of the key locations where the new standard was set.

Read More »

Microsoft will invest $80B in AI data centers in fiscal 2025

And Microsoft isn’t the only one that is ramping up its investments into AI-enabled data centers. Rival cloud service providers are all investing in either upgrading or opening new data centers to capture a larger chunk of business from developers and users of large language models (LLMs).  In a report published in October 2024, Bloomberg Intelligence estimated that demand for generative AI would push Microsoft, AWS, Google, Oracle, Meta, and Apple would between them devote $200 billion to capex in 2025, up from $110 billion in 2023. Microsoft is one of the biggest spenders, followed closely by Google and AWS, Bloomberg Intelligence said. Its estimate of Microsoft’s capital spending on AI, at $62.4 billion for calendar 2025, is lower than Smith’s claim that the company will invest $80 billion in the fiscal year to June 30, 2025. Both figures, though, are way higher than Microsoft’s 2020 capital expenditure of “just” $17.6 billion. The majority of the increased spending is tied to cloud services and the expansion of AI infrastructure needed to provide compute capacity for OpenAI workloads. Separately, last October Amazon CEO Andy Jassy said his company planned total capex spend of $75 billion in 2024 and even more in 2025, with much of it going to AWS, its cloud computing division.

Read More »

John Deere unveils more autonomous farm machines to address skill labor shortage

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Self-driving tractors might be the path to self-driving cars. John Deere has revealed a new line of autonomous machines and tech across agriculture, construction and commercial landscaping. The Moline, Illinois-based John Deere has been in business for 187 years, yet it’s been a regular as a non-tech company showing off technology at the big tech trade show in Las Vegas and is back at CES 2025 with more autonomous tractors and other vehicles. This is not something we usually cover, but John Deere has a lot of data that is interesting in the big picture of tech. The message from the company is that there aren’t enough skilled farm laborers to do the work that its customers need. It’s been a challenge for most of the last two decades, said Jahmy Hindman, CTO at John Deere, in a briefing. Much of the tech will come this fall and after that. He noted that the average farmer in the U.S. is over 58 and works 12 to 18 hours a day to grow food for us. And he said the American Farm Bureau Federation estimates there are roughly 2.4 million farm jobs that need to be filled annually; and the agricultural work force continues to shrink. (This is my hint to the anti-immigration crowd). John Deere’s autonomous 9RX Tractor. Farmers can oversee it using an app. While each of these industries experiences their own set of challenges, a commonality across all is skilled labor availability. In construction, about 80% percent of contractors struggle to find skilled labor. And in commercial landscaping, 86% of landscaping business owners can’t find labor to fill open positions, he said. “They have to figure out how to do

Read More »

2025 playbook for enterprise AI success, from agents to evals

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More 2025 is poised to be a pivotal year for enterprise AI. The past year has seen rapid innovation, and this year will see the same. This has made it more critical than ever to revisit your AI strategy to stay competitive and create value for your customers. From scaling AI agents to optimizing costs, here are the five critical areas enterprises should prioritize for their AI strategy this year. 1. Agents: the next generation of automation AI agents are no longer theoretical. In 2025, they’re indispensable tools for enterprises looking to streamline operations and enhance customer interactions. Unlike traditional software, agents powered by large language models (LLMs) can make nuanced decisions, navigate complex multi-step tasks, and integrate seamlessly with tools and APIs. At the start of 2024, agents were not ready for prime time, making frustrating mistakes like hallucinating URLs. They started getting better as frontier large language models themselves improved. “Let me put it this way,” said Sam Witteveen, cofounder of Red Dragon, a company that develops agents for companies, and that recently reviewed the 48 agents it built last year. “Interestingly, the ones that we built at the start of the year, a lot of those worked way better at the end of the year just because the models got better.” Witteveen shared this in the video podcast we filmed to discuss these five big trends in detail. Models are getting better and hallucinating less, and they’re also being trained to do agentic tasks. Another feature that the model providers are researching is a way to use the LLM as a judge, and as models get cheaper (something we’ll cover below), companies can use three or more models to

Read More »

OpenAI’s red teaming innovations define new essentials for security leaders in the AI era

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More OpenAI has taken a more aggressive approach to red teaming than its AI competitors, demonstrating its security teams’ advanced capabilities in two areas: multi-step reinforcement and external red teaming. OpenAI recently released two papers that set a new competitive standard for improving the quality, reliability and safety of AI models in these two techniques and more. The first paper, “OpenAI’s Approach to External Red Teaming for AI Models and Systems,” reports that specialized teams outside the company have proven effective in uncovering vulnerabilities that might otherwise have made it into a released model because in-house testing techniques may have missed them. In the second paper, “Diverse and Effective Red Teaming with Auto-Generated Rewards and Multi-Step Reinforcement Learning,” OpenAI introduces an automated framework that relies on iterative reinforcement learning to generate a broad spectrum of novel, wide-ranging attacks. Going all-in on red teaming pays practical, competitive dividends It’s encouraging to see competitive intensity in red teaming growing among AI companies. When Anthropic released its AI red team guidelines in June of last year, it joined AI providers including Google, Microsoft, Nvidia, OpenAI, and even the U.S.’s National Institute of Standards and Technology (NIST), which all had released red teaming frameworks. Investing heavily in red teaming yields tangible benefits for security leaders in any organization. OpenAI’s paper on external red teaming provides a detailed analysis of how the company strives to create specialized external teams that include cybersecurity and subject matter experts. The goal is to see if knowledgeable external teams can defeat models’ security perimeters and find gaps in their security, biases and controls that prompt-based testing couldn’t find. What makes OpenAI’s recent papers noteworthy is how well they define using human-in-the-middle

Read More »