A visit to IBM’s famous labs in upstate New York reveals a company attempting to reinvent itself with an R&D focus and ambitious bets on emerging technology.
The catchphrase in the IT departments of corporate America in the 1980s was that “you never get fired for buying IBM.” But the rise of personal computers and the internet in the 1990s signaled a period of decline for one of the country’s most respected technology providers.
So tied was IBM to its lucrative mainframe computer business and associated IT services that it failed to capitalize on its early role in the personal computer revolution. More recently, it was a late starter in both web services and cloud computing, allowing Amazon Web Services and Microsoft to establish themselves as the go-to platforms for millions of businesses migrating their applications and data online.
The outcome for Big Blue has been a long period of declining revenue and restructuring. They divested the computer hardware division behind the iconic “Thinkpad” laptops and the IT services business, which accounted for a quarter of IBM’s $55.2 billion in revenue in 2020 but was spun off in 2021 into a separate company called Kyndryl.
Hybrid Cloud or Bust
IBM is in the midst of reinventing itself under new CEO Arvind Krishna and it seems to be working. IBM outperformed many other large listed tech companies in terms of market performance last year and beat analysts’ revenue forecasts for the 4th quarter.
With the purchase of Red Hat in 2018 for $34 billion, IBM focused on being a hybrid cloud provider to those businesses that want to maintain some on-premises infrastructure while sending applications and workloads to cloud platforms as well. This means that IBM is no longer positioning itself in opposition to the large public cloud players AWS, Microsoft and Google. It will accommodate any cloud configuration a customer requires.
The Red Hat suite of enterprise open source software solutions, including the Kubernetes container orchestration system for automating software deployment, increasingly looks like a shrewd purchase by IBM.
While IBM isn’t mentioned in the same breath as AWS, Apple, Meta or Microsoft, its legacy as one of the most prolific investors in R&D could aid its turnaround efforts. On a wall in IBM’s Poughkeepsie labs in New York, there’s a charter that features in small print the thousands of patents the company secured in the previous year. IBM has held the record for most patents secured at the US Patents Office for 28 years running. The question is whether all of that cutting-edge R&D can translate into a market advantage.
Here are five areas of technology that I explored during a three-day trip to two IBM labs, the Thomas J. Watson Research Center in Yorktown Heights, NY and its IBM Z Systems Lab in Poughkeepsie, NY. They offered insights into Big Blue’s future vision and implications for software development outsourcing.
1. In Search of Quantum Advantage
Few companies have invested as heavily in quantum computing as IBM, which has one of the world’s most extensive clusters of operational quantum computers at its Yorktown Heights, New York laboratories. It’s fair to say that IBM has bet the future of the company on quantum.
In a room near the retro building’s lobby sits IBM’s Quantum System One, the world’s first integrated quantum computer system. It’s encased in thick glass and emits pneumatic hissing noises. Quantum computers are incredibly sensitive machines and need to be chilled with gasses to super-cold temperatures just above absolute zero to operate effectively. As such, you aren’t going to be buying a quantum computer any time soon.
A quantum computer denuded of its outer shell at IBM Research Labs, New York
Quantum computers are also completely different from the classical computers that the world currently uses. They exploit quantum physics principles such as superposition and entanglement to simultaneously allow particles to be multiple states. This moves us beyond the ‘binary’ world, exponentially increasing the computing power available to us.
We’re entering the era of "quantum advantage," where this new generation of computers is capable of performing in minutes what our best supercomputers would take hundreds of years to calculate. That has serious implications for how complex mathematical problems are solved, but also the integrity of the encryption that secures our data and internet connections (see below).
Out of its vacuum-sealed case, the quantum computer looks like an ornate chandelier, with one of IBM’s quantum processors manufactured on site at Yorktown, sitting amidst delicate metal tubing and gold-plated platforms - the better to avoid background interference.
IBM, Microsoft and Google are already offering quantum computing as a cloud service, where you rent time on their futuristic computers to run workloads. IBM has 400,000 registered users for its quantum cloud service and currently runs two billion circuit executions every day - mainly scientists conducting experiments.
IBM’s Q System One
Currently, quantum computers are only useful for a narrow range of applications. They are well-suited to simulating the real world, so pharmaceutical companies are experimenting with them to develop new molecules, and aircraft and automobile makers use them to explore materials science. IBM sees a major role for quantum machine learning and in the financial sector for calculating risk and probability.
They’re investing heavily in advancing quantum hardware because the larger their processing power, the more likely they will be useful for a wider range of real-world applications. The New York lab in 2021 produced the Eagle 127-qubit quantum computer (a qubit is the equivalent of a bit in classical computing), the world’s first 100-plus qubit computer. In November, IBM unveiled Osprey, a 400+ qubit quantum processor. Using multiple quantum processors running in parallel, as we do with classical computers, IBM expects to have the 4,158-qubit Kookaburra quantum computer available by 2025.
Still, a larger quantum computer with more qubits is only part of the solution. Equally important is reducing the error rate which remains significant for quantum computers. This “quantum noise” in computation limits their usefulness, which is why the real goal for IBM and its rivals is a high-capacity error-corrected quantum computer.
Implications for Software Outsourcing: Very few companies will actually own a quantum computer given their delicate nature and the significant cost of maintaining them. But in the coming years, many businesses will find diverse uses for quantum computers, splitting up demanding jobs between quantum and classical computers.
The reality is that we need to build a quantum workforce to develop new algorithms to run on these quantum computers. Several software development tools, such as IBM’s Qiskit, have been released to help developers gear up for the quantum world (Google Tensorflow Quantum and AWS Braket are other quantum software developer kits now available). Software development companies with capability in quantum computing algorithms and applications will have a competitive edge in the second half of this decade as high-qubit machines become more broadly available and start to change what is possible from computers.
2. AI Chatbots and Your Digital Sidekick
All the buzz in the tech world at the moment is about ChatGPT, the artificial intelligence chatbot from OpenAI, which has spurred a tech arms race between Microsoft and Google to lead the field in AI.
Microsoft has integrated ChatGPT into its Bing search engine and Edge web browser, while Google is preparing to release Bard, its ChatGPT equivalent based on LaMDA, its own large language model.
But IBM has been producing intelligent chatbots for years, based on its Watson artificial intelligence platform. Watson is perhaps best-known for its high-profile appearances against human contestants on the quiz show Jeopardy!. Just prior to lockdown, I watched Watson take on a world champion orator in a debate in front of a live audience in San Francisco. Watson narrowly lost the debate, but the way it formulated logical arguments and delivered them in a convincing speech was incredible to watch.
What’s the difference between ChatGPT and Watson? OpenAI aimed for a general AI chatbot with ChatGPT that excels in natural language processing to produce convincing conversations, text generation and written responses to answers.
IBM has aimed for something different, using Watson to assist with complex decision making across banking, healthcare, science and education. Watson is more tailored to the specific needs of the organizations using it. IBM’s head start in developing domain-specific models that help with problem solving in organizations means it will have an edge in the corporate world as the masses adopt ChatGPT. But IBM also faces formidable competitors in OpenAI, Microsoft and Google and will need to accelerate its AI development efforts to stay competitive.
IBM saw a major role for it in transforming healthcare by using AI to help determine patient treatments and tackle diseases like cancer by speeding up scientific progress. But in 2021, it sold off Watson Health for just US$1 billion, frustrated that it was unable to access useful health-related data in sufficient quantities to train its systems for healthcare applications.
Instead, Watson is being refocused on automation and AI-powered decision making inside large organizations. At Yorktown, I saw a demo of Watson Orchestrate, an application that effectively creates a digital sidekick for an employee, taking care of low-level admin tasks and learning as it goes. The example given was HR Sidekick, a Watson application that streamlines the process of making a new hire by collating resumes, setting up interviews and seeking the right approval from the relevant managers.
HR Sidekick plugs into apps such as Outlook, Slack and Asana, creating updates when it has completed tasks and leaving reminders for its human colleagues. Watson Orchestrate joins a host of platforms in the robotic process automation space, but is aiming to be effective for less-structured tasks that automate invoice or payroll processing, for instance. IBM wants Watson to become a digital buddy for the average office worker, increasing productivity by automating mundane tasks and making AI-driven suggestions to keep work on track.
Implications for Software Outsourcing: Automation of tasks in every business will accelerate with the uptake of AI-powered tools such as IBM Watson, ChatGPT and Google Bard. These tools will increasingly undertake low-level software coding tasks, freeing up developers to work on the more valuable aspects of projects that only humans can handle. Because automation tools are going mainstream, everyone will be able to take advantage of them to boost productivity. The competitive edge will therefore come from having skilled and experienced outsourcing partners who are best able to take advantage of these tools.
IBM has designed Watson Orchestrate to be configured by users with nothing required in the way of software programming. But to have deeper integration with the tech stack an organization relies on, including proprietary applications, developers will increasingly be called on to bridge the gap. Large software providers such as Salesforce have developed interfaces for Watson, allowing customer and sales data to be drawn on by a digital sidekick.
The rise of AI chatbots may automate tasks, costing some white collar jobs in the process, but are likely to create far more new types of roles overall. Software developers will have their hands full with interoperability projects as organizations seek to tap Watson’s brain for task automation at a more fundamental level across the business.
3. The Mainframe Lives!
In a massive room at IBM’s Poughkeepsie campus sit around 200 mainframe computers, probably the largest collection in one place anywhere in the world. Here, IBM technicians test hardware and applications on the company's Z-series of mainframes, which are used by banks, credit card companies, government departments and other large businesses all over the world.
During my visit, the team was preparing to start shipping the z16, IBM’s newest mainframe computer. IBM claims that “two- thirds of the Fortune 100, 45 of the world’s top 50 banks, 8 of the 10 insurers, 7 of the top 10 global retailers and 8 out of the top 10 telcos” rely on IBM mainframes. That equates to “70% of global transactions, on a value basis,” according to IBM.
IBM z16 mainframes being tested at IBM’s Poughkeepsie, NY labs
We’ve heard a lot about the rise of hyperscale public cloud providers and private data centers stacked with Intel-powered servers. But the reality is that the modern mainframe is still the backbone of large organizations that require minimum downtime and maximum security for processing high-volume financial transactions.
These days, a mainframe doesn’t look much different to a regular server. Its components have shrunk down just as its processing capabilities have increased. IBM makes its own Telum processor to power its mainframes, the latest processor coming with an on-chip AI accelerator for analyzing real-time transactions. The Telum roadmap allows for faster 2nm (nanometer) computer chips, which will serve to keep the mainframe relevant as more demanding applications are required. The z16 has also been built with an eye to the quantum world, billed as the industry’s first quantum-safe system.
Quantum-safe: As IBM points out, while encrypted data may not be accessible now to cybercriminals, powerful quantum computers could be drawn on to unlock that sensitive data in the future. A whole industry has therefore sprung up involving hacking encrypted data sets now to put aside to be decrypted in the future.
IBM claims the z16 is “underpinned by lattice-based cryptography, an approach for constructing security primitives that helps protect data and systems against current and future threats.” It basically means that IBM has made it incredibly hard to get malware onto its mainframes, but also encrypts data in ways that make it next to impossible to decrypt if it falls into the wrong hands, even if quantum computers can be applied to the task.
Implications for Software Outsourcing: Developing applications to run on mainframe computers in industries ranging from banking to healthcare is a high-value subset of software development. While the cloud computing giants have captured the narrative in recent years and indeed, hyperscale data centers are processing more and more data and hosting applications, there’s still plenty of life left in the mainframe business.
Cybersecurity experts who are proactive in developing quantum-safe systems will enjoy a competitive advantage as organizations big and small overhaul their IT infrastructure to deal with the threat of ‘steal now, decrypt later’ attacks.
4. Application on the Edge
One of my meetings at IBM’s Yorktown labs was disturbed by the sound of mechanical footsteps pounding their way up the hallway outside. A yellow robotic dog called Spot, the brainchild of Boston Dynamics, walked through the door, its sensors taking in the room around it.
IBM has been working with Boston Dynamics to develop an AI-powered edge computing platform that literally sits on Spot’s back. The computer uses cameras and sensors to gather information around Spot and process the data in real-time without any need to use a cloud computing platform.
Boston Dynamic’s Spot robot with IBM’s edge computing device, cameras and sensors
The use case IBM is pursuing is deploying Spot around industrial sites, such as natural gas plants or automobile factories, to take human workers out of harm’s way. Spot’s cameras can read a gauge and listen to the sound of machinery to detect faults. An additional sensor could even let it detect gas or chemical leaks. Spot also has a strong use-case in providing perimeter security on large premises. However, IBM is careful to point out that it has withdrawn from using facial recognition technology to identify people.
Instead, more straightforward video footage or heat-sensing technology could identify the presence of people on a property and send an alert to security. The AI analyzes the sensor and camera information to make decisions, such as alerting a manager if a gauge reading is unusual.
Spot costs around $75,000 before the cost of adding IBM’s technology on its back. But if one or two robots can cover a site, returning to their charging docks to replenish their batteries automatically, it can prove a cost-effective alternative to teams of humans performing monotonous and potentially dangerous safety checks.
Implications for Software Outsourcing: Staff shortages, Covid restrictions and health and safety regulations pose problems for businesses that need to perform regular safety and security checks on plant, equipment and premises. This has opened up opportunities for edge computing devices in the form of robots, drones and remote sensing devices. Software developers with an understanding of edge computing applications can leverage this technology for a wide range of use cases integrating with platforms such as Maximo, IBM’s enterprise asset management software.
5. AI for the Environment
With sustainability issues climbing higher on the corporate agenda, organizations have to figure out how to get a baseline snapshot of their current carbon emissions and monitor progress towards reducing them over time.
Any company without an ESG (environmental, sustainability, governance) strategy in place that doesn’t include concrete goals for emissions reductions and credible processes for evaluating progress, now risks missing out on commercial contracts. IBM itself has a plan to reach net- zero greenhouse gas emissions by 2030.
The problem is that accurately calculating emissions across large and complex businesses is very difficult, particularly when it comes to “Scope 3” emissions that take into account an organization’s entire supply chain. A host of software vendors has entered the fray, trying to make the job easier. Microsoft’s Sustainability Cloud is one such product, another is IBM’s Envizi, which is based on the software developed by the company Envizi, which IBM acquired in January.
IBM’s Envizi service helps businesses track their emissions
Envizi automates the collection of over 500 types of data and supports major sustainability reporting frameworks. It also is compatible with IBM’s Environmental Intelligence Suite, which draws in information such as weather data (IBM bought the Weather Company in 2016) to give clients a real-time view of environmental factors affecting their business.
IBM says Envizi helps businesses “operationalize sustainability,” extending the life of physical assets, creating more efficient and resilient supply chains, and analyzing and reporting on ESG commitments. That’s all well and good, but such platforms are only as good as the data fed into them. A huge effort is required to bring infrastructure and systems up to standard to make the fancy dashboards and graphs Envizi produces meaningful to decision-makers.
Implications for Software Outsourcing: Software developers will be called on increasingly to develop software apps, APIs and connectors to get important data out of systems that will feed into data analytics programs for ESG purposes. With most large organizations setting net-zero emissions goals and compliance obligations tightening up, the next few years will see high demand for developers who understand the methodologies involved and can integrate with platforms such as Envizi and Microsoft’s Sustainability Cloud.
Want more tech insider information?
Check out our Insights thought leadership hub with articles, guides, studies, enewsletters and webinars with trending topics and tips.
Peter Griffin is a freelance science and technology writer who has covered how innovation is changing the world for over 20 years. Follow him on Twitter @petergnz.