Pages

Monday, April 22, 2019

IBM aims to speed international payments while reducing costs!



                                            Image courtesy of IBM, Inc. 
Blockchain technology has the potential to fundamentally change many traditional business procedures. IBM  is a major sponsor of and contributor to Blockchain technology. IBM identified an area of international finance, i.e. cross-border payments, that they believe can be disrupted by Blockchain and with digital assets. 

So, no surprise, IBM created a new platform to facilitate the creation and sending of global payments based on Blockchain, IBM World Wire. This article describes the IBM offering and our opinions on the subject. Note that we are computer industry analysts, not experts in the fintech[1] area. Nor are we particularly expert on the intricacies of cross border payments.  However, our background does give us significant understanding of technology and the difficulties of introducing new technology. Now, our thoughts on blockchain and international financial transfers. 

In the U.S. today, moving money internationally is typically done through commercial banks and can also involve several financial intermediaries, including the use of a messaging system run by  SWIFT[2], a Belgian company. Several years ago, we did a transaction that involved transferring money from Australia to the U.S. To our surprise, we discovered that the transaction was not only surprisingly expensive but would also take more than a week before the transaction was settled. We investigated alternatives with other banks and even a credit union. We were unable to find a faster, less expensive process.

We investigated further to see if this was unique to Australia. It was not. We spoke to the U.S. branch of a Canadian bank about doing a similar transaction. Still no difference. That bank also expressed that there would be a delay that exceeded a week at essentially the same cost. We were frustrated and suggested we could almost walk to Toronto[3] and get our money far more rapidly than the  banks’ process. The banker was not amused.

Let’s summarize. The current correspondent banking system is slow and expensive. You lose the use of your money for up to a week plus pay a significant fee to initiate the transfer. Another cost is incurred as the currency needs to be converted in the process; in our case from Australian to U.S. dollars. Exchange rates fluctuate, frequently and sometimes rapidly and significantly depending on the currency. When converting currencies, there is always a cost incurred to the sender, which needs to be taken into consideration to ensure one is sending the correct amount to the receiver. Typically, a bank can generate a significant amount of revenue by managing the buy/sell foreign exchange rates being charged as part of these cross-border payments.

Introducing IBM Blockchain World Wire


World Wire will transform the way banks and other Money Transfer Operators initiate cross border payments. See Figure 1. First, with the use of a digital asset being used for settlement, the transaction can be completed in virtually real time; probably only taking less than a minute in total. Thus, the money is readily available and accessible to the recipient for reinvestment or whatever the use with no time lost. Second, because of the efficiency of the platform, the transaction cost is much lower than what one would normally see from banks using the more traditional correspondent bank, SWIFT initiated transaction. IBM estimates total transaction costs would be cut by more than 50%. From our viewpoint, IBM World Wire appears to be an excellent idea that we believe may very well succeed.

Let’s compare the Blockchain process to SWIFT.  First, IBM World Wire allows point-to-point transfers to take place versus the SWIFT system’s reliance on multiple intermediary banks, which adds to the time and cost it takes to complete the transaction. Second, IBM World Wire would be more efficient and less error prone. Finally, the IBM’s system is based  on Blockchain technology, therefore IBM World Wire easily handles digital assets such as stable coins. SWIFT indicated it was studying the path to the use of Distributed Ledger technology[4] within its messaging service.   Digital assets represent a huge growth area for payments. As near as we can tell today,  the advantage goes to IBM World Wire.

What about SWIFT?


Let’s take a closer look at SWIFT as the competitor. Their website indicates that they are planning to upgrade their service to satisfy their banking customers. (NOTE: the statement appears to be a few years old. No details are included.) The large banks are significant shareholders of SWIFT. IBM’s presentation noted that payments represent more than 30% of global banking revenue. Thus, the incumbent that IBM World Wire is competing with is not only closely (financially) tied to the banks but also represents an important part of banking revenue. Therefore, it comes as no surprise that none of the largest global banks appear in IBM’s list of those involved in the World Wire project. 

Some other considerations. Generally, it is very difficult to displace a well-established incumbent based only on cost savings and better technology. This case has occurred many times in computers. There are various reasons for this; (not-so-) simple inertia is one reason. Another reason is that change typically involves risk and many organizations are very risk adverse. If the potential change risks or even threatens a significant erosion of a bank’s revenue, that is an excellent reason to delay or resist the change.

However, if the new technology company identifies a growth area where the incumbent is not strong, they have a more advantageous (level) playing field. This is a good entry-point and opportunity for a challenger to grow revenue. Several trends in the payments industry appear to provide such an opportunity for IBM. These include the entry of non-banks into payments. Also, developing countries without strong existing ties to SWIFT have significant growth potential. Europe (since the advent of the euro) itself may be tempted to explore this new technology.

A Final Word


A 2018 quote from a McKinsey[5] web site, appears to be quite germane to our discussion.


“While cutting-edge technology is exciting, it can also be complex; demand is also untested, which can result in long lead times with little opportunity to validate the business model. As an example, consider cross-border money transfer, a market that has traditionally been dominated by large incumbents such as Western Union. Despite much hype about fintech—particularly blockchain-based solutions—entering the space, no start-up has gained anywhere near the scale of TransferWise[6], a digital business built on top of traditional payments rails, rather than a reinvention using the latest tech. TransferWise used great user experience and distinctive marketing campaigns to grow rapidly, enabling it to successfully disrupt the space, and to report £117 million in revenues in March 2018”

While we are enthusiastic about the prospects for IBM World Wire, we believe IBM will have to battle for success. It may take a while for IBM World Wire to gain enough market share for the offering to become profitable. We hope that IBM is in this project for the long haul. As the reader might expect, we believe that McKinsey’s caution is justified.







[1] Fintech is a term new to us. It stands for the application of technology to financial services.
[3] We hope the reader can forgive our sarcasm.
[6] There is an interesting article on TransferWise in Wikipedia that is worth reading. https://en.wikipedia.org/wiki/TransferWise

Friday, April 19, 2019

Quantum Computing - act now for tomorrow’s advantage

By Rich Ptak

We’ve been spending more time with quantum computing (QC) vendors lately. Partially because we find QC to be fascinating and intellectually challenging and partially and somewhat more importantly, because we’ve concluded that the time for enterprises to start educating some employees on and about quantum computing is now.
The number of enterprises, institutions and individuals involved in quantum computing has increased dramatically over the last year. Enterprises that delay team development and education will be seriously handicapped when QC becomes a commercial reality. We discuss Rigetti Computing[1] one of two firms currently providing public access to real quantum computers.
There exists a global competition to capitalize on the potential power of quantum computing. This includes major efforts in Australia, Canada, China, India, Japan, and UK. (China is arguably furthest along with its $11Billion budget for a new QC research center.) In December 2018, the US Congress and President Trump authorized some $1.2B for coordinating quantum research and development efforts. This is in addition to a Department of Energy $218 Million project budget.
Yes, we know that the actual day of a fully commercialized quantum computer lies in the somewhat hazy future. And, yes it remains important to maintain focus on today’s rapidly evolving, global competition and the technical challenges and benefits to be realized with proper application of AI, blockchain, etc.
All that said, you really need to start preparing for that ‘hazy day in the future’. Because, visible progress is being made from theory toward practical application. The fact is solution providers and developers face a significant learning curve to acquire the quantum computing understanding they need to effectively leverage quantum computer systems for commercial applications. Quantum computing systems require a radical change in thinking from current DevOps in order to effectively apply quantum systems in a useful way.

What’s going on

There exists a global competition to capitalize on the potential power of quantum computing. This includes major efforts in Australia, Canada, China, India, Japan, and UK. (China is arguably furthest along with its $11Billion budget for a new QC research center.) In December 2018, the US Congress and President Trump authorized some $1.2B for coordinating quantum research and development efforts. This is in addition to a Department of Energy $218 Million project budget.
Yes, we know that the actual day of a fully commercialized quantum computer lies in the somewhat hazy future. And, yes it remains important to maintain focus on today’s rapidly evolving, global competition and the technical challenges and benefits to be realized with proper application of AI, blockchain, etc.
All that said, you really need to start preparing for that ‘hazy day in the future’. Because, visible progress is being made from theory toward practical application. The fact is solution providers and developers face a significant learning curve to acquire the quantum computing understanding they need to effectively leverage quantum computer systems for commercial applications. Quantum computing systems require a radical change in thinking from current DevOps in order to effectively apply quantum systems in a useful way. 
For now, we think Rigetti provides sufficiently significant differentiation to make them interesting and a particularly attractive partner for an introduction to quantum computing.  
Figure 1  Rigetti Quantum Cloud Service model
Rigetti Computing stands out from most of the quantum computing field for a number of reasons. Here are a few that caught our attention:
1.      Pragmatic, action-oriented focus – organized and dedicated to deliver pragmatic, productive entrepreneur- and user-oriented systems, services and solutions. Experimentation and knowledge advancement shared with and through the network of research and engineering partners with marketability the goal.
2.      Engineering and research & development operate in parallel Closer interaction allows them to more rapidly engineer and integrate research advances into the production process. Thus engineering (pragmatic, production oriented) and development (leading-edge in knowledge and access) complement and support each other. This operational model allows Rigetti to more efficiently develop and distribute custom circuits and production-oriented tools; the intent is to advance not only the knowledge and understanding of quantum computing, but to increase and facilitate access to quantum computing on its the path to full commercialization and problem resolution.
3.      Rigetti Quantum Cloud Services will help users as they develop and run their models and algorithms on Rigetti’s quantum-classical hybrid computers. They currently have 2-16 qubit computers, will increase to 32 qubits later this year and plan 128  qubits within 18 months.
4.      Training and mentorship program targeted to aid early career engineers and technicians with internship get quantum computing experience via their Junior Quantum Engineer (JQE) program.
5.      Rigetti Quantum Advantage Prize –  of $1M for developing an algorithm that when run on Rigetti’s QCS platform, will either reach a faster solution time, improve the quality of the solution or at a lower cost compared to the best classical algorithm.
Figure 2 Rigetti Full Stack capability
Rigetti’s success is seen in the demonstrable evolution of full-stack systems, strong support services and a growing library of shared apps for solution development. Figure 2. Their In-house foundry facilitates rapid (4-6 week) turn-around (design-to-delivery) of customized integrated circuit boards.
Rigetti is one of only two vendors providing widespread direct public access to their quantum systems. Their unique access model gives users exclusive real-time access to Rigetti’s quantum-classical system. Rigetti users schedule sessions on quantum-classical systems. Users load their experiments/apps onto classical computers co-located with the quantum system; then have dedicated, uninterrupted access to the quantum processor during their scheduled session. Notable also because the job is not queued nor submitted remotely; but run as a dedicated real-time job. This allows more experiments or iterations of experiments to be run. Local access also eliminates remote access induced delays. Rigetti support service teams are available to work with customers if they desire.
One more observation, we found Rigetti documentation and writing to be clear and concise. Of course, there are deeply detailed, highly technical articles which communicate experimental details with great precision. Most of which is distinctly unhelpful during the very early part of an introduction to aspiring developers or engineers interested in learning about quantum computing. Rigetti staff understand the necessity for clear, concise communication and they deliver just that. Kudos to them. 

Recommendations

Significant effort and investments in quantum computing are currently underway worldwide. All manner of institutions, aerospace, energy, even retailers have organized teams at work.  Through networks, partnerships and cooperative arrangements, these  teams are working to develop applications usable in chemistry (modeling), financial (risk analysis), medicine, artificial intelligence(AI)/machine learning (ML) as they work to optimize existing and develop quantum unique algorithms.
 Finally, the reality of global competition increases the importance of technological leadership in quantum computing. Those who lag in quantum knowledge and expertise will find themselves at a significant competitive disadvantage. Multiple efforts are underway to accelerate the maturation of application development tools beyond esoteric physics and quantum mechanics for use by a wider audience.
It is critical to act today to facilitate the growth of a larger, dispersed base of quantum knowledgeable and quantum-literate staff. Some will be early adopters while others will actually pioneer the earliest at commercialization of a quantum advantage. Those that put off acquiring any knowledge in this area, will find themselves at a serious competitive disadvantage when full quantum commercialization arrives.


[1] Find more info on Rigetti here: https://www.rigetti.com/

Wednesday, April 10, 2019

Intel targets IBM as they battle for supercomputing supremacy


A recent rash of articles appearing in business and technical media have raised the visibility of supercomputers, and vendors involved in the competition. The articles shared some common characteristics chiefly being they provided little, if any, new information, or even specific product data. The articles were either revisiting well-known architectural details, speculating on the use of alternative technologies (without specifying details) or discussing areas of potential benefits in application scenarios. A fair amount focused on the benefits brought to supercomputing by use of AI, Machine Learning and Nvidia GPUs. We aren’t in a position to provide details. Our attention was on the newish confrontational stance of Intel emphasizing the seriousness of their challenge to IBM. 

Intel has clearly decided to be more publicly aggressive, emphatically reiterating Intel’s intent to replace IBM’s OpenPower-based systems to return to super computer dominance. That Intel and IBM have been battling each other for dominance in the supercomputer space is not exceptional news. IBM has made no effort to disguise their intent to undercut Intel’s overwhelming market dominance. In fact, IBM’s OpenPower chips currently hold the top spot as the world’s fastest supercomputer. On the other hand, in overall numbers of hi-performance computers, Intel chips are found in many highly competitive supercomputer systems. 

Intel operations have been somewhat hindered recently, possibly because of organizational missteps and a shakeup in upper management. But, neither of these prevented last fall’s confirmation of a 2017 decision by the US Department of Energy (DoE) that selected Intel as prime contractor, and Cray as subcontractor for the team to deliver the exascale Aurora system to DOE’s Argonne National Lab in 2021.

While we have no insight into internal Intel decisions, this looks to us as if the new CEO at Intel decided that it is high time to rattle the cage of their competitor. Both IBM and Intel have committed to delivery of new machines in the next two years. Today, IBM holds the top two spots with Summit at Oak Ridge Nat’l Laboratory.  and Sierra at Lawrence Livermore Nat’l Laboratory. But, in the overall world of supercomputing, Intel chips appear more often. We look forward for more substantive information from both players before potential clients have to make serious decisions. Neither IBM nor Intel responded to requests for comments.