Welcome to the August 2023 edition of neXt Curve Insights. This monthly newsletter is a compilation of articles, media, and news that have been curated by the research team at neXt Curve with contributions from partner analysts as well as business and technology leaders.
The goal of neXt Curve Insights is to provide our readers with a regular cadence of coverage of the industry and tech trends and events that matter with the intent of fostering constructive discussion and debate on the future of technology, innovation, and the continuous reinvention of enterprise, industry, society, and our lives.
I hope that you find this edition informative and inspiring.
Leonard Lee, Executive Analyst of neXt Curve
Special Announcement
neXt Curve and Transforma Insights, two leading technology research firms, announce strategic partnership to offer Essential Validated Insights to senior management teams, drawing on both companies’ formidable respective strengths.
Top Posts
Check out the top social media posts on LinkedIn by neXt Curve analysts, associates, and partners. Follow neXt Curve and Leonard Lee on LinkedIn and be part of the conversation. Click on the image to view the post.
I’m listening in on the AMD earnings call. Interesting results. Looking for evidence of a significant #generativeai bump. Lots of evidence of industry adjustment and bumpiness. Interesting compare to Intel Corporation‘s results.
I’m listening to the Qualcomm earnings call. Tough quarter coming off of a tough YoY compare given the smartphone & PC malaise and challenging macro especially China. Automotive is a bright spot with steady but moderate growth YoY.
I stumbled on this piece by David Linthicum in CIO Online who is a long-time thought leader in cloud computing. I thought it was interesting that he wrote this piece which outlines some of the shifts in thinking about cloud computing toward hybrid and edge cloud.
Thanks to Linda Hardesty for letting me share some of my thoughts about Tareq Amin’s departure from Rakuten Symphony. Tareq is arguably THE pioneer of Open RAN deploying the first nationwide ORAN based network with Rakuten Mobile.
All the best, Tareq in your next endeavor.
This is the big news of the day. Intel Corporation cancels the Tower Semiconductor acquisition. My take, the industry situation has changed and many of the assumptions going into the CHIPS Act, nor relevant. Policy and investments of sync. Time to focus on leading edge.
Check out this great chat I had with Matthew Trifiro of Vapor IO in this great podcast program, “Over The Edge” sponsored by Dell Technologies.
You can go to the podcast by clicking right here 👉 https://lnkd.in/gRU8GuST. Please liked and share!
Thanks to Matt and the folks at Dell Technologies! That was one hell of a jam!
Well, there you go!
Top Headlines
These are the hot headlines in the tech and industry media that neXt Curve has curated for your consideration and attention. Executive analyst, Leonard Lee, provides a brief analysis of each story. Contact, Leonard at leonard.lee@next-curve.com for a briefing on the details of his take (clients only).
Headliner
Summary
Analysis
T-Mobile launches 5G network slicing beta for developers
by Monica Allen of Fierce Wireless
August 2, 2023
“T-Mobile is taking advantage of its first-to-market 5G standalone (SA) network status and launching a 5G slicing beta for developers.
It’s already got beta participants lined up that include Dialpad, Google, Webex by Cisco, Zoom and more.
In the early days of 5G, network slicing was touted as one of the big ways operators would benefit from the next generation of wireless technology. But most operators started with non-standalone (NSA) 5G networks before moving to 5G SA. T-Mobile was first with a nationwide 5G SA.
“The wireless industry has talked about 5G network slicing for years and at T-Mobile we’ve been putting in the work to bring it to life,” said T-Mobile President of Technology Ulf Ewaldsson in a statement.”
This network slicing beta is a good and early indicator of the differentiating capabilities and business opportunities operators will be able to realize as they modernization their networks toward 5G SA.
Maybe network slicing is not a myth but a potentially powerful feature. That will allow T-Mobile to differentiate its business services in the market.
The jury is out whether T-Mobile can build a developer community for its network. Will developers want to deal with one operator?
T-Mobile is a supporter of Project CAMARA, so there may be some interoperability in the future that will allow them to participate in a “neutral host” developer community if they need it.
“Deutsche Telekom (DT) teamed with specialist Mira on a trial of teleoperated driving using its 5G network, with a remote control shuttle service set to be used as transport between some of the operator’s sites in Bonn, Germany.
DT explained the pilot would offer in-depth insights into the requirements for teleoperated driving, including on 5G infrastructure, which it indicated would be important for the future development of driverless mobility.
In the test the operator’s 5G network will be used to transmit large amounts of data in near real time between the shuttles and the control station.
Mira MD Klaus Kappen stated he was ‘convinced that teleoperated driving will play an important role in improving the efficiency and sustainability of transport” adding in the trial it could develop “solutions for the mobility of the future and test them on public roads today.'”
In recent years, the idea that autonomous vehicles will be 5G dependent has diminished as it also has become apparent that we are a long way off before we see SAE level 3 vehicles in the mainstream.
Five years ago, we had excluded autonomous vehicles from our list of “tech” to look out for the second half of this decade in our Technology Futures research for Ofcom for the reason we now know today.
We did include remotely controlled, autonomous robots as a potential network-dependent application type that could benefit from the promised features of 5G.
It turns out that many of the robotic deliver systems today are in remote controlled, not autonomous as are the drone that the Ukrainians are using to hit Russian targets with the Starlink satellite network.
AT&T makes first 5G IoT data call using new RedCap technology
by Linda Hardesty to Fierce Wireless
August 3, 2023
“AT&T says it conducted the first 5G Reduced Capability (RedCap) data call in both its lab and in the field on a live 5G standalone network. The carrier says RedCap technology holds a lot of promise for the burgeoning internet of things (IoT) ecosystem.
While it seems kind of amusing to brag about “reduced capability,” RedCap fits between the high capability of 5G networks and the low capability of IoT networks that provide connection for devices such as simple sensors. RedCap is sometimes referred to as “New Radio Light,” according to a blog post today by Jason Sikes, AVP of Device Architecture at AT&T.”
I’m not sure what an “IoT” data call is, but the use of RedCap in communications is interesting for the possibility of standardizing voice services for wearable devices such as standalone smartwatches on 5G networks.
Today, devices such as the Apple Watch come in cellular versions that connect via LTE. RedCap or 5G NR Light is an evolution of LTE Cat-1bis and Cat-4.
The demonstration highlights the possibility of more economical devices such as a new generate of “feature” smartphones having voice services.
Bosch, Infineon, Nordic, NXP, Qualcomm drive open-source RISC-V hardware
by James Blackman of RCR Wireless
August 4, 2023
“Heavyweights in the semiconductor industry have joined together to create a new company to accelerate production of products based on the new open-source RISC-V (RISC-five; reduced instruction set computer-V) architecture. Robert Bosch, Infineon Technologies, Nordic Semiconductor, NXP Semiconductor, and Qualcomm Technologies will jointly invest in the new venture, which is headquartered in Germany. The initial focus will be on the automotive sector, but this will expand to include mobile and IoT, they said.
RISC-V is an open-source instruction set architecture (ISA) based on established reduced instruction set computer (RISC) principles. Unlike most other ISA designs, it is provided under royalty-free open-source licenses. Open source operating systems with RISC-V support are available, and the instruction set is supported in several popular software toolchains. A number of companies are offering RISC-V hardware.”
This move by some of Arm’s largest customers (licensees) is an apparent hedge against what feels like an uncertain commercial regime going forward with arguably the most important IP core company in the world – Arm.
Qualcomm is currently in litigation with Arm over its acquisition of Nuvia, a chip design firm started by ex-Apple and Google engineers, and Qualcomm’s application of contested IP. Given allegations by Qualcomm that Arm will change its licensing model makes this partnership a good hedge.
The combined market power and technical prowess of this newly minted consortium will accelerate the adoption of RISC-V and ecosystem expansion eventually bringing about a viable threat to Arm and its dominance in mobile.
Companies like Ventana are developing server-class CPUs based on RISC-V ISA to challenge Arm’s data center ambitions.
“Apple (AAPL.O) and Samsung Electronics (005930.KS) will invest in SoftBank Group (9984.T)-owned chip designer Arm at its initial public offering (IPO), expected in September, Japan’s Nikkei newspaper reported on Tuesday.
Reuters reported in June that Arm was in talks with some ten companies – including Apple, Samsung and Intel (INTC.O) – with the aim of bringing on one or more anchor investors in the offering.
Last month, Reuters and other media reported that Arm was in talks to bring in U.S. chip designer Nvidia (NVDA.O) as an anchor investor for the New York listing.
Apple, Samsung, Nvidia and Intel all plan to invest in Arm as soon as it is listed on the market, the Nikkei said. The SoftBank-owned firm will officially apply to the U.S. Securities and Exchange Commission for the listing later this month, the newspaper said.“
It didn’t take long for Arm to muster some of its largest and most strategic licensees to ante into their IPO pot. With the addition of Apple and Samsung to its flock of anchor investors, it appears that Arm is getting a nod of support for not only its IPO but the future of its ecosystem.
Not surprisingly, this news comes shortly after a consortium of Arm’s other strategic customers, most notably Qualcomm and NXP, announced their intent to join forces in accelerating the evolution of RISC-V as an alternative to Arm.
It’s clear, Arm ecosystem is becoming bifurcated. On one hand, you have solid Arm allies, many who have invested in developing their own custom chips based on IP such as Apple, Google, and Nvidia. On the other, chip designers who see risk ahead in their future with an emboldened post-IPO Arm.
Nvidia brings HBM3e to GH200 Superchip to tackle generative AI
by Dan O’Shea of Fierce Electronics
August 10, 2023
“In less than a year, generative AI has become a dominant influence in enterprise computing, so processor innovations need to move quickly, too. Less than three months after announcing its new GH200 Superchip, basis for its DGX GH200 supercomputing system, Nvidia already is giving the GH200 “a boost,” according to Nvidia co-founder, president, and CEO Jensen Huang, who at SIGGRAPH 2023 this week unveiled a “next-generation” version of the chip with more robust memory capabilities.
Many of these developments driven by ever-larger AI models, are going to require ever-greater memory performance.
While the Nvidia GH200 Grace Hopper Superchip announced back in May at Computex and now in production has an HBM3 memory processor, the latest version, due to be in production in the second quarter of 2024, has the faster HBM3e processor. Huang said in a keynote speech at SIGGRAPH, ‘We’re gonna give this processor a boost with the world’s fastest memory called HBM3e.'”
GPUs are not the sole ingredient to the generation of AI computing. High bandwidth memory (HBM) is also key to the supercomputing systems for LLM (Large Language Model) training and inference.
As the leading AI systems company, Nvidia continues to transcend the GPU and has clearly gotten into the business of engineering modular supercomputing blocks. The AI system business is clearly becoming a larger portion of Nvidia’s data center revenues and the largest driver of growth based on Nvidia’s Q2 24′ earnings call.
In retrospect, Project Grace proved to be a key investment by Nvidia into their data center and AI future. As it turns out, the Grace Hopper superchip comprised of the Grace CPU and H100 GPUs coupled with HBM memory is proving instrumental to Nvidia’s current “AI market” dominance in the data center. As Nvidia ships more of HGX AI HPC systems, the HMB3e will go with them.
Cloud Infrastructure Services Market Size & Share Analysis – Growth Trends & Forecasts (2023 – 2028)
by ReportLinker of Yahoo! Finance
August 10, 2023
“Reportlinker.com announces the release of the report “Cloud Infrastructure Services Market Size & Share Analysis – Growth Trends & Forecasts (2023 – 2028)”
The Cloud Infrastructure Services Market size is expected to grow from USD 72.69 billion in 2023 to USD 100.15 billion by 2028, at a CAGR of 6.62% during the forecast period (2023-2028).
The primary growth drivers for the market include low costs, scalability, flexibility, and security. The cloud infrastructure service offerings accelerate Time-to-Market (TTM) and rapid application development and running processes. Moreover, the expanding need to decrease the operational costs and maintenance of the IT infrastructure also boosts several organizations’ adoption of cloud infrastructure services.”
The cloud market has always been a difficult one to segment, size, and understand. The definition of the cloud market continues to morph especially in this era of hybrid cloud.
The argumentation for primary growth drivers of the cloud market runs counter to what we have observed in multi-hybrid cloud circle. There is a growing realization by IT buyers that cloud infrastructure and platform services are costly and have resulted in what many in the industry have dubbed “cloud repatriation”.
Indeed, large cloud players such as AWS have admitted to working with their customers to streamline their spend as an investment in customer trust. The outstanding question is if buyer have gotten wiser to the cloud economics.
Biden signs order to stop new US investments in China in chips, quantum and AI
by Matt Hamblen of Fierce Electronics
August 10, 2023
“President Biden signed an executive order Wednesday to restrict or prohibit some new US investments in China in sensitive advanced technologies including semiconductors, quantum and some AI systems.
While the impact is far from clear, analysts said it appears it could have a greater impact on the world’s biggest GPU maker, Nvidia, than other US companies, although Nvidia and GPU competitors Intel and AMD did not comment immediately.
A 45-day comment period on the order is expected to generate interest by a wide range of tech investors and players, since the order is presumably forward-looking. However, the text of the order still has not been entered into the Federal Register, and the clock on the comment period has not even started.”
This executive order is another in an ongoing series of “de-risking” measures against China. These actions will clearly be perceived by Beijing as further escalation by the Biden Administration of what it considers efforts to decouple from China, which based on rhetoric on both sides of the U.S. political divide is a matter of rare consensus.
Nvidia’s Q2 2024 China revenue increased $1.15 billion or 72% sequentially suggesting increased data center purchasing by Chinese hyperscalers against a rising tide of sanction risk.
We saw significant stockpiling of chips by Chinese firms, notably Huawei, with the initial onset of U.S. tech sanctions. A recurrence of stockpiling will color the demand outlook for GPUs in the future.
“Huawei said its net profit margin was 15% during the period, up from 5% in H1 2022
Chinese vendor Huawei recorded revenues of CNY310.9 billion ($43.1 billion) in the first half of the year, an increase of 3.1% compared to the same period the previous year, the company said in a statement.
In the first half of 2022, Huawei’s revenues had declined by 5.9% while revenues in the first half of 2021 had declined by 29.4%
Huawei said it experienced growth in China and overseas during the period.
Huawei also informed that its net profit margin was 15% during the period, up from 5% in H1 2022. The vendor did not disclose its net profit for the period.
The company’s ICT infrastructure business contributed CNY167.2 billion, while its consumer business CNY103.5 billion. Meanwhile, Huawei’s cloud business recorded revenues of CNY24.1 billion during H1 2023, while its digital power business generated revenues of CNY24.2 billion. Finally, Huawei’s intelligent automotive solution (IAS) business recorded revenues of CNY1 billion.”
Despite years of suffering an expansive campaign of sanctions by the U.S. government, Huawei continues to demonstrate tremendous resilience as neXt Curve has admonished for several years as it transforms into a “solution integrator” leveraging its formidable chops in CT, IT, and OT.
Consequently, Huawei has engineered out Western components from its products as U.S. government sanctions on the company ramped up back in 2019. The company is sourcing with local Chinese suppliers.
Huawei continues to lead the market in mobile network equipment with 50% and 35 to 40% revenue share in and outside of China respectively according to the Dell’Oro Group‘s Stefan Pongratz.
To the chagrin of the U.S. government, Huawei is growing its enterprise revenues and reconstituting its consumer group despite being banned from 5G modem-RF systems from Qualcomm, MediaTek and other Western vendors, and leading-edge manufacturing from TSMC for their in-house SoCs by HiSilicon.
Chart of the Month - eSIM SGP.32
neXt Curve is proud to announce our strategic partnership with Transforma Insights, the leading digital transformation market intelligence and advisory firm who have contributed the Chart of the Month for August 2023.
Analysis by Matt Hatton, Co-founder of Transforma Insights
The arrival of eSIM technology promises slimmed down devices, greater power efficiency and a more seamless mechanism for managing cellular connectivity. The latest addition, so-called “SGP.32”, arrived in May 2023 and will dominate the landscape, with significant implications for how connectivity is delivered.
Until 2016, cellular connected devices were authenticated using a removable plastic SIM card. This wasn’t appropriate for many IoT use cases, for which a more ruggedised chip, or eSIM, to be soldered onto the device circuit board. This further evolved with the advent of iSIM in 2018, which moves the SIM onto another processor as a virtual element.
As a result of this change in the physical form factor, it was necessary to develop eSIM Remote SIM Provisioning (RSP) to remotely switch profiles on the SIM card without needing to access it physically.
The first two standards for eSIM architecture were developed by the GSM Association as SGP.02 (“M2M”) in 2014 and SGP.22 (“Consumer”) in 2016. Technical specifications of a third variant, SGP.32 (“IoT”), were finalised by the GSMA in May 2023, with devices due in the second half of 2024.
Today around 20% of shipped cellular devices are RSP-capable, a figure dominated by non-standard approaches. By 2032, that figure will be 45%, of which most will use the SGP.32 IoT variant.
However, we also note that many mobile network operators have expressed concerns that SGP.32 represents a significant loss of control for them over managing the customer connection. If every connection is ‘up-for-grabs’ at any time by way of the enterprise customer changing operator profiles, we will see increasing competition to secure contracts for existing deployments, typically through aggressive price discounting.
During 2023, Transforma Insights will continue to provide support for adopters and vendors on understanding the implications and likely dynamics in the use of eSIM/RSP.
If you are interested in finding out more about the neXt Curve + Transforma Insights partnership, check out the press release announcing our partnership to bring Essential Validated Insights to our clients and reach out to Leonard Lee.
reThink Insights
Check out the articles and the research notes that neXt Curve published this month as well as press quotes by the media on topics related to our research agenda.
- Tareq Amin departs from Rakuten (August 7, 2023) by Linda Hardesty featured in Fierce Wireless
- How video comms could become a critical part of network slicing (August 8, 2023) by Dan Jones featured in Silverlinings
- Managing & Scaling Edge Infrastructure (August 11, 2023) by Leonard Lee featured in neXt Curve
- China-US tech fight figures in scrapped Intel deal to buy Tower Semi (August 16, 2023) by Matt Hamblen featured in Fierce Electronics
- How much longer will Nvidia’s AI winning streak last? (August 24, 2023) by Dan O’Shea featured in Fierce Electronics
Go to our neXt Curve reThink research portal for more content and insights associated with our research agenda.
Orchestrating & Managing The "Edge"
In our recent engagements with VMware at VMware Explore and a number of Red Hat analyst briefings this year, workload and resource orchestration and management are important areas of develop that will be essential in driving the evolution of edge cloud infrastructures.
One of the key challenge areas of edge infrastructure management is fleet management especially given the growing diversity and intersects of “edge” environments. We are seeing emerging edge software vendors right sizing their fleet management offerings to cater the nuances of various edges from central offices to shop floors or campuses.
With cloud-native technologies being introduced across edge infrastructures and into what you might call edge devices such as a vehicle (think software defined vehicles or SDV), the lines between fleet management and endpoint or edge device management are becoming blurred sharing similar aspirational features such as zero-touch provisioning, configuration management, observability, remote administration and firmware/system updates.
One of the tricky things for end users will be dealing different levels of edge infrastructure fleet management and their respective edge device management requirements especially when instituting zero trust security principles into their edge environments.
Speaking of security and trust, trust frameworks and confidential computing paradigms are becoming key considerations as enterprises look to deploy their data and applications across the edge. Technical leaders should consider a processor-to-application holistic view on root-of-trust and the chain of provenance of devices in the portfolio and across edge nodes.
The evolution of this edge orchestration and management space in the next couple of years will be a determinant factor in how quickly enterprises adopt new edge computing models enabled by edge-native infrastructures.
Keep an eye out for developments in vRAN as it transitions to cloud RAN. This is where a lot of the edge orchestration and management innovation is happening at the moment.
Pretty exciting stuff!
Media Highlights
This month, neXt Curve participated in the following internally produced and third-party media events. More media content featured by or featuring neXt Curve is available on our reThink YouTube channel and our media center.
The State & Future of Open RAN with Patrick Lopez
Strategies for Scaling Edge Compute Companies with 5G, IoT, AI, etc.
Leonard is joined by special guest, Patrick Lopez, CEO and founder of {Core Analysis}, who advises some of the leading operators and vendors in the telecommunications industry. In this episode, Patrick and Leonard talk about the state and future of Open RAN.
Ever wondered how to distinguish the beneficial from the redundant in an over-hyped tech world? Leonard shares practical strategies, emphasizing the importance of understanding the entire technology stack from semiconductors to applications.
Event Highlights
This month, neXt Curve participated in the following virtual and in-person industry and technical events. For our full schedule of industry events refer to our event calendar. We also encourage you to follow neXt Curve’s LinkedIn company page.
Mountain Connect 2023
Date: August 7 to 9, 2023
Location: Denver, CO
Event Summary & Takes
neXt Curve was invited to attend Mountain Connect taking place this year in Denver. In years past, this event has been held in remote resorts more than a stone’s throw distance from the center of the city. This year, the event took place in at the Downtown Sheraton Hotel.
Mountain Connect is North America’s largest independent broadband conference that focuses on supporting the expansion of broadband infrastructure and services across the United States. This year marks an important moment for the conference that drew nearly 900 attendees and a wide range of vendor exhibitors with the recent allocation of $42.4 billion in BEAD (Broadband Equity and Deployment) funding to states administered by the NTIA (National Telecommunications and Information Administration.
neXt Curve’s Leonard Lee was invited to the conference by Mountain Connect CEO, Jeff Gavlinsky. The big theme for this year was what is next for the BEAD program? We wanted to get the nuances of the BEAD application, planning, and proposal processes that States and applicants need to go through to start applying for billions in Federal funding to bridge digital divides and bring digital equity to unserved and underserved communities and areas across the U.S.
Leonard also moderated a panel joining fellow IoT Coffee Talk co-hosts Rob Tiffany and Bill Pugh on stage to talk about the equity challenges beyond the connectivity that the BEAD-funded infrastructure buildouts will bring. Check it out here (link).
Here are our key takes from Mountain Connect 2023 event.
- 5G was not the talk of the town. Fiber is king with all of its deployment challenges and cost. We were surprised to see wireless, in particular cellular, treated as a worst-case scenario if at all.
- Are 100 mbps downlink and 20 mbps up sufficient to foster equity in a multi-Gigabit world?
- BEAD is different from prior efforts to bridge the digital divide. Its mandate is to provide “Internet for all”, meaning the funds are intended to connect every household in the U.S. The applicant matching for funds to get BEAD funding can be a challenging hurdle for many communities. The funds also come with stringent terms and risk of clawback.
- Mapping data is essential to the entire lifecycle of the BEAD program from applications, fund allocation, deployment planning, through to assurance. Both fiber and wireless options for connectivity come with their unique challenges. In this regard, Fabric is the FCC map based on Hexagon that strives to provide the essential data that will guide the targeting of funds to the qualification of services quality in a funded location.
- Longevity is key for the new BEAD infrastructure. How do you sustain the infrastructure over 30 to 40 years, many in remote and difficult to maintain areas, once the implementation is complete? Where does the funding come from? This are some big open questions that remain open for service providers, communities, and the U.S. government.
- While the funds are there, it is difficult to determine if they are sufficient to meet the objectives of BEAD. Wireless options such as 5G FWA and satellite may become more prominent discussion items in the months to come.
Related Media & Press Releases
- Mountain Connect 2023 Event site (link)
- LinkedIn: neXt Curve August 2023 Event Agenda (link)
- LinkedIn: On The Way to Mountain Connect (link)
- LinkedIn: Mountain Connect 2023 Banner (link)
- LinkedIn: Hanging Out With Jeff, Rob, and Bill (link)
- LinkedIn: IoT Coffee Talk at Mountain Connect 2023 (link)
Companies Engaged: Nokia, Google Fiber, Calix, Logic Networks, Ciena, Arista Networks, CommScope, Hexatronic U.S., Biarri Networks, Siklu Communications Ltd., Zyxel Communications, Ready.net, Ribbon Communications
VMware Explore 2023
Date: August 21 to 24, 2023
Location: Las Vegas, NV
Event Summary & Takes
After years of virtual attendance thanks to the pandemic and schedule conflicts, neXt Curve finally made it to VMware’s Explore event, formerly VMworld, to partake in one of the most intense analyst programs in the industry. Our schedule was packed with executive sessions, analyst one-on-ones, and networking events.
This year, one of the big themes for VMware Explore was what they call “multi-cloud”. In VMware’s vernacular, multi cloud means both multi-vendor cloud and hybrid cloud and a vision that expands the cloud concept beyond the hyperscaler’s data centers across the galaxy of enterprise and industrial “edge” environments.
The bulk of the Day 1 agenda was filled with an industry analyst summit and breakout sessions that peppered the afternoon. Raghu and Paul Fazzone, SVP of Strategy who presented the six key announcements and stated that the acquisition of VMware by Broadcom is expected to close on the 30th of October with the transaction receiving U.K. regulatory approval a day before the VMware Explore event.
The other inevitable topic du jour was generative AI. Unsurprisingly, Jensen Huang, CEO of Nvidia, made a cameo appearance during VMware CEO Raghu Raghavan’s keynote touting a new partnership between the two companies called VMware Private AI Foundation with Nvidia.
With the advent of ChatGPT, it was refreshing to see VMware emphasize a privacy first approach to enterprise generative AI system enablement leveraging Nvidia’s generative AI frameworks such as NeMo on top of VMware’s Cloud Foundation. VMware told us that they want to provide enterprises with the flexibility that the cloud players don’t or can’t provide to source data, train models and inference across a “hybrid AI” infrastructure.
VMware also introduced three generative AI applications dubbed Intelligent Assist for Tanzu, NSX+ and Workspace ONE. Intelligent Assist agents are similar to what Google and Microsoft have announced with Duet AI and CoPilots respectively. Largely characterized as a beta offering, we will have to wait until next year to see how well this offering attaches.
On the edge front, VMware announced their VMware Edge Cloud Orchestrator (VECO) for fleet management as a new and essential layer in their VMware Edge Compute Stack that they assert will bring about the “software-defined edge”.
VECO is built on Project Keswick which provides a trusted framework for managing and orchestrating workloads and resource configurations across edge infrastructure nodes leveraging an open-source framework, hardware-level trusted execution environments (TEE), encrypted virtual machine memory, and single operation model across chipsets.
There was a ton of stuff that we gleaned from the three-day event. Contact us for an analyst briefing for more details.
Going into VMware Explore, our primary research points of interest included:
- Emerging edge-native architectures
- Edge cloud management and orchestration
- Edge cloud security and trust
- Telco cloud and cloud RAN
- Edge AI and accelerated edge infrastructure
- Hybrid cloud and neutral host edge cloud
- SASE (Secure Access Service Edge)
Here are our key takes from VMware Explore 2023 event.
- VMware is setting a strong counterpoint to the “cloud” with its multi-cloud narrative. While not precluding cloud (meaning hyperscalers), VMware is expanding its portfolio of edge cloud products and services on the back of the work they have done in virtual RAN (Radio Access Network) with DISH Networks. As VMware evolves its Edge Compute Stack it will be well positioned to modernize edge infrastructures thus play an important role in the future of hybrid cloud computing.
- We love the idea of “Cloud Smart”. While this might seem like something new, it is something that neXt Curve has been advocating and advising our clients for six years. Cloud first was never a smart strategy. As an ambitious hybrid cloud leader, VMware could very well provide the tools and the tech enterprises need to make their IT and OT infrastructures edge-native and hybrid cloud enabling.
- VMware is making a big push toward SaaS largely for the predictable subscription-based revenue. They seem to be taking an asset-light approach (meaning not hosting the infrastructure) for their SaaS offerings. Three-fourths of their revenue is from subscriptions today.
- Generative AI, while a key talking/selling point of the VMware Explore event, didn’t seem fully baked. Like so many other companies in “tech” VMware seemed pressured into having a Gen AI story to tell. As we wrote about early this year, VMware is adopting a confidential AI approach for their enterprise customers. It will be interesting to see how the generative AI hype plays out for the company.
- I made this as a joking comment, but it ends up being true. VMware is becoming the middleware company for the edge cloud.
Related Media & Press Releases
- VMware Explore 2023 Event site (link)
- LinkedIn: VMware Explore 2023 Event Banner (link)
- LinkedIn: VMware Explore 2023 Day 0 (link)
- LinkedIn: VMware Explore 2023 Day 2 (link)
- LinkedIn: Jensen Huang cameo at VMware Explore 2023 (link)
- LinkedIn: Hanging with Evan and Sean at VMware Explore 2023 (link)
- LinkedIn: It’s a Wrap at VMware Explore 2023 (link)
Companies Engaged: VMware, Lenovo, IBM, Hitachi Vantara, Intel, AMD, Dell Technologies, Veeam, Nvidia, Tech Mahindra, Wipro
Google Cloud Next 2023
Date: August 29 to 31, 2023
Location: San Francisco, CA
Event Summary & Takes
Google Cloud Next 2023 was held in-person this year but neXt Curve attended virtually. Last year, the event has help in a hybrid “24 hours of global engagement” format across five locations in New York City, Sunnyvale, Tokyo, Bengaluru and Munich. Needless to say, Google does a great job virtual and hybrid events, but with this year’s big topic of generative AI, attending virtual did feel a bit distant.
We did get a very good impression of what was top of mind for Google Cloud as Sundar Pinchai, CEO of Alphabet, and Thomas Kiran, CEO of Google Cloud, which was clearly overweighted toward generative AI. No surprise given the fanfare the company has corroborated in the so-called “AI war” that has erupted since the advent of ChatGPT.
As expected, Jensen Huang, CEO of Nvidia, made a cameo appearance on what seems to be a world tour. Jensen is doing his best to be at every industry and partner event to tout the future of AI that he and his company believes generative AI. Google Cloud acknowledged that they would support both Nvidia’s GPU-based solutions as well as their own proprietary TPU offerings.
The matter top of mind for neXt Curve is responsible AI which we view as the umbrella these that encapsulates ethical AI, confidential AI, sustainable AI, private AI, and secure AI. These themes contribute to what neXt Curve views as safe AI which requires trustworthy engagement by technology firms, enterprises, organizations, and consumers/citizens in the use of AI applications.
Here are our key takes from Google Cloud Next 2023 event.
- Google Cloud Next felt more about generative AI (very little about other varieties of AI applications and technologies) than it was about “cloud”. Google Cloud’s SaaS portfolio seemed to take the fore with Duet AI-infused Google Workspace. Duet is Google’s Gen AI plug-in.
- Google, much like Microsoft is looking to take up the open AI banner and support as many Gen AI models and frameworks as available. It’s like a race to see who is the most open which I don’t think really matters. Much of this is ecosystem posturing.
- Google BigQuery Object Tables are super cool and exemplify some highly practical and valuable application of ML techniques applications to contextualize existing unstructured data sets and make them machine and human readable. Google continues to advance the capabilities with generative AI augmentation with Vertex AI integrations for both data contextualization and NLU interfaces.
- Many of the Gen AI features seemed incremental in value as well as well scripted as is typical with most. As neXt Curve has observed with many Gen AI demos behind the scenes, lack of reliability and determinism are the pervasive deficiencies of generative AI applications.
- The AI monetization opportunity for Google outside of providing compute capacity for LLM model development and operational inference is not clear. There is little doubt that Google will attempt to charge users for Duet AI on Google Workspace, but will this be a way of easing a subscription rate hike in or a material generative AI uplift based on realized productivity improvements. The jury is still out.
- Despite some follow ups with Google, there are significant concerns about how and if Google is protecting enterprise and consumer confidentiality and privacy. The simple answer on the enterprise side it boils down to private LLMs in VPCs. For SMBs and consumers, things are a bit more complex and unanswered and what Google means by personal data and confidentiality in its user agreements. Where and how does the data flow? Does Google offer a virtual private LLM for consumers for its public SaaS offerings infused with Gen AI features?
Related Media & Press Releases
Companies Engaged: Google, Nvidia
Upcoming Industry & Media Events
Next month, neXt Curve will participate in the following virtual and in-person industry and technical events. For our full schedule of industry events refer to our event calendar. We also encourage you to follow neXt Curve’s LinkedIn company page.
Subscribe to the neXt Curve Insights monthly newsletter to be notified when the next newsletter is published. Go to www.next-curve.com to be added to our mailing list. You will also be notified when we publish new research notes and media content.