Apple Intelligence’s launch in Beta has been met with mixed reactions as people realize how limited the release truly is. The reactions here are similar to many of the earlier instances where AI hype quickly met reality, especially given that a huge component of Apple’s user experience, Siri, is not coming with Apple Intelligence until next year.

Apple Intelligence’s launch in Beta has been met with mixed reactions as people realize how limited the release truly is. The reactions here are similar to many of the earlier instances where AI hype quickly met reality, especially given that a huge component of Apple’s user experience, Siri, is not coming with Apple Intelligence until next year.

Investor discontent over CapEx — Many tech firms announced earnings this week, and overall performance was good to excellent, depending on the company and the metrics being watched. Interestingly, one area that came up again and again was the amount of future capital investment earmarked for AI workloads. While Alphabet elected to hold steady, other firms including Microsoft and AWS are committing to increase their rate of investment.

It’s a tricky proposition from both a strategic and investor point of view. For investors, most tech firms are not sharing a lot of data about AI-specific revenues and expenses. Microsoft does share a bit more financial data than other firms and was able to point to a healthy growth trajectory. However, given that Microsoft is talking about spending potentially more than $50 billion on CapEx next year, the outlay still dwarfs whatever AI revenue growth they have seen to date. AWS took a more holistic view in noting that there was growth across many types of workloads including AI to justify the continued investment. It also did not hurt that AWS said that it now has a $150 billion backlog. So demand for cloud services—including AI—is clearly there.

In spite of that, investors have a right to question when the ROI will happen, given the limited data being reported. On the vendor side, we are seeing what is effectively an arms race to be the biggest player in the next wave of technology. But given the infancy of the AI market, the cost of infrastructure, and the inefficiency of AI models, that race is not only expensive but also open to future cost disruption. This should not be a big surprise, and it’s going to dictate a tense balancing act for a while.

And just like that, foundation models are commodities —  A year ago, everyone was talking about which LLM was superior under what circumstances. And while we still are seeing new features and model sizes on what feels like a weekly basis, we have also seen a breakneck pace toward commoditization. This is being driven by three key factors.

First, there is broader choice and distribution. All major cloud providers acknowledge that there is no one perfect model, so they now host multiple models on their respective clouds. Second is the open-sourcing of some models, which reduces both costs and training hurdles. Like most things open source, the presence of a viable open alternative has a reverberating effect on all vendors, open or closed.

Third is the availability of end-to-end AI stacks that are optimized for certain types of performance. The inclusion of specialized instances on custom-designed chips in clouds is generating new cost-per-performance metrics. So while you may be able to choose any model you want from a cloud provider, those providers are also doing what it takes to optimize the infrastructure to fit the model they would most like you to use, sometimes leading to dramatic pricing differences.

Atlassian State of Developer Experience 2024 research shows some positive signals from devs — Atlassian recently conducted a broad-based developer survey to help better understand the requirements in providing a good developer experience. The report is free and available from Atlassian. It is worth reading, but here are a couple of quick highlights.

First, there is a big disconnect between developers and leadership on how developers could be more productive. While leaders see the hurdles to developer productivity as being about capacity and shifting roles (sins of the present), devs see barriers such as technical debt and poor documentation (sins of the past). Second, despite the disconnect with leadership, there are positive feelings about the future of developer experience. For instance, while many devs are seeing fairly low benefit from AI tools today, they also predict improvements leading to a much bigger impact in the next two years.

Finally, Devs are also saying that the developer experience is getting more attention because it is increasingly a factor in developer retention—and attrition. So while things aren’t necessarily great, they are improving. Not surprisingly, all of these findings tie into Atlassian’s value proposition and vision of the future; however, it would be unfair to write off these findings as mere marketing. As I pointed out in my recent piece on agile development, another recent survey pointed to psychological safety as a major factor in developer morale. So, if Atlassian has a solution that can better bridge the gap between devs and leadership, organizations should at least look into it.

AMD’s datacenter quarter impressed. $2.8 billion in revenue represented 115% growth year over year, bolstered by sales of MI300 chips that seemingly far outperformed expectations. The company announced extensive use of the GPU by Microsoft to help power its GPT Turbo and many Copilot services. And as the company announced that the MI300 crossed the $1billion per quarter mark for the first time, it upped its revenue forecast for the product line from $4 billion to $4.5 billion for 2024.

GPU wasn’t the only big winner in the datacenter segment. EPYC also saw strong double-digit growth in both the cloud and enterprise, with notable wins such as Netflix, Uber, Boeing, Siemens, and Adobe. Further, AMD announced that more than 900 public cloud instances are now available on this CPU. As a nice bonus, more than one-third of the company’s quarterly enterprise bookings for EPYC were new customers.

What is driving this? It’s a symbiotic dynamic between the MI300 and EPYC. While EPYC stands strong on its own, it’s also being deployed along with every GPU being sold into the cloud and into the enterprise—perhaps into enterprise datacenters that were previously Intel customers. The adoption of MI300 for AI and HPC-like workloads, accompanied by EPYC CPUs, will lead to broader adoption of EPYC for general-purpose compute needs, because enterprise IT prefers a single CPU supplier for the sake of easier provisioning and management.

In contrast to AMD’s stellar datacenter quarter, Intel’s struggles are real. While the company has been executing against its “five nodes in four years” strategy, it continues to lose datacenter share across all market segments—hyperscale, enterprise, mid-market, and below. Whereas its competition saw a 115% gain in datacenter revenue, Intel’s DCAI (Xeon CPUs and Gaudi AI accelerators) saw decreases in revenue, operating income, and operating margin year over year.

Some of this is to be expected as Intel continues to claw its way back to performance relevance. However, the pricing tactics Intel has employed to stave off the competition in the enterprise are beginning to stall. This indicates that Intel has lost its lock on enterprise IT—a market resistant to change, and one that AMD had been unable to penetrate until just recently. Once this market’s momentum shifts, it is very difficult to reverse. The next quarter is going to be critical for Intel’s DCAI business segment: Can the company somehow stem the flow with the promise of its next-generation Granite Rapids server processors?

Commvault reported positive fiscal Q1 2025 results, with total revenues of $225 million, up 13% year over year, and ARR growing by 17% to $803 million. Subscription revenue increased 28%, reaching $124 million. The company also repurchased $51.4 million worth of shares. The data management and data protection industry will undoubtedly continue to grow. As global enterprises adopt more SaaS applications and generate more data, the risk of cyberattacks increases, requiring advanced data security technologies.

Dynatrace, an observability and security platform, just celebrated the fifth anniversary of its IPO. Since its market debut in 2019, its stock has increased from $16 to $44. The company has exceeded $1 billion in annual revenue and now has more than 4,000 organizations using its observability capabilities. How much of this can be attributed to the proliferation of SaaS applications, and how much to the features and functionality of the platform? I will analyze the data protection industry in an upcoming article.

Adobe released data on its partnership with the California State University (CSU) system and Adobe Creative campuses. Recognizing the importance of digital literacy and creative skills in today’s workforce, CSU partnered with Adobe to provide students and faculty access to Adobe’s Creative Cloud suite of applications. Through the program, students get the hands-on experience with creative and generative AI tools that employers are increasingly looking for. According to Adobe’s case study, at Fresno State 100% of the students completing the capstone broadcast journalism class in Spring 2023 found jobs immediately after graduation, many working in newsrooms as reporters, producers, or designers. However, the program is not only for creative positions or marketing content. Finance students, for example, use Adobe Audition to produce podcasts that showcase their understanding of financial concepts. This demonstrates—and probably reinforces—their expertise and very likely enhances their communication skills.

Canva has announced its intent to acquire Leonardo.AI, an Australian AI content production platform, aiming to strengthen its in-house AI capabilities and accelerate the development of its AI-powered design tools.

Canva has also expanded its partnership with Getty Images, focusing on enriching Canva’s content library with Getty Images’ high-quality stock photos and ensuring fair compensation for creators and IP rights holders whose work is used in training Canva’s generative AI tools. This collaboration is part of Canva’s $200 million Content Fund.

I’m eager to see how these strategic decisions will impact Canva’s future growth and competitive positioning in the market. The question remains: are these moves sufficient for Canva to compete with Adobe Firefly and Adobe Stock in the enterprise arena?

Integration is a key data management element, and Informatica’s 2nd quarter 2024 results reflect industry trends with its iPaaS solution and Intelligent Data Management. To quote the company’s earnings release:

• Cloud Subscription ARR increased 37% year-over-year to $703 million

• Total ARR increased 7.8% year-over-year to $1.67 billion

• Results within or above all second-quarter 2024 guidance metric ranges

As technology stacks expand with SaaS applications, managing and understanding your data is essential. Solution providers like Informatica offer tools to analyze and maximize the value of your data, ensuring it supports business goals effectively.

HPE took one step closer to approval of its $14 billion acquisition of Juniper Networks when it received unconditional approval by the European Commission so the deal can go ahead in the EU. This development points to the likelihood of the same happening in the United States and the rest of the world. At least in prospect, the combined companies could become a powerful force, covering cloud, data center, and edge connectivity infused with technology from Juniper’s AI-oriented Mist Systems acquisition four years ago.

Blue Yonder closed its acquisition of One Network Enterprises for $839 million, its third acquisition since Q4 2023. This enhances Blue Yonder’s supply chain platform by enabling real-time data sharing across the supply chain. The integration provides AI-powered supply chain assistants and a unified view of inventory and capacity. This acquisition helps businesses respond faster to market changes, reduce costs, and improve service levels, aligning with Blue Yonder’s strategy to create interconnected supply chain ecosystems.

Popular PC shooter Valorant has finally landed on consoles after being a PC exclusive for years; you can now play Valorant on both Xbox and PlayStation 5. This is great for console gamers, who should have the opportunity to play the game, and it’s a great thing for Valorant’s maker, Riot Games, because the PC market has long been saturated.

This week’s update covers two open-source, industry-standard enabling technologies that will influence IoT and edge product development over the next few years: Thread for device networks and WebAssembly for application containers.

Thread is now ten years old, and it’s time for consumer product companies to take a fresh look at this home automation network standard. Thread is an Internet Protocol-based replacement for Zigbee, Z-Wave, and other specialized low-power, low-bandwidth connectivity schemes. Thread has broad industry support. In fact, it’s probably already in your home, built into products from Amazon, Apple, Google, and many other manufacturers.

These big players adopted Thread ahead of mainstream consumer demand for four reasons. First, the Matter interoperability standard, founded by many of the same big companies sponsoring Thread, supports only Wi-Fi and Thread wireless networks. Second, Thread is an IP network, so the same device commands that work over Wi-Fi also work over Thread without translation. Third, Thread hardware and software are available off-the-shelf from top-tier silicon providers, so the technology is mature and shipping in volume. Fourth, Thread products are already broadly deployed, and the product ecosystem is snowballing. The device connectivity wars are over. Matter and Thread won. From now on, consumer product developers should use Wi-Fi and Thread for wireless LAN device connectivity—Wi-Fi for high bandwidth and Thread for low power.

WebAssembly (Wasm) is a lightweight application container technology that can potentially change how developers build IoT applications. Wasm programmers compile code written in C++, Rust, or other languages into binary instruction code that runs on PC and smartphone browsers at nearly native speed. It’s a faster alternative to JavaScript, and can run alongside it.

But here’s the good news for IoT. The client-side runtime can be standalone rather than browser-hosted. In this case, WebAssembly is a lightweight way to deliver applications to IoT devices, even battery-powered products. This approach offers significant performance advantages over Docker or Kubernetes containers, potentially allowing IoT platforms to support a wide variety of applications. Instead of building custom hardware and software stacks for each IoT product, developers can use off-the-shelf platforms common to many products and customize only the container-based app software. The process would be comparable to building and deploying phone apps.

Containerized apps accelerate IoT product development, improve security, simplify long-term support, and reduce costs. I’m watching Atym, a startup that’s sponsoring Ocre, an LF Edge open-source project that aims to realize this vision. Commercialization is still a long way off, but industrial IoT product architects should monitor the project—and maybe contribute.

Qualcomm’s new Snapdragon 4s Gen 2 is the chip that the company claims will enable 5G in devices under $100 and could help expand 5G access to 2.8 billion people. I believe that this is good for both consumers and operators because it will push the industry further towards 5G standalone (5G SA), which is the version of 5G that will deliver the use cases people have been expecting from 5G.

Ookla’s new 5G SA report shows the significant performance improvements that 5G SA can deliver, as well as which regions and carriers are doing well versus which ones are lagging behind. India and China are leading the charge with the U.S. close behind thanks to T-Mobile, but Europe is lagging, as are places that don’t really have 5G yet, including much of Africa and Latin America.

RingCentral reported strong second-quarter 2024 results, exceeding expectations with a 10% year-over-year increase in subscription revenue and total revenue. This growth reflects the company’s ongoing focus on cloud communications and AI innovation. The company remains optimistic about its future prospects, highlighting a solid pipeline and continued investments in strategic areas. A couple of highlights from the report include: RingCentral total contact center business stands at $390 million of ARR, up 90% YoY, and Cox Communications has chosen RingCentral as its partner for upcoming UCaaS and CCaaS solutions, expected to be rolled out later this year. RingCentral continues to perform in a very competitive environment, and its investments in AI look to be paying off.

Microsoft’s Q4 2024 earnings show solid early adoption of 365 Copilot, including a surge in seats and large deployments. This is particularly evident in CX and contact center, where Microsoft sees substantial cost savings due to AI.

On the earnings call, Microsoft CEO Satya Nadella said he believes that Copilot for Microsoft 365 is transforming how knowledge and frontline workers approach their tasks, similar to the impact of GitHub Copilot on software engineering. He sees Copilot enabling a new design system for work, where tasks can be broken down into steps such as issue identification, planning, specification, and execution. Businesses have adopted this Copilot workflow approach across various functions, including marketing, finance, sales, and customer service.

Also notable in this quarter’s earnings was the skyrocketing adoption of Microsoft Teams Premium, which now has over 3 million seats in use, a YoY increase of roughly 400%. Major organizations such as Eli Lilly and Ford are opting for the advanced features that Teams Premium offers, including end-to-end encryption and real-time translation.

There have been lots of discussions and memes around Logitech’s desire to turn the mouse into a subscription product. This has created a lot of pushback from the industry, which has traditionally monetized on hardware but not on software. That said, I do believe that companies such as SteelSeries, which makes gaming peripherals, are doing a better job of monetizing on both fronts and creating added value rather than taking away value just to resell it.

It is becoming apparent that the flubbed CrowdStrike EDR update that triggered the feared Microsoft Windows blue screen of death (which I analyzed here) could equate to billions of dollars in lost revenue for many of its customers. Delta Airlines’ CEO recently disclosed that his company suffered $500 million in operational costs because of the CrowdStrike fiasco. From my perspective, the losses for Delta and other businesses could be far greater than that in terms of reputation damage, lost customers, and future revenue. I believe that CrowdStrike can recover, but it will be incumbent on the embattled cybersecurity solution provider to be transparent in its efforts to shore up deficiencies within its developer operations.

It was a busy week for earnings, with Arm, Qualcomm, AMD, Intel, Apple, and T-Mobile all issuing reports. Arm’s earnings beat expectations while its guidance aligned with investors’ expectations, yet the stock still slipped. Qualcomm’s earnings showed some of the same strength in the mobile market that Arm’s did, but with additional strength in automotive as well as improving conditions in IoT, which created a beat-beat-raise scenario that jolted Qualcomm’s stock upward. AMD also had strong earnings, especially with the consumer business showing 50% higher revenue and datacenters continuing to be a bright spot with Instinct GPUs for AI.

Intel’s earnings were disappointing, even though its consumer business keeps the company afloat. Intel has a strong future ahead of it, but laying off 15% of its workforce—as it just announced—makes this a tough stretch in that journey. Meanwhile, Intel is giving its 13th and 14th Gen desktop processor customers an additional two years of warranty as a result of the increased failure rates of those CPUs due to issues that have been identified recently by PC OEMs and reviewers.

T-Mobile had very strong earnings, adding 777,000 net postpaid customers and generating considerable profits thanks to its leading 5G network.

NVIDIA is facing two antitrust probes from the U.S. Department of Justice connected to its market dominance, which should surprise nobody considering the position that it holds. That said, it will be interesting to find out whether there have been any specific actions or tactics the company has taken to maintain its leadership position that might get it in hot water.

T-Mobile has launched its “Friday Night 5G Lights” program, which aims to bring enhanced 5G network infrastructure to high school football fields across rural America. This initiative complements T-Mobile’s efforts to improve connectivity in large venues such as MLB stadiums (something I just wrote about in connection with the MLB All-Star Game). I appreciate the company’s efforts to ensure high-quality mobile experiences for customers in both urban and rural areas. This initiative will also help establish brand affinity with a high-school demographic likely to be in the market for a mobile carrier provider soon—assuming they will either be first-time buyers or taken off of their parents’ plan. I am also a big proponent of reliable connectivity in places where increased capacity would be needed in the case of an emergency with a large crowd.

Qualcomm showed off its front-of-jersey sponsorship with Manchester United last week as the Snapdragon Stadium hosted the Snapdragon Cup featuring Manchester United vs. Real Betis. I had insightful conversations with Qualcomm CEO Cristiano R. Amon and CMO Don McGuire about the vast potential of this partnership, including how it could revolutionize the fan experience through enhanced connectivity and innovative digital offerings beyond the game itself.

While it’s challenging to quantify the precise ROI of such a high-profile sponsorship, the strategic alignment between Qualcomm and Manchester United is undeniable. Man U.’s global fanbase aligns exceptionally well with key Snapdragon markets, presenting a unique opportunity to showcase Qualcomm’s tech worldwide. I’m eager to see how Qualcomm continues to leverage this partnership in ways that are tangibly tied to the bottom line.

Last week, Game Time Tech (Melody Brue, Robert Kramer, and Anshel Sag) hit the road to take a closer look at how Qualcomm is influencing the world of sports technology.

Visit to Petco Park: Qualcomm has enhanced Petco Park’s operations and fan experience. Snapdragon-powered sensors collect real-time data to improve efficiency. Fans enjoy an AR experience in the Padres Hall of Fame and benefit from high-speed Wi-Fi throughout the park, enabling them to stay connected, share updates, and access real-time game statistics, creating a more connected and engaging experience.

Visit to Snapdragon Stadium: Qualcomm hosted a great event to kick off its historic Snapdragon sponsorship of the front of Manchester United’s jersey. Effective sponsorships go far beyond logo placement, integrating technology, data, and experiences to create meaningful connections. Data is leveraged to understand fans better and provide valuable information that enhances the overall fan experience. Qualcomm’s use of its Snapdragon technology demonstrates how sponsors can integrate their products directly into the event experience, transforming venues into smart stadiums.

Coming soon: Our podcast with Qualcomm CMO Don McGuire to discuss Snapdragon sponsorship of the Man United jersey.

I believe that the opportunity for mobile network programmability may finally come to fruition. Ericsson may have been ahead of the market, given its near-complete writedown of its $6 billion acquisition of Vonage, but rumors of a consortium of communications service providers backing the Vonage platform could breathe new life into the effort. Nokia also offers its network-as-code platform and recently announced an initiative with a new ecosystem partner to focus on application development that leverages the power of 5G for healthcare and utilities. It could ultimately turn into a two-horse race, but competition in the technology industry typically fosters innovation.