The UK’s antitrust regulator has concluded its investigation into Microsoft’s hiring of the majority of staff from Inflection and its licensing of the company’s technology.

The Competition and Markets Authority (CMA) published a summary of its decision Wednesday, finding that while Microsoft’s actions constituted a “relevant merger situation” and thus fall under its purview, they did not result in what it called “a realistic prospect of a substantial lessening of competition (SLC).”

In other words, the deal — which didn’t involve Microsoft buying the company — is a merger, but the regulator is OK with that.

This means that the CMA will not pursue a full-scale investigation into the deal, which poured an estimated US$650 million into Inflection’s coffers.

Not this time, anyway: Similar deals may, however, come under scrutiny for their effect on competition.

This was one of many regulatory looks at investments in AI startups by big tech companies hoping to escape regulatory scrutiny with what some have dubbed a quasi-merger: strategic investments and/or hiring key team members that gain the investor influence or control over the startup without actually buying the company.

At the same time the UK investigation into Microsoft was announced, the US Federal Trade Commission (FTC) began a look into Amazon’s hiring of key executives, including the CEO, from AI startup Adept, and its plan to license some of Adept’s technology. And in early August, the CMA announced that it is launching an inquiry into Amazon’s relationship with Anthropic to determine whether it, too, warrants a full investigation. A CMA inquiry into Google’s relationship with Anthropic is also underway.

When is a merger not a merger?

In its summary of the Microsoft-Inflection decision, the CMA said that that it assessed the criteria for a relevant merger situation under the Enterprise Act 2022, noting, “There is no particular combination of assets that constitutes an enterprise. As set out in the CMA’s guidance, it may include a group of employees and their know-how where this enables a particular business activity to be continued.”

Furthermore, it said, “In addition to hiring the core former Inflection team, Microsoft also acquired additional assets, including access to Inflection IP. The combination of acquiring the core team together with these assets was key to the value of the Transaction, as it enabled the former Inflection team to continue the pre-Transaction Inflection roadmap for consumer-facing AI product development within Microsoft.

“On this basis, the CMA believes that Microsoft has substantively acquired Inflection’s pre-Transaction FM [Foundation model] and chatbot development capabilities. Accordingly, the CMA has found that at least part of the activities of pre-Transaction Inflection has been brought under the control of Microsoft and, as a result, that two enterprises have ceased to be distinct such that the Transaction falls within the CMA’s merger control jurisdiction for review.”

The CMA said that the full text of its decision will be published “shortly” on the web page for the case.

Anthropic has launched the Claude Enterprise subscription plan, enabling businesses to securely leverage their own corporate data in their interactions with its Claude large language model. It’s a complement to Claude Work, Anthropic’s product aimed at small organizations, and a competitor for OpenAI’s ChatGPT Enterprise, released last year.

“The goal for us, for Claude Work and Claude for Enterprise, is really to enable and empower every team within an enterprise so that you can really become the most creative and most productive version of yourself,” said Nicholas Lin, Claude Enterprise product lead, in an interview.

Claude Enterprise features an expanded context window — 500,000 tokens, more than double the 200,000 previously offered — which Anthropic said is the equivalent of hundreds of sales transcripts, dozens of 100+ page documents, or 200,000 lines of code.

Dynamic workspaces

“Artifacts” — dynamic workspaces that, Lin said, let users “really see what’s going on in Claude’s head and to really iterate on outputs with Claude” — will assist users in creating data visualizations, documentation, presentations, and more.

“One thing I love to use artifacts for is a great way to just brainstorm with Claude, and using Claude to think about drawing diagrams and helping it visualize concepts,” he said. For example, a marketer could create an artifact in the form of a marketing calendar for a campaign, or generate content for the campaign, or a strategy document. In sales, Claude could analyze sales data, forecast trends, and generate collateral for a sales meeting.

The activity feed, he said, lets users draw inspiration from others in their organization. “Activity feed really enables you and others around you in the organization to share the most insightful pieces of feedback that you’re working with Claude on, so this is pieces of knowledge insights from your conversations with Claude in artifact outputs through the organization,” he said.

And, since manually uploading data is not sustainable at scale, the company is introducing native data integrations. The first, with GitHub, is now in public beta.

“We want to make sure that Claude is really well integrated into your everyday workflow,” Lin said. “This is the first of our native integrations. Many more will be coming in the coming months, and this is also the first of our software developer focused features. Many more will be also coming in the coming months.”

And, he promised, the uploaded data will not be used to train models.

Granular permissions

Anthropic says that, along with its AI features, Claude Enterprise contains enterprise-grade security controls including single sign-on (SSO) and domain capture, and role-based access with granular permissioning. Within a few weeks, Lin said, audit logs for security and compliance monitoring, and automated user provisioning and access control, known as the System for Cross-domain Identity Management (SCIM), will be available as well.

These features are long overdue, said Jeremy Roberts, senior research director at Info-Tech Research Group.

“It’s high time we got some general-purpose AI SaaS to compete with the likes of Microsoft Copilot,” he said. “When we think about new software, we focus a lot on its capabilities, but to be an enterprise solution, it must integrate nicely into the broader ecosystem. The announcements around SSO, RBAC, and audit logs are essential for this. Anyone worried about consumer technology in their businesses should be greatly heartened by the increasing competition in this space.”

Another analyst is curious whether the user experience will surpass those of other AI products.

“I’m excited to see the release of the Enterprise version of Anthropic,” said Terra Higginson, principal research director at Info-Tech Research Group. “Just like we saw in the search engine race of the early 2000s, the product with the best user experience and functionality dominated. Will Claude by Anthropic be the winner of the LLM race? Many of these systems are still offering subpar user experiences, and, to make matters even worse, the companies put a ton of restrictions that just make users lean towards private alternatives.”

Claude Enterprise is available today. Pricing was not announced; Lin said that each organization will be given a customized price based on its needs.

Intel’s plunging stock price, which as of noon New York time on Tuesday was the lowest it has been since 2010, could cost the chip giant its coveted spot on the Dow Jones Industrial Average (DJIA).

It comes at a very difficult time for Intel, as it is trying to maintain its enterprise relevance in the face of more effective generative artificial intelligence (genAI) campaigns from the likes of Nvidia.

Reuters reported that Intel, which was the second technology company to join the DJIA in the late 1990s, was “likely to be removed from the Dow” because of a “near 60% decline in the company’s shares this year that has made it the worst performer on the index and left it with the lowest stock price on the price-weighted Dow.”

Analysts and financial observers were mixed on the ultimate implications for enterprise IT executives. On the one hand, Intel’s installed enterprise base is so huge that it is not likely to face any imminent danger. That gives Intel a couple of years to turn things around.

But genAI is the perception problem. If they are seen as lagging in that space, that perception could hurt them severely.

However Ryan Shrout, president of Signal65, thinks Intel’s huge installed base will provide a buffer. He spent almost five years at Intel before departing in September 2023, with his final role being Intel’s senior director for client segment strategy in the graphics and AI group.

“Even though Intel appears to be so far behind in the world of technology based on their earnings report and the race versus Nvidia in the AI space, you have to keep in mind that something like 80% of the client market — laptops and PCs — use Intel chips,” Shrout said. “Even in the data center CPU space, 70% or so are using Intel Xeon processors. If Intel disappeared tomorrow, nobody has the capacity to fill that gap.”

But Shrout echoed analysts and pointed to AI strategy, or at least the perception of that strategy, as the overwhelming cause of Intel’s current difficulties. 

“The competition that’s come into the market was allowed to come in because Intel didn’t see the writing on the wall for the AI movement. That’s a self-inflicted blind spot,” Shrout said. 

Intel has taken various steps to try and strengthen its financial numbers, such as recently having suspended its dividend and laying off about 15% of its employees, along with splitting its foundry operations from its design teams. 

“Intel CEO Pat Gelsinger and key executives are expected to present a plan later this month to the company’s board of directors to slice off unnecessary businesses and revamp capital spending,” said a Reuters report. “The plan will include ideas on how to shave overall costs by selling businesses, including its programmable chip unit Altera, that Intel can no longer afford to fund from the company’s once-sizeable profit.”

Forrester senior analyst Alvin Nguyen, who oversees their Intel coverage, said that he is still a fan of Intel’s long term strategy, but he sees various problems with their execution.

“Foundry is very expensive. It’s capital intensive,” Nguyen said. “They have made a big bet on the foundry business. If it works, they will have the best semiconductor fab process [in the industry]. If they win the foundry battle, people will look at them differently.”

Some have questioned whether Intel was wrong to decline to invest in OpenAI, but Nguyen said that he thinks it might have been the right decision for Intel. Indeed, he saId, “I am wondering if Microsoft today is questioning the wisdom of their decision [to invest in OpenAI].”

Nguyen added that Intel’s “push towards AI everywhere seems like a smart bet.” He added that Intel’s lack of position within mobile and IoT devices is a problem.

As for the prospect of Intel being removed from the DJIA, Nguyen doubted it would make much of an impact. “It’s just a status symbol. If they lose their Dow status, it’s more of a reputational hit than anything else,” he said.

Nguyen agreed with Shrout that Intel’s massive current installed base will insulate the company for at least a couple of years, giving them time to turn things around. 

“Intel is still in danger and the more hits they take, the worse their position,” Nguyen said.

Another Intel industry analyst is Mario Morales, the IDC group vice president for semiconductors and enabling technologies. 

“There is an ongoing battle for survival at Intel,” Morales said, adding that he thinks that splitting the company and selling off divisions may be the best move. “The parts of Intel are more valuable as pieces than as a whole.”

Morales’ sources have reported that Intel is “actively talking with more than 100 customers, but none of them have yet committed” to more major purchases, he said.

A critical problem for Intel in the perception realm is that they have been outsourcing too much; the manufacturing of both its Lunar Lake and Arrow Lake CPUs were almost entirely outsourced to Taiwan Semiconductor Manufacturing Company Limited (TSMC).

“Even Intel’s own products are being built somewhere else,” Morales said, suggesting that such a move is sending the wrong message to enterprise CIOs. This is happening just as those executives are thinking about creating their own on-prem operations for genAI deployments, in an attempt to gain more control than they now have in the cloud.

“Intel has always had a lot of technology that can enable genAI. They simply had the wrong product mix,” Morales said. As the industry moved from CPUs to GPUs, Intel didn’t move quickly enough, he said.

On an optimistic note, Morales said that there is industry precedent for exactly such a turnaround. Some ten years, AMD faced similar issues and overcame them.

“In 2014, AMD was a month or two away from bankruptcy,” Morales said, stressing that “because AMD was so close to death,” its CEO halted a wide range of side projects that were not central to their customers. 

“Intel has to suffer the tough pill [and decide that] ‘If we can’t lead (in a segment), then we can’t be in those spaces,’” Morales said. “It is well beyond a wakeup call. They are already late.”

In another interesting move that hints at a symbiotic relationship, ChatGPT maker OpenAI has reportedly followed Apple to become a lead customer for TSMC processors. Given the industry lead Apple has achieved with Apple Silicon, the move could be seen as tacit enthusiasm, rather than symbiosis, but follows reports Apple might stake an investment in OpenAI.

These moves by the biggest names in tech underscore the profound difference generative AI (genAI) has made in artificial intelligence, which has taken what’s been part of the industry for decades and placed it at the forefront of the zeitgeist. That Open AI plans to work with TSMC can also be seen as justification of the integrity of Apple’s approach to silicon design, as it concedes the computational power these processors provide while meeting real world needs in terms of energy supply.

The first OpenAI chips under the purported deal are set to slip off the lines some time in 2026.

A new platform battle?

As Apple stands at the cusp of becoming the world’s biggest multi-platform AI ecosystem, the move also hints at new competition down the road.  After all, it was only earlier this year that OpenAI CEO Sam Altman was reported to be getting into chip manufacturing. Now, the company has booked into early production of chips using TSMC’s A16 process, which are expected to enter production in 2026. 

Despite using the same foundry, the processors won’t be the same as Apple’s and will be designed apparently by Broadcom and Marvell.

While it is very possible that OpenAI wants to use its chips inside its own servers, it is also plausible it might also have plans to introduce its own devices, or to offer up its AI inside chips as options to other computer hardware manufacturers.

It takes energy to make things happen

Everyone with a passing interest in genAI recognizes that the scale of energy consumption required to deliver server-based services using the tech is very, very high. Even at this point in genAI deployment, the energy being used is higher than that required by some smaller nations — and those demands will only increase.

With that in mind, Apple’s M-series chip message around computational performance per watt turns out to be even more prescient than earlier believed. After all, if genAI is to be woven into global use, it must meet those needs without using all the world’s energy; reducing energy demands is mandatory. This also implies tech firms will continue to make major investments in renewable energy supply to drive those server farms, and suggests the carbon offset market will be forced to prove its legitimacy, rather than continuing to be a kind of 21st century equivalent of Papal Indulgences (as George Monbiot once described it).

Power, profit, people

The chips Apple makes deliver excellent computational performance at significantly less power than rival processors. Once Apple’s production moves to TSMC’s A16 process, you’ll see another 8-10% spike in performance for up to 20% less power, a report claims.

That’s great for Mac, iPad, and iPhone users — who doesn’t want more powerful devices that use less energy? But for server-based services handling millions of requests daily, that power difference affects both environmental performance and operational costs in terms of energy bills.

With that in mind, OpenAI doesn’t need to be looking to become a hardware competitor to unlock value from chip design; its own running costs will be reduced dramatically through the introduction of more efficient chips — particularly as the number of people it serves grows from millions to billions.

While people in tech might see AI everywhere, most people haven’t begun using genAI tools and services just yet — something which is going to change within the next few weeks as Apple ships its AI-ready devices, starting with the next iPhone.

But if the direction of travel is anything to go by — a trajectory in which Apple and Microsoft seem set on investing in a company that could yet compete with both of them — it seems the people at the summit of Tech Power Mountain don’t merely see OpenAI as a service provider, but as a peer player in the future of IT. We just have to hope that neither they, nor the AI, are hallucinating.

Please follow me on LinkedIn, Mastodon, or join me in the AppleHolic’s bar & grill group on MeWe.

Download the September 2024 issue of the Enterprise Spotlight from the editors of CIO, Computerworld, CSO, InfoWorld, and Network World.

According to credible rumors, Snap and Meta will soon unveil their next-generation AI glasses.

Snap might introduce its fifth-generation Spectacles at the Snap Partner Summit on Sept. 17. Features like a wider field of view and improved battery life could grace a production run of fewer than 10,000 units for developers. (The current 4th-generation Spectacles are also for developers only; Snap never sold them to the public.)

Meanwhile, Meta could well unveil its “Orion” project at its Connect conference, scheduled for Sept. 25-26. According to the latest tech chatter, Orion glasses are expected to be highly advanced augmented reality (AR) glasses with immersive technology and a design that makes them look like standard glasses. While Orion glasses won’t be available for sale right away, Meta is prepping around 1,000 units for demonstration and early developer exploration.

Both companies hope to kickstart a third-party developer ecosystem for high-quality AI-driven AR glasses that anyone can wear every day in polite society as ordinary eyeglasses. And while the developers are working on apps, the companies will work hard to bring down the costs of manufacturing the products.

This generation of glasses combines an AI voice assistant with AR holographic visuals plus all the features and functionality of Bluetooth earbuds (calls, podcasts, etc.). In theory, this is the Holy Grail of wearables — fantastic power, rich user interfaces, and invisible and inaudible to everyone around you. Wearing such glasses will make you feel like you know everything and are constantly aided by powerful AI.

Meanwhile, Meta’s Ray-Ban Meta glasses are the surprise hit of the year. The glasses look more or less like regular Ray-Bans but offer multimodal AI and an audio interface. They’re inexpensive because they don’t attempt visual output, only sound.

While the Ray-Ban Meta glasses are well-designed and well-made and equipped with quality speakers and microphones, the basic concept is easily replicable. Chinese companies accessing lower-cost components can make much cheaper glasses containing batteries, Bluetooth connectivity, speakers, microphones, and an app that connects to the hardware and gains access to generative AI (genAI) chatbots through APIs.

In other words, good-enough AI glasses are relatively easy and inexpensive to produce. That’s why the success of Ray-Ban Meta glasses has Chinese companies taking notice.

Wait, what’s happening in China?

Smaller Chinese companies are focused on the growing market for AI-powered smart glasses, aiming to compete directly with Ray-Ban Meta smart glasses. They’re coming from companies you probably never heard of:

• Superhexa: Backed by Xiaomi, Superhexa is a Chinese startup that has launched “Jiehuan” branded AI glasses that provide access to large language models (LLMs) and offer voice-guided navigation and AI chat features.
• Solos: Hong Kong-based Solos has introduced its AirGo Vision smart glasses, which enable voice access to Open AI’s ChatGPT. The glasses also have a detachable camera, which enables multimodal AI via GPT-4o.
• Even Realities: This Shenzhen-based startup makes G1 glasses featuring LED microdisplays. Although they don’t have speakers, they do output information in the form of visible green text.
• Liweike: Based in Hangzhou, China, this company developed smart AR glasses, unfortunately branded as Meta Lens S3 glasses. They integrate sports functionality with AI-powered voice interaction with the company’s AI chatbot. One standout feature is an integrated 120-degree ultra-wide 2K high-definition sports camera.
• Sharge: This company’s OptoX AI Glasses have a camera, speakers, and all the trimmings. Users can access ChatGPT-4o by talking and listening. Also, they can function like a dashcam, constantly recording and deleting while retaining only the last 30 minutes of video, which you can watch or keep.

As you can tell from this list, some companies are making audio-only AI glasses, some of which will cost less than $100. Other glasses add holographic heads-up displays, which could cost a few hundred dollars a pair. At least two of these products offer both cameras and interaction with the advanced GPT-4o chatbot, able to essentially do all the stuff from the May 13 OpenAI Spring Update (everything, of course, except get Scarlett Johansson’s voice), but through glasses instead of a smartphone.

Great glasses or cheap glasses?

To oversimplify the coming AI glasses market, the American companies will make them great, and the Chinese companies will make them cheap. The result will be an incredible selection of variable features, quality, and styles.

As a result, we’ll quickly arrive at a place where the question won’t be, “Why buy AI glasses?” It’ll be: “If you’re going to buy glasses, why wouldn’t you buy AI glasses?”

The AI glasses revolution will also eviscerate the in-the-ear earbud market and might damage the smartwatch industry. With audio in your glasses, why put plastic in your ears? And with apps, notifications, and information hovering holographically in space in front of your eyeballs, who needs it on your wrist.

Smartphones could even be affected. If a big holographic display in the lenses provides the main interface, then a big-screen smartphone might be unnecessary.

The emergence of AI glasses as a ubiquitous category creates interesting and valuable possibilities for enterprise and business apps running on the platforms, everything from factory training and instructions to board room presentation teleprompters.

Beyond that, we might eventually see the rise of BYOG — bring your own glasses — policies. Companies will also need to cope with this generation of glasses’ privacy and security implications. The prescription glasses employees rely on to see clearly will often have cameras and microphones capable of secretly recording anything. (Even Ray-Ban Meta glasses have a bright light that indicates when the camera is taking a picture or recording video, a system easily foiled, according to hundreds of how-to videos on TikTok.)

Social implications will abound. Today, we’re still trying to figure out the social norms around looking at a smartphone during a conversation. What happens when people can be looking right at you but secretly distracted by online content only they can see?

AI glasses are about to significantly impact business, society, and culture. This will become clear when Snap and — more importantly — Meta will likely demonstrate the future of AI glasses to developers and the public.