The new Google Pixel 11 Pro XL doesn’t dramatically reinvent the company’s flagship smartphone formula. From a hardware perspective, this year’s model represents an incremental update. But after spending time with Google’s latest full-featured Pixel, I’ve found that some of its most meaningful improvements for business users aren’t necessarily the things that stand out on a spec sheet or benchmark chart.

Starting at $1,299 with 256GB of storage, the Pixel 11 Pro XL features a 6.8-in. Super Actua LTPO OLED display, Google’s new Tensor G6 processor, a 5115mAh battery and an updated triple-camera system. Google also pledges seven years of operating system, Android security and Pixel Drop updates.

The hardware is very good, but Google’s deeper integration of Gemini with Android and other thoughtful software features are what make this device more interesting.

For much of the past several years, Intel has been defined more by what went wrong than by what the company accomplished. Manufacturing delays, product roadmap slips and growing competitive pressure from AMD and NVIDIA reshaped both the company's reputation and investor sentiment. The prevailing narrative became almost self-perpetuating: Intel was continually rebuilding, while everyone else was executing.

I don't believe that narrative accurately reflects where Intel stands today.

Of course, Intel's turnaround isn't complete. The company still faces significant competitive challenges, particularly in AI infrastructure, and Intel Foundry must continue proving itself with external customers at scale. However, after following Intel's engineering roadmap and manufacturing strategy closely over the past several years, I believe we've reached an inflection point where tangible improvements in manufacturing, product execution and long-term strategy are beginning to strengthen the company's competitive position. This is no longer a turnaround built primarily on promises. It's increasingly supported by measurable progress.

AMD executives took to the stage at its Advancing AI 2026 event in San Francisco today to detail the company’s next generation of AI infrastructure solutions, from Instinct MI455X AI accelerator GPUs and 6th Gen EPYC “Venice” CPUs, to Pensando networking, ROCm.AI software and its Helios rack-scale platform that ties it all together.

AMD has been working towards rack-scale AI system solutions for years. Its ZT Systems acquisition last year added valuable engineering talent and intellectual property that is now finally bearing the real fruits. Its Helios AI platform is a major platform evolution for AMD, with shipments scheduled to begin in the second half of this year (which is here and now).

Nvidia has expanded well beyond GPUs over the years, building an increasingly more diverse and comprehensive AI platform spanning networking, DPUs, high-speed interconnects, rack and pod-scale systems and software. With its upcoming Vera CPU, the company is adding another strategic product, a custom Arm-based server processor featuring its own custom “Olympus” microarchitecture, designed specifically for emerging agentic AI workloads.

One of the recurring themes across the AI industry this year has been making powerful hardware easier to use and develop on. AI accelerators continue to improve in performance and efficiency, but getting applications from concept to deployment can still require significant engineering effort. As edge AI adoption expands across industrial automation, robotics, retail analytics, healthcare and physical security, developer productivity is becoming an increasingly important part of the conversation.

That is the backdrop for Axelera AI's Voyager Wingman announcement. Voyager Wingman is a new AI-powered assistant built specifically for developers targeting the company's Metis edge AI accelerators through Axelera AI’s Voyager SDK.

As systems and AI infrastructure get more complex, the engineering challenges required to design and bring them to market extend well beyond silicon. Printed circuit boards, advanced packaging, power delivery, thermal management, materials and mechanical structures are all critical to system performance and reliability as well, and co-optimization of all aspects of a design is effectively a requirement today.

Cadence Design Systems believes the next level of AI-assisted engineering resides at the system level, and to that end, the company just announced its new AuraStack AI Super Agent, an agentic AI platform for PCB and advanced chip packaging.

Qualcomm's Investor Day in Manhattan last week marked what CEO Cristiano Amon described as the company's next major transformation—expanding from a leader in connected edge computing into a company building AI infrastructure that spans everything from ultra-low-power IoT devices to hyperscale data centers.

That's a broad, bold vision, but unlike many AI narratives circulating across the semiconductor industry, Qualcomm backed it with a coherent technology roadmap, meaningful customer engagement and measurable financial targets.

The handheld gaming market has evolved from a niche category into one of the most competitive segments in PC hardware. For years, AMD has powered many of the devices that defined the category, but Intel's Arc G3 Extreme marks something different. It represents the company's first truly dedicated handheld gaming platform and demonstrates just how much progress Intel has made in graphics, software and power management.

The MacBook Neo's launch has sparked much debate, with some arguing that PC manufacturers shouldn't copy Apple. Despite the Neo's praised $599 starting price, it achieves this by using older iPhone-class silicon, limited RAM, and basic connectivity. While its premium aluminum chassis stands out, many affordable Windows laptops offer superior specifications, including newer processors, more memory, and better connectivity, often at similar or lower price points.

By integrating Ansys simulation technologies more deeply into its core EDA workflows, Synopsys is making a strategic bet that future gains in semiconductor performance and efficiency will come less from optimizing individual components and more from understanding how increasingly complex systems behave as a whole. Simultaneous co-optimization will be key.

For much of the last decade, the Windows on Arm ecosystem has felt like a market perpetually on the verge of a breakthrough. The underlying advantages have never really been in question. Potentially better power efficiency, tighter integration, increasingly capable graphics and AI accelerators all align well with where modern computing is headed. Yet despite steady progress from Microsoft and key silicon partners like Qualcomm with its excellent Snapdragon X platforms, Windows on Arm has remained a relatively small part of the broader PC market.

NVIDIA and its ever-tenacious CEO Jensen Huang think they can change this reality.

Qualcomm has been working for years to broaden how investors view the company, via an aggressive revenue diversification strategy, and lately the market appears willing to listen. The recent run-up in QCOM shares seems tied to a wider reassessment of its growth story, with Automotive, IoT, AI PCs and more recently data center AI giving investors more to consider beyond traditional handset cycles.

Tenstorrent stepped out boldly today in the AI infrastructure space, with the launch of its Galaxy Blackhole system powered by the company’s Blackhole accelerators. And while the company’s highlighted use cases span real-time AI video generation and large-language-model inference, perhaps the more substantive takeaway is architectural. Tenstorrent is positioning Galaxy as a system-level AI platform designed to compete on sustained inference performance, high-speed memory access and scalable networking, three factors that increasingly define real-world AI deployment efficiency.

At a high level, Tenstorrent is making a case that AI infrastructure performance is not determined solely by peak compute throughput.

At its annual CadenceLIVE event this week, Cadence Design Systems delivered a pair of announcements that signal an evolution in how the company sees its role in the AI era. One partnership deepens its long-standing relationship with Nvidia around accelerated computing, digital twins and large-scale system simulation. The other integrates Google’s Gemini models into Cadence’s ChipStack AI platform to advance agent-driven design automation in the cloud.

Individually, each announcement is incremental. Collectively, they outline a broader industry shift, where engineering workflows are becoming increasingly agent-orchestrated, simulation-first and tightly coupled to the infrastructure they ultimately run on.

SiFive has been a long-standing commercial force behind RISC-V processor IP, and the company announced today it has closed a $400 million Series G financing round, which actually was oversubscribed, valuing the company at $3.65 billion. The round was led by Atreides Management, with participation from Apollo Global Management, NVIDIA, Point72 Turion, and T. Rowe Price Investment Management, alongside returning backers Prosperity7 Ventures and Sutter Hill Ventures. More than a vote of confidence in SiFive as a business, this financing round signals where AI infrastructure is heading, and about the recently reinvigorated role of the CPU in agentic AI workloads.

As generative AI tools continue to mature, the practical question for many small businesses and creative professionals isn’t if they should adopt AI, but where it delivers measurable value. In this latest HotTech Vision and Analysis research report, we take a hands-on look at real-world creative workflows, from photo editing and client communications to marketing asset generation, and quantify the productivity impact of modern AI-accelerated PCs.

At this week’s Imagine event in New York, HP made a wide range of announcements spanning new hardware, software, and collaboration technologies. As expected, there were many refreshed PCs and workstations aimed squarely at the growing AI PC segment along with a myriad of AI-infused tools and technology. The overarching message, however, wasn’t simply tied to performance or laptop features, though there was a lot of that with over 30 new systems announced.

While much of the attention around Nvidia’s agentic announcements from GTC centered on its support of the OpenClaw agentic AI platform and the company’s NemoClaw software stack, the OpenShell component of its AI toolkit may ultimately prove to be one of the more important pieces of the puzzle, particularly when it comes to enterprise adoption.

Earlier in the week, Synopsys announced its Electronics Digital Twin Platform, which is an open solution designed to accelerate the creation, management, deployment, and use of digital twins to help engineers co-design and optimize silicon and software. The Synopsys eDT platform will Initially target automotive use cases to reduce vehicle development cost and speed time-to-market, by allowing customers to configure what Synopsys calls cloud-based eDT Labs. eDT Labs bring together an array of Synopsys technologies, along with open‑ecosystem tools, models, software, and scalable compute, to enable early evaluation of new System-on-Chips or microcontrollers, system validation, and early software development, even before physical hardware is available. And today, at its Synopsys Converge event currently underway in Santa Clara, California, the company announced a wide array of new systems, tools, and initiatives to further accelerate design and integration for a variety of use cases, including, Synopsys AgentEngineer technology, Ansys 2026 R1 and Synopsys Multiphysics-Fusion technology, and new Hardware-Assisted Verification platforms.

Dave Altavilla offers commentary on Nvidia's (NVDA) earnings minutes following the report. He says the company's record quarter shows that Nvidia is "powering" the AI trade on all fronts and helps ease fears of massive CapEx spending in hyperscalers like Amazon (AMZN), Microsoft (MSFT), and Meta Platforms (META). Dave also explains why he doesn't see margin compression in the cards for Nvidia.