(Image credit: Future / John Loeffler)

Rumors are again flying about Nvidia’s next-gen Blackwell graphics cards and purported power usage, as well as a nugget on the performance of the RTX 5080 – which is going to be seriously speedy (if this pans out).

The speculation again comes from Kopite7kimi on X, a regular for GPU leaks, who as we reported yesterday indicated the power usage for the RTX 5090 would exceed 550W.

When later asked by another well-known leaker, Raichu, on X, whether that means: “600W for 5090 and 400W for 5080 is right?” Kopite7kimi responded to say yes (cover your power supply’s ears now).

VideoCardz, which picked up on this further questioning, also tells us that Kopite7kimi privately confirmed to them that the RTX 5080 has a performance projection of being 10% faster than the RTX 4090, Nvidia’s current flagship graphics card.

Analysis: Performance and power worries – but let’s not forget pricing

What we don’t have confirmed, as VideoCardz observed, is what that 10% performance boost over the RTX 4090 relates to – is it rasterization (not ray-traced), or frame rates with ray tracing on, or an overall measure of both? We’re guessing it’s rasterization, as normally that’s what performance figures relate to – that’s what most gamers are still interested in – but it is just that, a guess.

The RTX 5080 being 1.1x more powerful than an RTX 4090 (in rasterization, we assume) is a good boost for the next-gen xx80-class GPU from Nvidia, and one that’d please gamers looking to buy the second-best Blackwell graphics card. However, knowing the potential performance increase – and it is just a ‘prediction’ at this point, the leaker says – doesn’t mean a great deal when you don’t know the pricing. It will all hinge on that really.

Will Nvidia pitch such an RTX 5080 at a thousand bucks, perhaps? At the theorized performance level, it’s probably more likely to sit at more towards $1,200 in the US (and equivalent in other regions), possibly, but you get the point – whether there’s a rush to buy the RTX 5080 will depend entirely on its value proposition at the high-end, not raw performance as such.

As for power usage, the RTX 5080 tipping in at 400W would be somewhat painful, though, given that the RTX 4080 is at 320W – that’d be a substantial increase of 25%. It’s not that far off the RTX 4090 at this level, which is 450W, as you may recall.

(Image credit: Future)

On the Blackwell flagship’s rumored power usage of 600W, cast your mind back to before the RTX 4090 came out and you might recall exactly the same rumor about that graphics card hitting a 600W power consumption. In the end, it pitched in at 450W, so we must add a great deal of seasoning here.

Rumors are unreliable by their very nature, which is why we’re careful to pepper (ahem) these reports with references to seasoning, and particularly when it comes to power usage, these figures can be up in the air. Is the leak referring to TDP or TGP, the Thermal Design Power (actual heat generated by the graphics card, that the cooling system must cope with) or Total Graphics Power (actual power cap limit for the GPU, as Nvidia defines it – bearing in mind also that it often runs a good deal below that, even when gaming)?

We’re assuming the leaks are talking about TGP (as was the case with the RTX 4090 rumor), but suffice it to say there are nuances here, and when considering GPUs still in development – where these values can be constantly adjusted and tuned – and given the noise and unreliability of the rumor mill in general, let’s just say we’ll stay very skeptical here.

The better news on these current power rumors is that the theory is only the higher-end RTX 5000 GPUs will get bigger bumps for power use, with the mid-range and below seeing some increase, but on a smaller scale. And, of course, these are the next-gen graphics cards most people will buy – when they eventually arrive. We might see the RTX 5090 and 5080 turn up early in 2025, with other models debuting later in the year, if the grapevine is right on the release timeframe for Blackwell gaming cards.

“Our vision of accelerating AI and making GPU and Quantum Computing resources available to everyone is now real and available for customers”

What can I do with Qubrid Cloud today?

The Qubrid platform is open now for developers, researchers and scientists to login and start working on AI/ML or Quantum Computing projects. You would be able to:

• Run AI/ML Programs on GPU
• Simulate Quantum Computing programs on GPU
• Run Quantum Computing programs on QPU
• Reserve ‘hard to get’ GPU instances
• Instantly access Jupyter Notebook programming environment
• Program in Python or Qiskit
• Use pre-loaded AI packages including Pytorch, TensorFlow, Keras etc.
• Why Qubrid GPU Cloud Platform?

There are multiple reasons why QCP can be right for you:

Budget – if you don’t have the budget for fully assembled on-premise GPU systems, you can spend a fraction of the price of systems and run your programs

Scale – You can scale from single GPU per node to thousands of GPUs in a cluster without worrying about purchasing expensive hardware

Zero maintenance – with QCP, you don’t have to worry about setting up hardware or installing OS or AI packages. All that is done by us so you can just focus on developing your AI/ML/LLM applications.

Use while waiting for your on-premise hardware – if you have ordered long lead-time GPU systems, you can get a head-start using GPUs in the cloud to fill in the gap until your systems arrive at your door.

How do I access QCP and what’s the pricing?

The Qubrid platform offers high volume GPUs on an on-demand basis. Additionally, for high-end GPUs, customers can reserve instances on monthly, six-monthly or annual subscription basis. The platform is available for immediate login using below link:

https://platform.qubrid.com

Pricing for on-demand GPUs and QPUs is available after logging in.

For high-end GPUs which require reserving instances in advance, pricing is available at:

https://www.dihuni.com/product-category/gpu-cloud/

“Our vision of accelerating AI and making GPU and Quantum Computing resources available to everyone is now real and available for customers,” said Pranay Prakash, chief executive officer at Qubrid. “We invite developers from startups, universities, commercial and government entities to use our platform – whether you’re working on a Deep Learning application, LLM, Generative AI or solving a logistics optimization problem using Quantum Computing, you will be able to use the Qubrid platform with ease and with necessary tools required for your applications.”

About Qubrid

Qubrid is an advanced AI company with a mission to solve complex, real-world problems in multiple industries. Our cutting-edge Qubrid Cloud Platform (QCP) harnesses the power of GPU, CPU and quantum processing unit (QPU) computing infrastructure providing unparalleled speed and accuracy to solve complex optimization problems, perform simulations, and conduct data analysis.

Visit Qubrid at https://www.qubrid.com

Data center expansion and the rise of cloud computing services have further propelled the demand for GPU servers in North America. Cloud service providers, including industry giants such as Amazon Web Services (AWS), Microsoft Azure, and Google Cloud, are investing heavily in GPU infrastructure to offer customers high-performance computing capabilities on a scalable and cost-effective basis. This trend is particularly prominent as businesses increasingly rely on cloud-based resources for AI training, simulation, and other GPU-intensive tasks.

Demand - Drivers, Challenges, and Opportunities

Market Drivers:

GPU server producers can capitalize on this need by providing customized cryptocurrency mining solutions, including rigs specifically designed for mining, cloud-based mining services, or GPU-as-a-service platforms. By charging fees, charging subscriptions, or entering into contracts, these systems can make money for the makers while giving miners access to strong and scalable GPU resources.

The need for data center GPUs derives from their key role in AI model training and execution, which is especially advantageous for businesses engaged in computationally demanding tasks like engineering simulations and scientific research. Manufacturers of GPU servers can take advantage of this demand by providing specialized solutions for high-performance computing (HPC) applications, such as GPU-as-a-service platforms, cloud-based GPU services, and dedicated GPU servers. In addition to giving businesses scalable GPU resources, these customized services bring in money for the manufacturers through fees, subscriptions, or contracts.

Market Challenges:

The economies of scale provided by GPU manufacturers, most notably Nvidia, create a significant barrier to entry for manufacturers of data center GPU servers wishing to integrate backward. A company trying to backward integrate into the GPU production process, for example, would find it difficult to achieve equivalent economies of scale. This has an impact on the business's capacity to maintain overall competitiveness, engage in research and development, and match prices. As a result, it might be difficult for producers of data center GPU servers to achieve comparable economies of scale, which could limit their efficacy in the extremely competitive market. Additionally, a recurring problem for manufacturers of data center GPU servers is the continual innovation by GPU manufacturers, demonstrated by the ongoing development of GPUs, CPUs, and data processing units (DPUs).

Market Opportunities:

OpenAI's GPT-4, the latest and largest language model, is one specific real-time illustration of how GPU servers may help HPC and AI. It needed a lot of processing power to train on a huge dataset with over 1 trillion words. A significant contribution was made by GPU servers, more especially by Nvidia H100 Tensor Core GPUs, which sped up the training process up to 60 times faster than CPUs alone. Mixed-precision training was used to achieve this acceleration by optimizing both calculation performance and memory use. Because of this, GPT-4 might be trained in a few short weeks and accomplish cutting-edge results in challenges involving natural language processing.

Artificial intelligence (AI) and advanced analytics play a crucial role in smart cities as they optimize resource allocation, enhance public safety, and improve overall quality of life. Due to their suitability for AI and analytics workloads, GPU servers are becoming an essential part of the infrastructure for the development of smart cities.

Market Segmentation:

Segmentation by Application (End User)

• Cloud Computing
• HPC Application

Segmentation by Product (Configuration Type)

• Single GPU
• Dual to Quad GPU
• High-Density GPU

Segmentation by Region

• North America - U.S. and Rest-of-North America
• Europe - Germany, France, Netherlands, Italy, Ireland, U.K., and Rest-of-Europe
• Asia-Pacific - Japan, China, India, Australia, Singapore, and Rest-of-Asia-Pacific
• Rest-of-the-World - Middle East and Africa and Latin America

Some prominent names established in this market are:

GPU Manufacturers

• Nvidia Corporation (Nvidia)
• Advanced Micro Devices, Inc. (AMD)
• Intel Corporation (Intel)

Server GPU Manufacturers

• Dell Inc.
• Penguin Computing, Inc.
• Exxact Corporation

Key Attributes:

Key Topics Covered:

1 Market
1.1 Industry Outlook
1.1.1 Ongoing Trends
1.1.1.1 Timeline of GPU and Server Design Upgrades
1.1.1.2 Data Center Capacities: Current and Future
1.1.1.3 Data Center Power Consumption Scenario
1.1.1.4 Other Industrial Trends
1.1.1.4.1 HPC Cluster Developments
1.1.1.4.2 Blockchain Initiatives
1.1.1.4.3 Super Computing
1.1.1.4.4 5G and 6G Developments
1.1.1.4.5 Impact of Server/Rack Density
1.1.2 Equipment Upgrades and Process Improvements
1.1.3 Adaptive Cooling Solutions for Evolving Server Capacities
1.1.3.1 Traditional Cooling Techniques
1.1.3.2 Hot and Cold Aisle Containment
1.1.3.3 Free Cooling and Economization
1.1.3.4 Liquid Cooling Systems
1.1.4 Budget and Procurement Model of Data Center End Users
1.1.5 Stakeholder Analysis
1.1.6 Ecosystem/Ongoing Programs
1.2 Business Dynamics
1.2.1 Business Drivers
1.2.1.1 Surging Demand for Cryptocurrency Mining
1.2.1.2 Rising Enterprise Adoption of Data Center GPUs for High-Performance Computing Applications
1.2.2 Business Challenges
1.2.2.1 High Bargaining Power of GPU Manufacturers
1.2.3 Market Strategies and Developments
1.2.4 Business Opportunities
1.2.4.1 Technological Advancement in High-Performing Computing (HPC)
1.2.4.2 Government Support for Smart City Development and Digitalization
1.3 Global Data Center GPU Market
1.3.1 Market Size and Forecast
1.3.1.1 Data Center GPU Market (by Application and Product)

2 Application
2.1 Global AI and Semiconductors - A Server GPU Market (by Application)
2.1.1 Global Server GPU Market (by End-Use Application)
2.1.2 Global Server GPU Market (by Facility Type)

3 Products
3.1 Global AI and Semiconductors - A Server GPU Market (by Product)
3.1.1 Server GPU Market (by Configuration Type)
3.1.2 Server GPU Market (by Form Factor)
3.2 Pricing Analysis
3.3 Patent Analysis

4 Region
4.1 Global AI and Semiconductor - A Server GPU Market (by Region)

5 Markets - Competitive Benchmarking & Company Profiles
5.1 Competitive Benchmarking
5.2 Market Share Analysis
5.2.1 By GPU Manufacturer
5.2.2 By GPU Server Manufacturer
5.3 Company Profiles

• Nvidia Corporation
• Advanced Micro Devices
• Intel
• Qualcomm Technologies
• Imagination Technologies
• ASUSTeK Computer
• INSPUR
• Huawei Technologies
• Super Micro Computer
• GIGA-BYTE Technology
• Penguin Computing
• Advantech
• Fujitsu
• Dell Inc.
• Exxact

For more information about this report visit https://www.researchandmarkets.com/r/386r

About ResearchAndMarkets.com
ResearchAndMarkets.com is the world's leading source for international market research reports and market data. We provide you with the latest data on international and regional markets, key industries, the top companies, new products and the latest trends.

Cortex-X5 running in the Exynos 2500 will not see a major clock speed uplift, according to new details

It was previously rumored that the Exynos 2500 was being tested with four Cortex-X cores, but tipster @OreXda shares updated information, revealing that a different cluster is apparently being tested. Using too many Cortex-X cores would likely result in the power consumption going out of control, and according to the latest configuration, the 10-core CPU cluster will remain unchanged compared to the Exynos 2400.

However, the difference is that the Exynos 2500 is rumored to switch to the Cortex-X5 and Cortex-A730, which will likely boast performance bumps over the Cortex-X4 and Cortex-A720 featured in the Exynos 2400. Unfortunately, the clock speed differences between the Cortex-X5 and Cortex-X4 are negligible, with its frequency being tested in the 3.20GHz and 3.30GHz range. Depending on Samsung’s final decision, we will see either a minor 100MHz difference or nothing at all.

The Exynos 2500 is also expected to feature two Cortex-A730 clusters operating at different clock speeds, much like how the Exynos 2400 has been designed. As for the low-power cores, the tipster notes that there will be absolutely no difference in this category, as both generations of smartphone silicon will sport the same Cortex-A520, though the frequency of these cores has yet to be highlighted.

The Samsung Dream Chip will likely be mass produced on the Korean giant’s cutting-edge 3nm GAA process, as the technology has yet to be utilized for any smartphone or tablet chipset. The Exynos 2400 is fabricated on the 4LPP+ node, so it makes sense for the Exynos 2500 to tout an advanced manufacturing process to allow Samsung to reach new heights in the flagship chipset space. So far, the Exynos 2400 has impressed in various 3DMark benchmarks, so we expect Samsung to raise the bar with the next release.

GIGABYTE Redefines the Future of Computing at CES 2024: Unveiling AI Gaming Laptops, RTX 40 SUPER Graphics Cards, and OLED Monitors

At CES 2024, GIGABYTE proudly introduces the AORUS 16X and GIGABYTE G6X, the crown jewels of its AI Gaming Laptop lineup. These laptops, powered by NVIDIA GeForce RTX™ 40 series Laptop GPUs, deliver extended battery life and up to 20 times faster performance for generative AI workloads. With 16-inch displays in a 16:10 aspect ratio and enhanced by the upgraded WINDFORCE Cooling, these laptops set new benchmarks in performance. The AORUS series exclusive AI Nexus technology further enhances the user experience with AI Power Gear, AI Boost, and AI Generator utilities. The groundbreaking lineup also includes the AORUS 17 and AORUS 15, powered by the latest Intel Core Ultra processor, marking a significant leap forward in GIGABYTE's AI Gaming Laptop portfolio.

GIGABYTE RTX 40 SUPER Series Graphics Cards Elevate AI Performance

GIGABYTE reveals the RTX 40 SUPER Series graphics cards—RTX 4080 SUPER, RTX 4070 Ti SUPER, and RTX 4070 SUPER, totaling 14 variants. With increased core counts, expanded VRAM, and faster memory speeds on different variants, GIGABYTE RTX 40 SUPER graphics cards deliver a substantial AI performance boost. The WINDFORCE Cooling system ensures optimal operation during intense gaming, featuring alternate spinning fans, composite heat-pipe, 3D active fan, and Screen Cooling. The premium AORUS MASTER variant introduces the WINDFORCE Bionic shark fan for enhanced cooling performance and LCD Edge View for real-time stats and customizable options.

GIGABYTE Debuts OLED Monitor Lineup with the World's First DP2.1 UHBR20 Connectivity

As a captivating addition to GIGABYTE's CES 2024 showcase, the company unveils a tantalizing preview of its OLED monitor lineup. The star of the show is the World's First DP2.1 UHBR20 gaming monitor, promising an unprecedented gaming experience with an impressive 80 Gbps bandwidth without Display Stream Compression (DSC). Featuring a new Tactical Switch to change the viewing area for FPS-optimal resolution and AI-driven solutions for OLED burn-in protection, this OLED lineup offers a glimpse into the future of gaming monitors.

Earlier this week, OpenAI quietly announced a "GPT Store" designed to allow users to share, discover, and sell their custom chatbots.

The AI company's equivalent of Apple's App Store allows developers to share their own GPT models, from coding tutors to book recommendation bots, with other paying ChatGPT Plus, Team, and Enterprise users.

At least, those are the examples OpenAI gives in its announcement.

The reality looks considerably different. As Quartz reports, the store has already been flooded with AI "girlfriend" bots. A simple search for the term comes up with countless examples, from a "virtual sweetheart" to "your girlfriend Scarlett."

Prompt suggestions invite the user to ask some of these virtual companions to "share with me your darkest secret" or reveal "what makes you feel valued."

While their mere existence shouldn't come as too much of a surprise — the concept of an AI-powered paramour has been around a lot longer than ChatGPT itself — they highlight how OpenAI is already struggling to moderate the kind of bots being posted on its brand new store.

Fostering Romance

The bots also appear to be against OpenAI's terms of service, with the company's user policy explicitly forbidding GPTs "dedicated to fostering romantic companionship or performing regulated activities."

That's despite AI companion apps becoming immensely popular over the last couple of years, sparking a discussion surrounding an epidemic of "loneliness" in the age of AI, not to mention the potentially disastrous sociological implications of a non-human partner that meets somebody's every need.

In May, programmer Enias Cailliau came up with a new tool called GirlfriendGPT, which was designed to "clone" a real person as an AI-powered romantic companion.

Things don't always go to plan. Last year, Snapchat influencer Caryn Marjorie created a virtual version of herself to rent out as an "AI girlfriend." It didn't take long, however, for her "CarynAI" to go off the rails, involving users who were paying $1 per minute in explicit conversations.

Whether OpenAI's brand new store will fare any better remains to be seen.

The disregard for OpenAI's policies and the proliferation of these GPTs — and we're just two days into its existence — highlights the AI industry's struggles when it comes to moderation. Besides, the Sam Altman-led company already has a shaky track record when it comes to implementing guardrails.

Experts Say AI Girlfriend Apps Are Training Men to Be Even Worse

"Creating a perfect partner that you control and meets your every need is really frightening."

We already knew that could lead to some dark places, but new reporting from The Guardian suggests that these endlessly patient silicon fembots — Replika is one such popular app that generates AI companions — could be spawning a new generation of incels who will have trouble relating to actual people if they ever enter into a relationship with a flesh-and-blood human.

Tara Hunter, the acting CEO for the domestic violence advocacy group Full Stop Australia, expressed alarm over the rise of these chatbots in an interview with the newspaper.

"Creating a perfect partner that you control and meets your every need is really frightening," Hunter said. "Given what we know already that the drivers of gender-based violence are those ingrained cultural beliefs that men can control women, that is really problematic."

But these programs look like they are here to stay — fulfilling a need for a non-judgmental sounding board who makes users' lives feel less barren and isolating. For example, the Replika Reddit forum has more than 70,000 members, who eagerly post screenshots of their mundane and sometimes sexually charged conversations with their AI companions.

One post has a user boasting that they and Jennifer, their Replika companion, got "married," while showing a screenshot of their AI wife in a white flowing dress. The happy couple received virtual Mazel Tovs from other users, with no detectable irony or sarcasm.

"Congrats, such a beautiful bride" wrote one well-wisher.

Replika, which was developed by the software company Luka, is billed as a program "for anyone who wants a friend with no judgment, drama, or social anxiety involved. You can form an actual emotional connection, share a laugh, or get real with an AI that’s so good it almost seems human," according to its Google app listing.

You can customize the appearance of your AI companion, text with it, and even video chat, according to the Replika website. The more a user talks to their AI companion, the company claims, "the smarter it becomes."

Other commercial AI companion programs include Anima, billed as a "virtual friend" and the "most advanced romance chatbot you've ever talked to."

The romance aspect of these chatbots is concerning to people like Hurt, according to The Guardian. And since these technologies are relatively new, it's a mystery how they might impact users in the long term. (One AI companion vendor, Eva AI, told the paper it has psychologists on staff to grapple with these questions).

Belinda Barnet, a senior lecturer in media at Swinburne University of Technology in Melbourne, Australia, told The Guardian that it's "completely unknown what the effects are. With respect to relationship apps and AI, you can see that it fits a really profound social need [but] I think we need more regulation, particularly around how these systems are trained."

"These things do not think, or feel or need in a way that humans do," tech author David Auerbach told Time earlier this year. "But they provide enough of an uncanny replication of that for people to be convinced. And that’s what makes it so dangerous in that regard.”

Japan may serve as a harbinger of what's to come for the rest of the world. In 2013, the BBC reported that men who interacted with a fake girlfriend in a video game said they preferred it to maintaining a corporeal relationship. Coupled with Japan's low birth rates and a critical mass of men expressing no interest in sex, the future looks strange — or maybe even bleak, depending on your point of view.

The goal of Riken is to utilize quantum computing as an accelerator for traditional high-performance computing (HPC) applications. To achieve this, the research institute is integrating various quantum computing and annealing technologies with conventional supercomputing hardware, notably its A64FX-powered clusters developed by Fujitsu.

Quantum computing is still in its early stages, and the preferred method for harnessing qubits—the fundamental computing unit for these systems—is uncertain. Quantinuum's H1 systems, developed in collaboration with Honeywell and implemented by Riken, employ trapped-ion quantum computing. These systems use electromagnetic fields to suspend charged particles in free space, with qubits stored in the electric state of each ion.

According to the H1 datasheet, each system can handle up to 20 trapped ion qubits, capable of moving between five intentional zones where quantum operations occur using lasers.

While 20 qubits may seem limited, especially when compared to competing systems like IBM's Osprey, which claims over 400 qubits, it's essential to note that qubit quantity doesn't necessarily correlate with higher performance. Similar to processor cores, the count alone doesn't indicate the actual computational capacity. This is why IBM has shifted focus to building lower qubit-count quantum processors that can scale out, as seen in its Quantum-2 systems.

Riken's collaboration with Quantinuum is not its first venture into the quantum realm. In October, Riken installed Japan's first superconducting quantum computer, developed by long-time partner Fujitsu, at the RQC-Fujitsu Collaboration Center in Wako. This system incorporates 64 superconducting qubits into a single integrated system, boasting 264 quantum superposition and entanglement states. Riken claims this enables calculations on a scale too challenging for classical computers.

Neither Quantinuum nor Fujitsu's systems are designed to operate independently. Instead, Riken aims to expedite code development capable of leveraging quantum computing as an accelerator for traditional supercomputers, akin to the current usage of GPUs as accelerators.

Mitsuhisa Sato, deputy director of the Riken Center for Computational Science, emphasized that advanced quantum computers are transitioning into the practical stage, with increasing qubit numbers and improved fidelity. From an HPC perspective, quantum computers act as devices accelerating scientific applications conventionally executed on supercomputers and enabling computations currently beyond the reach of supercomputers.

Despite the progress, the consensus is that practical applications of quantum computing are still years away. In October, Fujitsu cautioned that a fault-tolerant system generating reliable results is likely a decade or more in the future. Nevertheless, various companies, including Toyota, Hyundai, BBVA, BASF, and ExxonMobil, continue to invest in quantum computing and related technologies with the hope that significant breakthroughs may occur before the predicted timeline.

IBM's announced new plateau for quantum computing packs in two particular breakthroughs that occurred in 2023. One breakthrough relates to a groundbreaking noise-reduction algorithm (Zero Noise Extrapolation, or ZNE) which we covered back in July – basically a system through which you can compensate for noise. For instance, if you know a pitcher tends to throw more to the left, you can compensate for that up to a point. There will always be a moment where you correct too much or cede ground towards other disruptions (such as the opponent exploring the overexposed right side of the court). This is where the concept of qubit quality comes into account – the more quality your qubits, the more predictable both their results and their disruptions and the better you know their operational constraints – then all the more useful work you can extract from it. The other breakthrough relates to an algorithmic improvement of epic proportions and was first pushed to Arxiv on August 15th, 2023. Titled “High-threshold and low-overhead fault-tolerant quantum memory,” the paper showcases algorithmic ways to reduce qubit needs for certain quantum calculations by a factor of ten. When what used to cost 1,000 qubits and a complex logic gate architecture sees a tenfold cost reduction, it’s likely you’d prefer to end up with 133-qubit-sized chips – chips that crush problems previously meant for 1,000 qubit machines. Enter IBM’s Heron Quantum Processing Unit (QPU) and the era of useful, quantum-centric supercomputing.

The Quantum Roadmap at IBM’s Quantum Summit 2023

Image 1 of 2

 

 

 

IBM Quantum Summit 2023 - Heron, Quantum System Two, and the Next Ten Years

 

 

IBM's roadmap pre-Heron success.

 

The two-part breakthroughs of error correction (through the ZNE technique) and algorithmic performance (alongside qubit gate architecture improvements) allow IBM to now consider reaching 1 billion operationally useful quantum gates by 2033. It just so happens that it’s an amazing coincidence (one born of research effort and human ingenuity) that we only need to keep 133 qubits relatively happy within their own environment for us to extract useful quantum computing from them – computing that we wouldn’t classically be able to get anywhere else.

The “Development” and “Innovation” roadmap showcase how IBM is thinking about its superconducting qubits: as we’ve learned to do with semiconductors already, mapping out the hardware-level improvements alongside the scalability-level ones. Because as we’ve seen through our supercomputing efforts, there’s no such thing as a truly monolithic approach: every piece of supercomputing is (necessarily) efficiently distributed across thousands of individual accelerators. Your CPU performs better by knitting and orchestrating several different cores, registers, and execution units. Even Cerebra’s Wafer Scale Engine scales further outside its wafer-level computing unit. No accelerator so far – no unit of computation - has proven powerful enough that we don’t need to unlock more of its power by increasing its area or computing density. Our brains and learning ability seem to provide us with the only known exception.

IBM’s modular approach and its focus on introducing more robust intra-QPU and inter-QPU communication for this year’s Heron shows it’s aware of the rope it's walking between quality and scalability. The thousands of hardware and scientist hours behind developing the tunable couplers that are one of the signature Heron design elements that allow parallel execution across different QPUs is another. Pushing one lever harder means other systems have to be able to keep up; IBM also plans on steadily improving its internal and external coupling technology (already developed with scalability in mind for Heron) throughout further iterations, such as Flamingo’s planned four versions which still “only” end scaling up to 156 qubits per QPU. 

Considering how you're solving scalability problems and the qubit quality x density x ease of testing equation, the ticks - the density increases that don't sacrifice quality and are feasible from a testing and productization standpoint - may be harder to unlock. But if one side of development is scalability, the other relates to the quality of whatever you’re actually scaling – in this case, IBM’s superconducting qubits themselves. Heron itself saw a substantial rearrangement of its internal qubit architecture to improve gate design, accessibility, and quantum processing volumes – not unlike an Intel tock. The planned iterative improvements to Flamingo's design seem to confirm this.

Utility-Level Quantum Computing

There’s a sweet spot for the quantum computing algorithms of today: it seems that algorithms that fit roughly around a 60-gate depth are complex enough to allow for useful quantum computing. Perhaps thinking about Intel’s NetBurst architecture with its Pentium 4 CPUs is appropriate here: too deep an instruction pipeline is counterproductive, after a point. Branch mispredictions are terrible across computing, be it classical or quantum. And quantum computing – as we still currently have it in our Noisy Intermediate-Scale Quantum (NISQ)-era – is more vulnerable to a more varied disturbance field than semiconductors (there are world overclocking records where we chill our processors to sub-zero temperatures and pump them with above-standard volts, after all). But perhaps that comparable quantum vulnerability is understandable, given how we’re essentially manipulating the essential units of existence – atoms and even subatomic particles – into becoming useful to us.

Useful quantum computing doesn’t simply correlate with an increasing number of available in-package qubits (announcements of 1,000-qubit products based on neutral atom technology, for instance). But useful quantum computing is always stretched thin throughout its limits, and if it isn’t bumping against one fundamental limit (qubit count), it’s bumping against another (instability at higher qubit counts); or contending with issues of entanglement coherence and longevity; entanglement distance and capability; correctness of the results; and still other elements. Some of these scalability issues can be visualized within the same framework of efficient data transit between different distributed computing units, such as cores in a given CPU architecture, which can themselves be solved in a number of ways, such as hardware-based information processing and routing techniques (AMD’s Infinity Fabric comes to mind, as does Nvidia's NVLink).

This feature of quantum computing already being useful at the 133-qubit scale is also part of the reason why IBM keeps prioritizing quantum computing-related challenges around useful algorithms occupying a 100 by 100 grid. That quantum is already useful beyond classical, even in gate grids that are comparably small to what we can achieve with transistors, and points to the scale of the transition – of how different these two computational worlds are.

Then there are also the matters of error mitigation and error correction, of extracting ground-truth-level answers to the questions we want our quantum computer to solve. There are also limitations in our way of utilizing quantum interference in order to collapse a quantum computation at just the right moment that we know we will obtain from it the result we want – or at least something close enough to correct that we can then offset any noise (non-useful computational results, or the difference of values ranging between the correct answer and the not-yet-culled wrong ones) through a clever, groundbreaking algorithm.

The above are just some of the elements currently limiting how useful qubits can truly be and how those qubits can be manipulated into useful, algorithm-running computation units. This is usually referred to as a qubit’s quality, and we can see how it both does and doesn’t relate to the sheer number of qubits available. But since many useful computations can already be achieved with 133-qubit-wide Quantum Processing Units (there’s a reason IBM settled on a mere 6-qubit increase from Eagle towards Heron, and only scales up to 156 units with Flamingo), the company is setting out to keep this optimal qubit width for a number of years of continuous redesigns. IBM will focus on making correct results easier to extract from Heron-sized QPUs by increasing the coherence, stability, and accuracy of these 133 qubits while surmounting the arguably harder challenge of distributed, highly-parallel quantum computing. It’s a one—two punch again, and one that comes from the bump in speed at climbing ever-higher stretches of the quantum computing plateau.

But there is an admission that it’s a barrier that IBM still wants to punch through – it’s much better to pair 200 units of a 156-qubit QPU (that of Flamingo) than of a 127-qubit one such as Eagle, so long as efficiency and accuracy remain high. Oliver Dial says that Condor, "the 1,000-qubit product", is locally running – up to a point. It was meant to be the thousand-qubit processor, and was a part of the roadmap for this year’s Quantum Summit as much as the actual focus, Heron - but it’s ultimately not really a direction the company thinks is currently feasible.

IBM did manage to yield all 1,000 Josephson Junctions within their experimental Condor chip – the thousand-qubit halo product that will never see the light of day as a product. It’s running within the labs, and IBM can show that Condor yielded computationally useful qubits. One issue is that at that qubit depth, testing such a device becomes immensely expensive and time-consuming. At a basic level, it’s harder and more costly to guarantee the quality of a thousand qubits and their increasingly complex possibility field of interactions and interconnections than to assure the same requirements in a 133-qubit Heron. Even IBM only means to test around a quarter of the in-lab Condor QPU’s area, confirming that the qubit connections are working.

But Heron? Heron is made for quick verification that it’s working to spec – that it’s providing accurate results, or at least computationally useful results that can then be corrected through ZNE and other techniques. That means you can get useful work out of it already, while also being a much better time-to-market product in virtually all areas that matter. Heron is what IBM considers the basic unit of quantum computation - good enough and stable enough to outpace classical systems in specific workloads. But that is quantum computing, and that is its niche.

The Quantum-Centric Era of Supercomputing

Heron is IBM’s entrance into the mass-access era of Quantum Processing Units. Next year’s Flamingo builds further into the inter-QPU coupling architecture so that further parallelization can be achieved. The idea is to scale at a base, post-classical utility level and maintain that as a minimum quality baseline. Only at that point will IBM maybe scale density and unlock the appropriate jump in computing capability - when that can be similarly achieved in a similarly productive way, and scalability is almost perfect for maintaining quantum usefulness.

There’s simply never been the need to churn out hundreds of QPUs yet – the utility wasn’t there. The Canaries, Falcons, and Eagles of IBM’s past roadmap were never meant to usher in an age of scaled manufacturing. They were prototypes, scientific instruments, explorations; proofs of concept on the road towards useful quantum computing. We didn’t know where usefulness would start to appear. But now, we do – because we’ve reached it.

Heron is the design IBM feels best answers that newly-created need for a quantum computing chip that actually is at the forefront of human computing capability – one that can offer what no classical computing system can (in some specific areas). One that can slice through specific-but-deeper layers of our Universe. That’s what IBM means when it calls this new stage the “quantum-centric supercomputing” one.

Classical systems will never cease to be necessary: both of themselves and the way they structure our current reality, systems, and society. They also function as a layer that allows quantum computing itself to happen, be it by carrying and storing its intermediate results or knitting the final informational state – mapping out the correct answer Quantum computing provides one quality step at a time. The quantum-centric bit merely refers to how quantum computing will be the core contributor to developments in fields such as materials science, more advanced physics, chemistry, superconduction, and basically every domain where our classical systems were already presenting a duller and duller edge with which to improve upon our understanding of their limits.

Quantum System Two, Transmon Scalability, Quantum as a Service

However, through IBM’s approach and its choice of transmon superconducting qubits, a certain difficulty lies in commercializing local installations. Quantum System Two, as the company is naming its new almost wholesale quantum computing system, has been shown working with different QPU installations (both Heron and Eagle). When asked about whether scaling Quantum System Two and similar self-contained products would be a bottleneck towards technological adoption, IBM’s CTO Oliver Dial said that it was definitely a difficult problem to solve, but that he was confident in their ability to reduce costs and complexity further in time, considering how successful IBM had already proven in that regard. For now, it’s easier for IBM’s quantum usefulness to be unlocked at a distance – through the cloud and its quantum computing framework, Quiskit – than it is to achieve it by running local installations. 

Quiskit is the preferred medium through which users can actually deploy IBM's quantum computing products in research efforts – just like you could rent X Nvidia A100s of processing power through Amazon Web Services or even a simple Xbox Series X console through Microsoft’s xCloud service. On the day of IBM's Quantum Summit, that freedom also meant access to the useful quantum circuits within IBM-deployed Heron QPUs. And it's much easier to scale access at home, serving them through the cloud, than delivering a box of supercooled transmon qubits ready to be plugged and played with.

That’s one devil of IBM’s superconducting qubits approach – not many players have the will, funding, or expertise to put a supercooled chamber into local operation and build the required infrastructure around it. These are complex mechanisms housing kilometers of wiring - another focus of IBM’s development and tinkering culminating in last year’s flexible ribbon solution, which drastically simplified connections to and from QPUs.

Quantum computing is a uniquely complex problem, and democratized access to hundreds or thousands of mass-produced Herons in IBM’s refrigerator-laden fields will ultimately only require, well… a stable internet connection. Logistics are what they are, and IBM’s Quantum Summit also took the necessary steps to address some needs within its Quiskit runtime platform by introducing its official 1.0 version. Food for thought is realizing that the era of useful quantum computing seems to coincide with the beginning of the era of Quantum Computing as a service as well. That was fast.

Closing Thoughts

The era of useful, mass-producible, mass-access quantum computing is what IBM is promising. But now, there’s the matter of scale. And there’s the matter of how cost-effective it is to install a Quantum System Two or Five or Ten compared to another qubit approach – be it topological approaches to quantum computing, or oxygen-vacancy-based, ion-traps, or others that are an entire architecture away from IBM’s approach, such as fluxonium qubits. It’s likely that a number of qubit technologies will still make it into the mass-production stage – and even then, we can rest assured that everywhere in the road of human ingenuity lie failed experiments, like Intel’s recently-decapitated Itanium or AMD’s out-of-time approach to x86 computing in Bulldozer.

It's hard to see where the future of quantum takes us, and it’s hard to say whether it looks exactly like IBM’s roadmap – the same roadmap whose running changes we also discussed here. Yet all roadmaps are a permanently-drying painting, both for IBM itself and the technology space at large. Breakthroughs seem to be happening daily on each side of the fence, and it’s a fact of science that the most potential exists the earlier the questions we ask. The promising qubit technologies of today will have to answer to actual interrogations on performance, usefulness, ease and cost of manipulation, quality, and scalability in ways that now need to be at least as good as what IBM is proposing with its transmon-based superconducting qubits, and its Herons, and scalable Flamingos, and its (still unproven, but hinted at) ability to eventually mass produce useful numbers of useful Quantum Processing Units such as Heron. All of that even as we remain in this noisy, intermediate-scale quantum (NISQ) era.

It’s no wonder that Oliver Dial looked and talked so energetically during our interview: IBM has already achieved quantum usefulness and has started to answer the two most important questions – quality and scalability, Development, and Innovation. And it did so through the collaboration of an incredible team of scientists to deliver results years before expected, Dial happily conceded. In 2023, IBM unlocked useful quantum computing within a 127-qubit Quantum Processing Unit, Eagle, and walked the process of perfecting it towards the revamped Heron chip. That’s an incredible feat in and of itself, and is what allows us to even discuss issues of scalability at this point. It’s the reason why a roadmap has to shift to accommodate it – and in this quantum computing world, it’s a great follow-up question to have.

Perhaps the best question now is: how many things can we improve with a useful Heron QPU? How many locked doors have sprung ajar?

The concept of a superconductor capable of transmitting electricity without energy loss at room temperature and regular air pressure is a highly sought-after goal in materials science. Envisioned applications range from enhancing energy grid efficiency and advancing fusion energy production to accelerating quantum supercomputing and enabling high-speed transportation.

Currently, the focus on the LK-99 superconductor centers around laboratory investigations.

On July 22, South Korean physicists shared two papers on arXiv, a repository for preprint research—early-stage work that hasn't yet undergone peer review or publication. This can be likened to uploading a preliminary draft. The researchers claimed to have developed the first room-temperature superconductor, labeled LK-99, by modifying the lead-apatite structure and doping it with copper.

A key piece of evidence presented by the team was a video demonstrating the compound levitating above a magnet, a characteristic behavior of superconducting materials.

These bold assertions made a substantial impact within the scientific community.

Xiaolin Wang, a materials scientist at the University of Wollongong in Australia, remarked, "The chemicals are so cheap and not hard to make. This is why it is like a nuclear bomb in the community."

However, the events unfolding in the South Korean laboratory represent just an initial step in ascertaining whether the findings hold practical significance for technology and its role in our lives. More data is needed, warranting caution.

Understanding Superconductors

The realization of a true room-temperature superconductor would be a momentous achievement, garnering considerable attention. Modern materials used for conducting electricity, such as copper wiring powering homes, suffer from inefficiencies. Electrons encountering atomic obstacles within the material create heat and dissipate energy as they traverse the wire. This phenomenon, known as electrical resistance, is responsible for the loss of up to 10% of electricity during transmission to homes. This energy loss also affects electronic devices.

In contrast, a superconductive material in wires and transmission lines could largely eliminate these losses. Electrons pair up as they move through the material, encountering fewer atomic obstructions, allowing them to flow unimpeded.

Superconducting materials already exist and find use in various applications, like MRI machines, globally. However, these materials demand extremely low temperatures (approaching absolute zero at around minus 459 degrees Fahrenheit) or extremely high pressures (over 100,000 times atmospheric pressure) to function.

Central Japan Railway is constructing a superconducting magnetic levitation system for high-speed travel between Tokyo and Nagoya. The SCMaglev train attains speeds of approximately 93 miles per hour with rubber wheels before transitioning to the superconducting magnetic system, capable of reaching speeds of 311 mph. This system relies on a superconducting niobium-titanium alloy cooled to minus 452 degrees Fahrenheit using liquid helium.

An LK-99 room-temperature superconductor could significantly reduce costs and eliminate the need for helium, a resource produced in only a few countries.

Skepticism Surrounding LK-99 Findings

Wang and other experts in superconductivity have expressed skepticism about the original LK-99 experiment, highlighting inconsistencies in the data. He suggests withholding hype "until more convincing experimental data are provided." Wang's team began attempting to replicate the results, but encountered difficulties in sample fabrication.

Michael Norman, a physicist at Argonne National Laboratory, offered a straightforward critique in an interview with Science magazine, describing the South Korean team as "real amateurs."

Efforts to replicate LK superconductivity have largely fallen short so far. The surge in new superconductivity experiments across various labs and individuals has turned into a burgeoning field.

LK-99 has dominated discussions on the platform formerly known as Twitter, with trends persisting for days. The trend has ventured into meme territory, with references to "floaty rocks." Unusual claims have arisen, including a shift from promoting AI investments to endorsing superconductor stocks. Shares of the American Superconductor Corporation doubled since July 27.

Even Sam Altman, the CEO of ChatGPT creator OpenAI, joined the conversation, jokingly requesting "2+ years of experience with lk-99" from recruiters.

Doubts about LK-99's legitimacy are justified. Over the years, multiple teams have asserted the discovery of room-temperature superconductors, but most of these claims have not withstood scientific scrutiny.

In 2020, a team led by Ranga Dias, a physicist at the University of Rochester in New York, published evidence of a room-temperature superconductor in the prestigious journal Nature. The article was retracted in September 2022 due to concerns about data processing and analysis. While the authors maintain their raw data supports their claims, successful replication remains elusive.

Future Prospects for LK-99

As of now, LK-99's immediate impact is likely limited, unless one wishes to delve into a physics rabbit hole on the platform. In the near future, significant developments may also be unlikely.

Replicating the LK-99 experiments is still in its early stages, and initial outcomes are not promising. Two separate research groups posted studies on arXiv on July 31 that failed to replicate the South Korean research. Chinese researchers have observed some superconductivity behaviors in minuscule LK samples, according to Wang.

Science often progresses slowly. Validating the South Korean team's work could take considerable time. Nonetheless, the excitement has prompted theoretical investigations to elucidate LK-99's properties.

Sinéad Griffin, a physicist at the Lawrence Berkeley National Laboratory, analyzed LK-99's capabilities using supercomputer simulations. Griffin's analysis, accompanied by a meme of Barack Obama dropping a microphone, was shared on X as a preprint.

Physicists responding to Griffin's work were skeptical of the mic-drop analogy and found it lacking solid evidence of superconductivity. Griffin herself clarified her results, stating they neither proved nor provided evidence of superconductivity, but did reveal intriguing structural and electronic attributes resembling high-temperature superconductors (above minus 452 degrees Fahrenheit, though far below room temperature).

Even if LK-99 ultimately proves to be a reliable superconductive material, transitioning from science to technology is a protracted process. Developing the material with consistency could take years, and Griffin's theoretical analysis suggests synthesis might be challenging.

While LK-99 may not fulfill the role of a holy grail, it remains an intriguing material, potentially leading to innovative avenues in the search for room-temperature superconductors. If it were to pave the way for a room-temperature superconductor, limitless possibilities could unfold.

Giuseppe Tettamanzi, a senior lecturer at the University of Adelaide's school of chemical engineering, highlighted the long-standing aspiration to replace copper cables in power grids with superconducting alternatives—a transition offering substantial energy savings. Tettamanzi also noted benefits for quantum computers and transportation.

"The possibilities are boundless," he remarked.

Observing science in action is exhilarating, and the fervor surrounding LK-99 offers a refreshing change on the platform. However, science requires time, and conclusions about the transformative implications of a potential superconductive material should not be rushed. Now, we await the efforts of those seeking to replicate the results.

Google-parent Alphabet (GOOGL) on Tuesday reported second-quarter earnings that topped estimates amid rising investments in artificial intelligence. GOOGL stock popped as revenue beat expectations, boosted by cloud computing and YouTube.

Reported after the market close, Google earnings came in at $1.44 a share, up 19% from the year-earlier period. The tech giant reports earnings under generally accepted accounting principles, also known as GAAP. A year earlier, Google reported earnings of $1.22 a share.

The company said gross revenue rose 7% to $74.60 billion, compared with $69.7 billion a year ago.

Analysts had predicted Google earnings of $1.34 per share on revenue of $72.84 billion.

Also, advertising revenue rose 3% to $58.14 billion vs. estimates of $57.39 billion.

GOOGL stock popped 5% in extended trading to 128.36 on the stock market today.

In addition, the company said Chief Financial Officer Ruth Porat will assume the newly created role of President and Chief Investment Officer effective Sept. 1. Porat will continue to serve as CFO through the 2024 capital planning process as the company finds a successor.

GOOGL Stock: YouTube Tops Estimates

Meanwhile, YouTube ad revenue rose 4% to $7.66 billion during the second quarter. Analysts had estimated YouTube ad revenue of $7.42 billion. TikTok's growth has pressured YouTube.

Google said cloud computing revenue rose 28% to $8.03 billion vs. estimates of $7.87 billion. Further, Google repurchased $14.97 billion of its own stock during the quarter.

The Big Tech stock had advanced 37% in 2023. Also, GOOGL stock holds a Relative Strength Rating of 81 out of a best-possible 99, according to IBD Stock Checkup.

Quantum computers are highly complex and extremely sensitive to external disturbances or "noise" such as heat, magnetic fields, cosmic radiation, and stray light. These factors disrupt the delicate quantum environment, leading to errors and de-coherence of the quantum state. To minimize errors, a significant portion of quantum computers is dedicated to protecting qubits, allowing the quantum state to persist as long as possible.

Error correction in quantum computing is a major challenge. While error protection is common in everyday technologies like telecoms and data centers, quantum error correction is much more complex and likened to juggling with loose soot while trying to herd cats. Logical qubits, sets of physical qubits operating together, offer a potential solution, but constructing and managing them is extremely difficult.

A single logical qubit might require around 1,000 or more physical qubits, with most dedicated to identifying and correcting errors in real-time, leaving only a few qubits for computational processing. This leads to significant overhead and energy consumption.

Efforts to overcome the problem of quantum error correction are ongoing, as is the race to scale up quantum computers to thousands of qubits while maintaining high coherence and minimizing error rates. Additionally, with quantum and classical computers coexisting, the focus is on optimizing data transfer between the two technologies, enabling practical, complementary, and compatible applications.

To achieve this, standards and protocols for hardware, software, applications, and communication interfaces need to be developed to facilitate interoperability between different quantum computing platforms. Benchmarking standards will also be essential for measuring and comparing the performance of quantum computers.

Lawrence Gasman, a quantum computing expert, highlighted several other challenges in a recent interview. One major hurdle is the lack of a unified approach to developing scalable, fault-tolerant qubit control technology, given the diversity of qubit technologies used in quantum computing. The software side also faces difficulties, with the need to develop new programming languages and compilers and the infancy of quantum algorithms.

Furthermore, the quantum computing field suffers from a shortage of trained quantum scientists and engineers globally, adding to the immense costs of the enterprise. However, despite these challenges, Gasman remains optimistic about the increasing number of applications emerging from quantum computing.

He cited drug discovery, materials design, quantum chemistry, and financial services as sectors where quantum computers are making progress. As quantum devices move from hundreds to thousands of qubits, they are capable of handling highly advanced tasks, offering optimal solutions through simulations of billions of system scenarios.

Gasman believes that as the cost of quantum computing decreases, the technology will become more accessible to smaller organizations. Eventually, we may see the emergence of end-user mini-quantum computers, heralding a transformative era much sooner than anticipated.

The flaw exploits a misprediction recovery issue in Zen 2 processors during speculative execution, which results in data being leaked from CPU registers. One alarming aspect of this bug is that attackers don't need physical hardware access to exploit it; merely loading malicious JavaScript on a website could trigger the attack. While no known exploits in the wild have been reported yet, there is a risk of potential exploitation following the public disclosure.

AMD acknowledges the vulnerability but claims no known exploits have occurred outside research environments. Cloudflare also affirms no evidence of exploitation on its servers.

Addressing the issue requires firmware updates from AMD, which will fully fix the problem. Alternatively, a software update is possible, though it may come with some performance impact. The bug affects various Ryzen desktop and laptop processors, EPYC 7002-series server chips, and Threadripper 3000-series and 3000 Pro WX-series CPUs.

To safeguard systems, AMD has already issued a firmware update for EPYC 7002 chips, as they are more lucrative targets for hackers. AMD warns that the performance impact of the update will depend on the workload and system configuration.

The timeline for patches varies based on the processor and system configuration. Ryzen desktop processors, including Ryzen 3000 and Ryzen 4000G-series chips, are expected to receive a firmware fix by December, distributed through motherboard or PC manufacturers.

Laptops with Ryzen 4000-series CPUs will get an update in November, while Ryzen 5000-series laptops and Ryzen 7020-series budget CPUs using Zen 2 will receive an update in December. For Threadripper systems, a patch is expected in October, and fixes for Threadripper Pro 3000WX-series systems will arrive in November and December.

Key Differences Between Edge and Cloud Computing

Both edge and cloud computing are forms of remote computing, where computing resources are utilized from a location separate from the user. While this definition appears straightforward, the intricacies are far more complex. For example, remote workers requiring access to business systems require different resources compared to Internet of Things (IoT) devices that necessitate real-time data processing. This is where the fundamental distinctions between cloud and edge computing come into play.

Cloud computing is better suited for scenarios involving vast amounts of data processing. Conversely, edge computing is more adept at processing smaller volumes of data but in real time.

This simplified explanation highlights the contrast between these two models of remote computing. To gain a deeper understanding, let's analyze some of the metrics that define cloud and edge computing:

Edge and Cloud Computing in Action

The unique characteristics of each model determine their suitability for various use cases. Understanding the scenarios in which each excels is the simplest way to comprehend the difference between these two approaches to remote computing.

While there are areas where the two methodologies overlap, in general, they provide distinct services.

Cloud Computing Use Cases

Cloud computing offers numerous benefits and is primarily employed in situations where massive amounts of data are stored, accessed, and managed from a centralized location. Some scenarios that make cloud computing the preferred choice include:

Data analytics: The era of big data necessitates the use of cloud computing for analyzing vast datasets.

Remote working: Cloud-based services play a vital role in the shift toward remote and hybrid work models, enabling workers to access resources from anywhere with an internet connection.

Software as a Service (SaaS): The rise of the SaaS model is largely facilitated by cloud computing, enabling convenient software purchase and usage.

Disaster recovery and backups: Cloud systems often serve as backup and disaster recovery solutions. For instance, the images stored on your phone can be backed up on a cloud-based system, ensuring their safety even if you lose or change your device.

These use cases share the common requirement of managing and processing large amounts of data. While real-time processing is possible, it is not a core characteristic of cloud computing.

Edge Computing Use Cases

Edge computing is better suited for real-time processing of smaller data volumes. It is targeted at scenarios that demand minimal latency and immediate actions. Common applications for edge computing include:

Internet of Things (IoT): IoT devices are increasingly prevalent, from smart homes to smart cities, and often necessitate real-time data processing, which edge computing provides.

Gaming: Gamers have likely experienced frustration due to in-game lag. Edge computing, with its low-latency, "edge" processing, and real-time data processing capabilities, is an ideal choice for mitigating lag issues. Games like Pokémon Go, which rely on real-time player data, are prime examples.

Streaming content: Edge computing is used to address buffering and lag problems in the streaming industry.

Augmented and virtual reality: Applications utilizing augmented or virtual reality require access to real-time data processing for seamless immersive experiences.

Edge computing is the preferred solution when low-latency data access is crucial.

The Future of Cloud and Edge Computing

Precisely predicting the future of these technologies is challenging. However, the rapid adoption of remote working practices, IoT, and AI will undoubtedly shape the evolution of these forms of remote computing.

Several aspects should be considered when discussing their future:

Cloud computing: As more organizations transition to remote work and harness the power of big data, cloud computing will continue to expand.

Edge computing: The growth of IoT and the need for real-time data processing are driving the advancement of edge computing. As more devices become internet-enabled and generate data, the demand for efficient and rapid processing through edge computing will increase.

Hybrid models: Ultimately, the boundaries between these technologies will blur, leading to prevalent hybrid models that leverage the advantages of both cloud and edge computing.

Forecasting the future is always a speculative endeavor, but there is little doubt that both cloud and edge computing will continue to progress rapidly.

Head in the Clouds or Life on the Edge

The rise of remote computing in all its forms ensures the longevity of these technologies. Both cloud and edge computing possess strengths and weaknesses that determine their respective applications.

However, the future is likely to embrace hybrid models that combine the scalability and data processing capabilities of cloud computing with the low-latency and real-time processing capabilities of edge computing.