Cisco has revealed its intent to acquire Working Group Two (WG2), a Norwegian company that is said to have pioneered a cloud-native mobile services platform that’s fully application programming interface (API)-consumable and highly programmable.

Emerging from a Telenor laboratory in 2013, WG2 began after being frustrated with the inability to use mobile networks to conduct innovation, prompting a rebuild of the mobile network for a new age. The overarching goal of the firm was to create a “radically innovative, radically simple and radically affordable” modern mobile core, something the company said was a key component in advancing the mobile industry.

WG2 said there’s at least one reason why mobile users haven’t seen much product innovation coming out of their telecom operator over the past few decades. It claimed mobile network machinery is “not just a beast, it’s also a beauty”, adding that mobile telephony is perhaps the most successful, distributed and standardised technology the world has ever seen. The big idea behind WG2 was to “bring out the beauty in the mobile networks and revitalise the entire industry”, making connectivity truly valuable again.

Backed by Telenor Group and Digital Alpha, a fund with a long-standing, strategic Cisco partnership, WG2 was spun out as an independent entity from Telenor in 2017. Today, the company comprises a team of 90 dedicated professionals, primarily software engineers based in Europe, Japan and the United States. In 2022, it helped customers deliver 527 million SMSs, make 354 million voice calls and transfer 33 petabytes of data.

To grow and prosper, WG2 said the telecom industry must strive for simplicity, global scale and programmability, and that it must also nurture marketplaces and ecosystems of developers to build global platforms, just like the tech industry at large.

Putting the acquisition into context, WG2 noted that Cisco will build on the as-a-service model to realise benefits for the full mobile industry, having already seen success with internet of things (IoT) and private networks. It said the as-a-service delivery model allows for consistent deployments, reduced complexity, global scale, programmability, and continuous innovation and security management.

By expanding this model to cover the entire spectrum of mobile use cases, WG2 was confident Cisco could provide simplified access to the company’s technology while enabling customers to focus resources on their core business and deliver new use cases faster.

For its part, Cisco said millions to billions to trillions of connected devices coming online was making IoT the single greatest digital business transformation that most of us will see in our lifetime. It said communication service providers must scale to support it, which means connectivity must be simplified for both deployment and management, and the mobile business service must evolve to become a much more lucrative offering.

Cisco noted that WG2’s platform uses the web-scale playbook and operating models, which makes it a natural fit with its Mobility Services Platform.

“WG2’s technology and team beautifully align with the same approach: simplifying the mobile network architecture to deliver a radically innovative mobile service,” said Masum Mir, senior vice-president and general manager at Cisco Provider Mobility, in a blog post. And with WG2 and the Cisco Mobility Services Platform, we’ll be able to boost our service edge deployment and API-first strategy for application development partners, enterprise customers and service provider partners.

“WG2 and Cisco have a unique synergy,” he said. “We share a common ambition to deliver a global, programmable mobile core as a service, and are remarkably aligned to this shared vision.

“The goal? To enable our Communication Service Provider customers to cover ground faster, as we work together to monetise their 5G investments,” said Mir. “The consumption model is simple and comprehensive, supporting all stakeholders. The end result is a mobility services platform that can dramatically simplify mobile network deployments, provide enhanced edge experiences, enable new and advanced use cases, as well as support simple application development.”

Copyright © 2023 IDG Communications, Inc.

Is the inevitable backlash against generative artificial intelligence at hand? Some early signals suggest the possibility, but you’d never know it from all the fundings and new products streaming out this week, as SiliconANGLE documented in a raft of stories this week.

We also covered the better-than-expected quarterly earnings from Cisco Systems Inc., which gave investors hope that a tech spending lull might be easing. Meantime, though, some cybersecurity companies are struggling as generative AI could be siphoning off information technology investments. Crypto isn’t dead yet despite the ongoing wintry conditions. And Intel Corp. dropped its bid for Tower Semiconductor after China’s regulators essentially nixed it — another in a string of bad news for the struggling chipmaking giant — even as Arm Ltd.’s IPO moves ahead.

For an insightful and entertaining take on this and other news in enterprise and emerging tech, check out this week’s theCUBE Pod, John Furrier’s and Dave Vellante’s weekly podcast. And this weekend, look for Vellante’s Breaking Analysis, his weekly deep dive on an enterprise tech trend.

I’ll kick things off this week with a look at the current state of the cybersecurity industry through the lens of one of its top leaders, Palo Alto Networks Inc., which announced earnings today.

As cybersecurity hits a slower patch, Palo Alto Networks charts plan to stay ahead

Palo Alto held a highly unusual presentation on its earnings and a review of its strategy today — unusual in that Friday announcements are usually reserved for bad news. Not so today, since in Palo Alto’s case, the news was plenty good. (As Chief Executive Nikesh Arora apologetically explained on the call, which was kicked off with a remix of Rebecca Black’s 2011 earworm “Friday,” he wanted to leave time to talk one-on-one with analysts over the weekend before a planned company sales meeting that starts Sunday, on top of a board meeting this week.)

Anyway, the company reported its fiscal fourth-quarter profit before certain costs such as stock compensation jumped 90% from a year ago, to $482.5 million, or $1.44 a share. Net profit hit $227.7 million, or 64 cents a share. Revenue rose 26%, to $2 billion.

The outlook was positive as well. For its fiscal first quarter, the company expects adjusted earnings of $1.15 to $1.17 a share, up 40% at the midpoint, on a 16% to 18% rise in revenue, to as much as $1.85 billion. For the full year, the revenue forecast is a tick higher, between 18% and 19%, with adjusted earnings up 19% to 22%.

Nikesh Arora

Investors liked the results. Palo Alto’s shares were rising more than 8% in after-hours trading. Shares were already up 80% on the year through the start of August.

“The report is better than feared,” Ivana Delevska, founder and chief investment officer of investment adviser Spear Invest, told me. “Guidance is light, but given the timing of the report many investors were expecting something much worse like an accounting restatement, or management change.”

Arora touted the “changing environment” that drove more customers toward “platformization,” meaning customers that buy multiple product lines. To that end, Arora said the company was surprised by the strength of its extended security intelligence and automation management, or XSIAM, product, with bookings of more than $200 million in its first year.

XSIAM, in its Cortex line, combines endpoint detection and response or EDR, security orchestration automation and response or SOAR, attack surface management or ASM, and security information and event management or SIEM technologies into a single solution. Many other companies offer these separately, which can be a pain point for customers that often must juggle many different cybersecurity products from different companies.

Despite the positive results, it’s no easy sledding these days, given high interest rates that are tamping down spending. “There is more scrutiny” on deals, Arora said. “There are some that get stretched or get canceled.”

During a 90-minute presentation, Arora dug into the evolution of what he says is a $213 billion cybersecurity market. There are new segments such as SASE, cloud security and internet of things security that contribute $29 billion, transforming segments such as endpoint and XDR as well as SIEM and network security at a collective $72 billion, steady segments such as identity, app security, data and email security at a total of $31 billion, and $81 billion in services.

“We were told customers don’t want platforms, they wanted best-of-breed solutions,” Arora noted. Instead, he said, Palo Alto is aiming to do both, through what the company calls a “build and buy” strategy, to become the largest pure-play cybersecurity provider. Cisco and Microsoft are larger, but of course cyber is just part of their businesses.

The upshot of all this is that Palo Alto looks to continue as a consolidator in this industry, along with a few others such as CrowdStrike Holdings Inc., Check Point Software Inc. and Cisco Systems Inc. The industry does seem to be splitting between larger companies continuing to grow and roll up smaller companies and others struggling to maintain traction either because of aging technology or because IT departments are diverting more spending toward AI. Just in the past week or so, Rapid7 and Secureworks laid off workers. And Arora noted that there are 3,000 cyber startups out there — clearly unsustainable.

“We believe M&A will pick up significantly in the second half and 2024 and companies like Palo Alto that have the capital availability are in a solid position to benefit,” Delevska said. “We see ‘shift left’ as a major theme for M&A, that is, cybersecurity embedded earlier in the software development cycle.”

She also thinks there will be only a limited number of industry consolidators, in particular Palo Alto and CrowdStrike. “M&A has been hit-and-miss in cybersecurity and therefore track record is key,” she said.

Going forward, Arora sees a need for, and shift to, more real-time and autonomous security. “We will see a standard platform for security,” he vowed. “That’s the only way we’re going to get to the future we need for real-time and AI-based security.”

More cyber news below.

Around the AI universe

A well-known AI skeptic raises doubts about generative AI that are worth memorizing even if the essential critique is, as often with Gary Marcus, points to inflated expectations more than a complete lack of utility: What if Generative AI turned out to be a Dud? Also, Benedict Evans is unimpressed so far, though it’s a bit mystifying he can’t find any good use for it. Plus, gen AI has landed right at the top of Gartner’s Hype Cycle and you know what that means.

Axios talks to larger companies struggling to implement generative AI.

Some observers even think the AI startup Gold Rush is ending: A.I. startups are losing their bloom for seed investors, argues one VC

But that seems slightly early, many startups keep on coming: Former Google researchers launch startup to build nature-inspired neural networks and Elemental Cognition, led by IBM Watson’s former head, raises $60M

It’s getting competitive on the gen AI infrastructure front: Continuing AI investment is why the GPU battle, for one, keeps intensifying — and the software matters as much as the hardware, which is why Modular could raise so much money: AI software startup Modular seeks bumper Series A round to challenge Nvidia. Of course, the big guys such as Amazon are already well into this chips-and-software race: How Amazon is racing to catch Microsoft and Google in generative A.I. with custom AWS chips. Meantime, per the New York Times, the GPU shortage rages on.

And then there’s the endlessly fundraising Databricks: Databricks looking to raise ‘hundreds of millions’ in fresh funding to fuel generative AI push But Furrier notes in theCUBE Pod podcast that it’s not out of need for cash but striking while the AI iron is hot.

For better or — in the case of the Iowa school board that wants to ban books they can’t be bothered to read — for worse, harnessing chatbots for content moderation: OpenAI finds GPT-4 can Boost online platforms’ content moderation efforts

Even the big large language model creators are looking to provide industry-specific AI models: Anthropic raises $100M from SK Telecom to build AI for telecommunications Do they become AI superclouds? And in the same vein: Arthur launches open-source tool to help companies make data-driven decisions about LLMs

About the worry over those LLM data leaks — someone’s working on that: DynamoFL raises $15.1M to tackle language model data leaks

“Hey Google, what should I do?” Google DeepMind reportedly developing at least 21 new generative AI features Not scary at all, nope.

And the hardware suppliers see an opportunity to bundle things to make AI development easier, though it’s not clear how big a market there is for this kind of thing outside the big cloud providers: Nutanix offers quick-start approach to AI development

OpenAI makes its first acquisition: OpenAI acquires digital products company Global Illumination for undisclosed price

Amazon finds some low-hanging gen AI fruit: Amazon adds AI-powered review summaries to its e-commerce marketplace

More cyber ups and downs

Cybersecurity companies still cutting costs perhaps as AI steals some budget: Cybersecurity provider Secureworks to let go 15% of its workforce, on the heels of Rapid7 layoffs last week.

Strom looks at the latest CPU attacks and concludes they will be tough to fix quickly: Mitigating the latest processor attacks will be a chore on many levels

Same deal with phishing: New reports show phishing is on the rise — and getting more sophisticated

And another security issue in IoT (maybe take the stairs next time): New widespread IoT compromise could affect millions of logic controllers

Finally, here’s detailed advice on how to avoid security fatigue: How to prevent multifactor authentication fatigue attacks

Tech economy pulse

Cisco earnings beat estimates and its stock rises a bit after-hours. At first a seemingly weak forecast from the industry bellwether raised doubts about tech infrastructure spending, but subsequent company comments and analysis indicate a good quarter and equally good outlook after all: In a positive sign for tech spending, Cisco’s stock rises on strong earnings and prospects for AI

More China economic war fallout: Intel scraps its $5.4B acquisition of Tower Semiconductor Vellante and Furrier think Intel’s in deep trouble, as they describe on theCUBE Pod. Tower wasn’t a game-changer in itself, but it’s hard to see how CEO Pat Gelsinger can revive Intel’s fortunes anytime soon.

Arm’s IPO to get rolling next month: SoftBank reportedly buys back 25% Arm stake from its Vision Fund unit

Crypto ain’t dead yet: Despite crypto downturn, BitGo raises $100M at $1.75B valuation and ZetaChain raises $27M. Plus Coinbase secures approval to offer crypto futures to US customers. And Crypto hardware wallet maker Ledger teams up with PayPal

Random bits

That was quick: SUSE to be taken private by its majority shareholder

Eliminating the decryption tax for querying encrypted data: MongoDB unveils data encryption tech for developers to boosting data privacy and compliance

Supercomputers in the cloud: Harvard researchers clone supercomputer on Google Cloud

Sassine Ghazi to become CEO of chip design software firm Synopsys

What’s coming

Big changes in Europe for big tech: Doing business in Europe? Time to focus on its new Digital Services Act – now

VMware Explore runs Aug. 21-24, and SiliconANGLE and theCUBE will be covering it in Las Vegas.

Earnings from Zoom, Nvidia, Snowflake, NetApp, Splunk, Autodesk and Workday

Then the following week, Aug. 29-31, is Google Cloud Next, also to be covered onsite at Moscone Center in San Francisco by SiliconANGLE and theCUBE.

Image: Bing Image Creator; photo: SiliconANGLE

• Cisco CEO surprised analysts by mentioning it had already sold half a billion dollars of AI gear.
• Cisco is trying to woo cloud companies away from offerings by Nvidia.
• One analyst says it's a promising start on a potentially huge new market.

Cisco CEO Chuck Robbins dropped a bit of a bombshell during its earnings conference call Wednesday. That's when he told analysts the company had already sold over $500 million worth of equipment to the largest cloud computing companies for their AI needs.

The cloud companies (or hyperscalers as the tech world calls them) are buying from Cisco because they are experimenting with using Ethernet — the networking technology that Cisco is known for — for their AI needs instead of Infiniband. Infiniband is the high-speed networking technology more typically used by data centers, a market cornered by Nvidia when it bought leader Mellanox Technologies in a $6.9 billion in 2019.

So Cisco, along with AMD, Arista, Broadcom, HPE, Intel, Meta and Microsoft, formed a technology coalition last month called the Ultra Ethernet Consortium to work on making Ethernet twice as fast as the fastest 400 gigabits speeds available today, and tweaking the tech to better handle AI.

"To date, we have taken orders for over half a billion dollars for AI Ethernet fabrics. We are also piloting 800 gig capabilities for AI training fabrics," Robbins said.

True, half a billion is an insignificant amount of revenue for Cisco, which reported a record $57 billion in sales for its fiscal year 2023 on Wednesday and $3.07 earnings per share.

But Robbins says this new AI cloud market will be triple or more the size of the original and, this time, he plans to make sure Cisco gets its share.

"This would probably be three to four times the opportunity size of the original cloud build out. And, you know, unfortunately for us, as it's been well documented, we missed the original cloud build out," Robbins said.

He's referring to how cloud companies that run huge data centers like Google and Facebook built their own data center networking tech or chose data center specialists like Arista Networks.

"But I can say with every bit of confidence right now that as we go through this AI transition to Ethernet we are super well positioned," he said adding that "we now are installed in 21 use cases across the top six of these providers. And we expect that that momentum will just continue over the next few years."

Morgan Stanley analyst Meta Marshall is intrigued but cautiously skeptical. In a post-earnings research note, she pointed out that there's "an intense investor focus" on companies positioned for the AI infrastructure build-out and Cisco's investors want to hear its AI story. After all, at about a $225 billion market cap, Cisco's valuation is one-quarter's of Nvidia's $1 trillion.

While the $500 million sold so far is currently "small in terms of revenue, growing relevancy within this customer base is important, particularly given scale of opportunity," she wrote. "While we still think Cisco has smaller share within these hyperscalers, increasing penetration is a positive."

Are you Cisco employee with insight to share? Reach Julie Bort through email, jbort@insider.com, encrypted chat app Signal at 970-430-6112 or DM on Twitter @julie188 using a non-work device.