Close Menu
  • AI
  • Content Creation
  • Tech
  • Robotics
AI-trends.todayAI-trends.today
  • AI
  • Content Creation
  • Tech
  • Robotics
Trending
  • Tesla Rolls Out FSD V15 with Actuality View and Imaginative and prescient Upgrades – Unite.AI
  • Potential Cognition AI Spherical Would Elevate $2B-Plus at $75B Valuation – Unite.AI
  • OpenRouter Cuts Three Suppliers After AI Immediate Publicity Findings – Unite.AI
  • A Group of Researchers from Columbia and CertiK Introduce TensorLean for Formal Verification of GPU Kernels
  • Collectively AI Discloses Immediate Leak After Coding Agent Uncovered Credential – Unite.AI
  • Dragos Finds Immediate Leaks and Stolen Keys in AI Provide Chains
  • Lean and Mathlib Communities Certify Claude-Assisted Riemann Proof – Unite.AI
  • AWS Routes HyperPod Inference Throughout Areas From One Endpoint – Unite.AI
AI-trends.todayAI-trends.today
Home»Robotics»Fredrik Nilsson, Vice President, Americas, Axis Communications – Interview Collection – Unite.AI

Fredrik Nilsson, Vice President, Americas, Axis Communications – Interview Collection – Unite.AI

Robotics By Gavin Wallace02/10/202612 Mins Read
Facebook Twitter LinkedIn Email
Nordic Startup IntuiCell Unveils World's First Digital Nervous System for
Nordic Startup IntuiCell Unveils World's First Digital Nervous System for
Share
Facebook Twitter LinkedIn Email

Fredrik Nilsson, Vice President, Americas, Axis Communications, oversees the corporate’s operations throughout North and South America, main groups centered on creating and delivering network-based safety options for companions and finish customers. A longtime Axis government, Nilsson has performed a major function within the safety trade’s transition from analog closed-circuit tv to IP-based community video. Since taking accountability for the Americas in 2003, he has helped develop regional income from roughly $20 million to effectively over $1 billion and supported the enlargement of greater than 20 Axis Expertise Facilities throughout the Americas. He joined Axis in 1996, beforehand serving as Product Supervisor and Division Supervisor in Sweden and later as Director of Enterprise Growth, the place he helped set up partnerships in the course of the early improvement of IP video. Earlier in his profession, he was a Product Supervisor at ABB, and he at present serves on the boards of MSAB and the Swedish-American Chamber of Commerce in Washington, D.C.

Axis Communications is a Sweden-based know-how firm specializing in clever bodily safety and community options, together with community cameras, video surveillance techniques, entry management, intercoms, community audio, radar, and analytics. Based in 1984 and headquartered in Lund, Sweden, Axis employs roughly 5,000 individuals throughout greater than 50 international locations and has been a part of the Canon Group since 2015. The corporate launched what it describes because the world’s first community digital camera in 1996 and has since expanded from community video right into a broader portfolio of related safety applied sciences. Its present merchandise more and more incorporate edge-based analytics and synthetic intelligence to rework video, audio, and sensor knowledge into actionable insights for safety, security, operational effectivity, and enterprise intelligence.

Your profession has spanned distant monitoring at ABB, the early improvement of network-attached storage and Web of Issues applied sciences, and the transition from analog surveillance to community video at Axis Communications. How has that background formed the best way you concentrate on immediately’s shift towards AI-powered cameras and clever bodily infrastructure?

I see the shift towards AI-powered cameras as the following step in an evolution I’ve been part of by way of my profession – from related know-how and IP convergence to immediately’s clever edge. There’s a frequent thread all through my profession, which helps prospects perceive what related know-how can do for them. ABB was an awesome college for me. It was a big world firm, so I had the chance to journey the world, together with spending a 12 months in Canada, and work with massive enterprise prospects, particularly energy utilities, on advanced tasks. My function was product supervisor for a brand new IP-connected distant monitoring and management system. At Axis, I later labored on a brand new class of IP-connected storage gadgets and helped unfold the message globally about the advantages of related know-how. Then, within the early days of IP convergence, I used to be constructing on those self same rules, serving to enterprise prospects perceive the worth of network-connected video and the alternatives created by IP know-how. So, once I have a look at AI-powered cameras immediately, I see one other step in that very same evolution. The know-how retains getting extra succesful, however the actual alternative continues to be about connecting gadgets, producing helpful knowledge and serving to prospects do extra with it.

What kinds of laptop imaginative and prescient workloads are finest suited to run immediately on a digital camera, and which functions nonetheless profit from being processed in a neighborhood knowledge middle or the cloud?

The workloads finest suited to run immediately on a digital camera are these the place you want real-time selections, low latency, and environment friendly use of bandwidth. For instance, detecting and classifying individuals or objects, counting, monitoring motion, or figuring out particular occasions. With applied sciences reminiscent of Axis Object Analytics, cameras can precisely detect individuals, autos, colours, path of journey, and different traits. We will additionally prepare a digital camera for very particular objects or eventualities by way of Axis Customized Analytics. As analytics proceed to enhance, extra of that intelligence can occur on the edge. For a lot of use instances, working the analytics on the digital camera and sending solely the ensuing knowledge or metadata to a server, or the cloud is necessary for scalability and price. The cloud and servers nonetheless have an necessary function, notably for issues like centralized administration, long-term storage, and analyzing metadata throughout massive numbers of gadgets. I don’t see it as an either-or resolution. The sting and the cloud can complement one another, with every doing what it’s best suited to do.

Deploying AI on the edge requires balancing mannequin accuracy towards limitations in processing energy, reminiscence, vitality consumption, and warmth. How are advances in specialised chips altering what can realistically be achieved inside a digital camera?

There are actually two traits driving this. First, the AI capability of the processor on the edge is bettering in a short time. ARTPEC-9, our in-house-developed system-on-chip, has thrice the AI processing capability of ARTPEC-8, giving us considerably extra room to run subtle fashions immediately on the digital camera. Future generations will proceed to extend that functionality. Second, the dimensions and effectiveness of the fashions are bettering. That’s the place we’ve seen plenty of progress over the previous 12 months. Our R&D crew has achieved a tremendous job of getting extra of the mannequin to run immediately on the edge, so in lots of eventualities you solely want to make use of the info from the digital camera somewhat than sending the video elsewhere for processing. A 12 months in the past, we had been speaking about roughly 50% of the wants being dealt with within the digital camera. At this time, it’s nearer to 90%, and we anticipate that to achieve 95% or extra. The true breakthrough isn’t simply that we’ve extra highly effective chips. It’s that we’re combining extra succesful {hardware} with extra environment friendly AI fashions, permitting an rising quantity of intelligence to occur immediately the place the info is generated. That makes surveillance techniques a lot much less depending on servers and the cloud for AI.

Cameras are more and more being positioned as operational sensors somewhat than solely as safety gadgets. Which functions in retail, transportation, healthcare, manufacturing, or logistics are at present delivering the clearest enterprise worth?

Manufacturing is an excellent instance as a result of the worth may be tied on to the operation. BMW Group, for instance, is utilizing community cameras as a part of its AI-driven high quality inspection system to seize detailed photos of autos in actual time and determine defects. That could be a very completely different function for a digital camera than merely recording an incident. We’re additionally seeing rising use of video for operational effectivity and enterprise intelligence extra broadly. In transportation and logistics, cameras will help organizations perceive how individuals, autos, and items are transferring by way of a facility. In retail, video can present insights into buyer conduct, area utilization, and staffing. The frequent theme is that the digital camera is producing info that can be utilized to make a course of higher, not simply recording what occurred.

How does combining video with audio analytics, radar, access-control techniques, and environmental sensors create insights that might not be potential from visible knowledge alone?

The extra context you’ll be able to carry collectively, the higher the system can perceive what is occurring. Once you mix completely different inputs, you will get insights that might be troublesome, and generally not possible, to derive from video alone. For instance, you’ll be able to mix what an object appears to be like like with how it’s transferring, or use thermal and visible info collectively when lighting or visibility is difficult. Radar can add details about an object’s motion and pace, audio can determine sounds or present two-way communication, and environmental sensors can present further context about what is occurring within the surrounding setting. A number of sensors may enhance confidence in what the system or consumer is detecting. Combining know-how is all extremely necessary to guard key areas of life like important infrastructure.

We’re already transferring on this path. At GSX, for instance, we launched options that mix visible, thermal, radar, audio, and environmental sensing. The objective is to carry collectively the fitting knowledge so organizations could make sooner, better-informed selections. That is notably necessary for functions reminiscent of defending important infrastructure, the place having a number of sources of knowledge can present a extra full image of what’s occurring.

Edge processing can cut back the quantity of uncooked video that must be transmitted or saved, however AI-enabled cameras nonetheless elevate necessary privateness and governance questions. What rules ought to organizations comply with when designing techniques which might be helpful with out accumulating extra private knowledge than mandatory?

The start line must be to ask what the system truly wants to perform and gather solely the data mandatory for that goal. Privateness and transparency are additionally necessary. Individuals need to perceive why video is getting used and the way their info is being dealt with. Our latest analysis reveals that individuals are changing into extra comfy with video surveillance, however that acceptance can’t be taken with no consideration. The largest issues round AI-enabled surveillance are false positives or inaccurate info and the potential misuse or mishandling of information. That places an actual accountability on organizations and know-how suppliers to construct privateness and safety into techniques from the start. Edge processing will help right here as a result of you’ll be able to analyze info regionally and ship solely the info that’s wanted, somewhat than transferring steady uncooked video across the community.

Bodily infrastructure could stay deployed for a few years, whereas AI fashions and cybersecurity threats evolve quickly. How ought to cameras and different edge gadgets be designed in order that their analytics stay safe, maintainable, and adaptable all through their operational lives?

This is without doubt one of the causes I consider the digital camera as a part of the broader IT and operational infrastructure, somewhat than a standalone safety gadget. Cyber safety must be constructed into the gadget from the beginning, with safe software program, common updates, and robust cybersecurity protections all through its lifecycle. At Axis, for instance, Edge Vault gives a hardware-based basis for shielding gadget integrity, securing cryptographic keys, and establishing a sequence of belief. It additionally helps capabilities reminiscent of safe boot and signed software program, serving to make sure that solely trusted software program runs on the gadget.

The gadget additionally wants an structure that may adapt as analytics and buyer necessities change. Open platforms are necessary right here as a result of they permit organizations to combine new functions and applied sciences somewhat than having to switch a whole system each time one thing evolves. On the identical time, organizations want good governance round their related gadgets. As extra video, sensors and enterprise techniques come collectively, cybersecurity and knowledge safety must be handled as a part of the general system design, not as one thing added later.

Axis helps an open utility ecosystem that permits third-party builders to construct and deploy specialised analytics on its gadgets. How necessary will open platforms and interoperability be as organizations search to run their very own fashions throughout massive fleets of cameras?

I feel will probably be extraordinarily necessary. Prospects don’t need to be locked into one know-how or a method of doing issues, particularly as AI and analytics proceed to evolve so rapidly. An open platform makes it simpler to combine cameras with different techniques, herald specialised functions and scale throughout a big deployment. It additionally offers prospects extra flexibility as their wants change. By eliminating vendor lock-in and simplifying integration, open, interoperable ecosystems make it simpler to scale clever edge options whereas adapting to altering enterprise calls for. In the end, the worth isn’t just within the digital camera itself. It’s in how the digital camera, sensors, analytics, and enterprise functions work collectively.

How ought to organizations consider the efficiency of laptop imaginative and prescient techniques after deployment, notably when lighting, digital camera positioning, human conduct, or working situations can change and doubtlessly have an effect on accuracy?

All of it is dependent upon the use case, however a very powerful factor is to check the cameras and AI fashions in the actual setting the place they’ll truly be used. You’ll be able to have a mannequin that performs very effectively in a managed setting, however the actual take a look at is what occurs with the precise lighting, digital camera place, climate, background, and conduct you see on the web site. That’s the reason we encourage prospects to guage analytics reminiscent of Axis Object Analytics or Axis Customized Analytics in the actual setting and ensure the accuracy is excessive sufficient for the appliance. The BMW manufacturing facility is an effective instance of this method. The cameras are being utilized in an actual manufacturing setting, the place the standard of the photographs and the ensuing analytics must help an precise manufacturing course of.

As edge AI turns into extra succesful, how do you anticipate the function of the community digital camera to evolve over the following decade, and what new capabilities might emerge as cameras change into lively elements of clever, responsive environments?

The community digital camera is changing into way more central to the system. It’s not only a digital camera producing video. It’s a sensor producing knowledge that may be analyzed and shared throughout completely different techniques. Because the processing energy continues to extend, I anticipate cameras to change into even higher at understanding what is occurring of their setting and offering helpful info in actual time. That’s the concept behind Scene Intelligence, the place the digital camera can use a deeper understanding of the scene to offer extra significant details about objects, their traits, and what’s occurring round them.

That might imply extra superior object and conduct detection, larger integration with different sensors and enterprise techniques, and extra automation round how organizations reply to occasions. The larger shift is that video turns into one a part of a broader knowledge and intelligence layer. The digital camera continues to be capturing what is occurring, however more and more it could additionally assist clarify what is occurring and what a company could need to do subsequent.

Thanks for the good interview, readers who want to be taught extra ought to go to Axis Communications.

AI communications Interview x
Share. Facebook Twitter LinkedIn Email
Avatar
Gavin Wallace

Related Posts

Tesla Rolls Out FSD V15 with Actuality View and Imaginative and prescient Upgrades – Unite.AI

22/11/2026

Potential Cognition AI Spherical Would Elevate $2B-Plus at $75B Valuation – Unite.AI

22/11/2026

OpenRouter Cuts Three Suppliers After AI Immediate Publicity Findings – Unite.AI

22/11/2026

Collectively AI Discloses Immediate Leak After Coding Agent Uncovered Credential – Unite.AI

22/11/2026
Top News

Age Verification Is Sweeping Gaming. Are you ready for AI Fakes in the Age of Gaming?

Meta Pushes Its New AI Agent on Employees—but Eases Off on Tokenmaxxing

“Create a replica of this image. Don’t change anything” AI Trend Takes Off

The AI wearable from ex-Apple engineers looks like an iPod Shuffle

AI Fashions Can Be Simply Modified to Help Terrorist and Extremist Exercise

Load More
AI-Trends.Today

Your daily source of AI news and trends. Stay up to date with everything AI and automation!

X (Twitter) Instagram
Top Insights

Epoch AI Reviews Open-Weight GLM 5.3 Exploits Firefox Patch Hole – Unite.AI

30/10/2026

The Coding implementation to master GPU computing with CuPy and Custom CUDA kernels. Sparse matrices, Streams.

15/05/2026
Latest News

Tesla Rolls Out FSD V15 with Actuality View and Imaginative and prescient Upgrades – Unite.AI

22/11/2026

Potential Cognition AI Spherical Would Elevate $2B-Plus at $75B Valuation – Unite.AI

22/11/2026
X (Twitter) Instagram
  • Privacy Policy
  • Contact Us
  • Terms and Conditions
© 2026 AI-Trends.Today

Type above and press Enter to search. Press Esc to cancel.