Intel has unveiled the Tunnel Falls research chip. This processor is intended for analysis and study to look for its applicability. Despite the revolutionary nature of quantum computing, Intel is trying to maintain as many existing manufacturing techniques as possible.
Specifically, we’re talking about a 12-qubit chip baked on a 300mm piece of silicon. While that number does not sound too impressive to a layman, it is essential to note that qubits are exceptionally capable. For example, Intel states that a current proof of concept with more than 50 qubits can theoretically run rings around existing supercomputers regarding computing power.
The new chip would allow for more experiments regarding the fundamentals behind qubits and quantum dots. In addition, it should allow researchers to see what can be done by unleashing multiple qubits on a task.
Not commercial (yet)
Intel is mainly focused on democratizing quantum technology for researchers.
During the announcement, several journalists wondered how Intel views the competitive landscape with parties such as IBM. After all, there are already some players in the quantum computing field, although it is still far from a genuine market. According to quantum chief James Clarke, Intel itself is not at all sure about the commercial nature of quantum yet. He thinks significant development steps are still needed first, which should be able to keep up with the pace of Moore’s Law. However, he does not see much point in IBM’s famous roadmaps, which tend to plot out the next five years. Clarke thinks those roadmaps soon appear dated in the face of ever-shifting challenges.
This view is reinforced by the fact that Intel will develop its next research chip next year already. Thus, it hopes to surpass the timeline of commercializing classical computing. The first transistor was developed in 1947, while a commercial microprocessor did not emerge until 1971 with the Intel 4004. The first physical quantum chips were not seen until the 1990s. For the future, Intel is also trying to build on existing techniques that chips currently support. For example, Tunnel Falls is based on CMOS, so it has a transferable manufacturing process with current transistor-based chips.
Why quantum?
Although we continue to see the miniaturization of manufacturing processes, many problems cannot be solved practically with classical binary chips. Intel cites chemistry and cryptography, among others, that can only be figured out in a timely manner with current commercial processors if the code makes certain “shortcuts. These are assumptions of specific uncertainties so that we don’t have to wait years for a computational result.
Quantum is extremely difficult to grasp in its entirety, so we depend on the abstractions that experts like James Clarke can come up with. A well-known one he brings up is the analogy of the state of a coin: in the classical computing world, a coin is either showing heads or tails. In quantum, this coin can assume a “superposition,” where it retains both states. Anyone who wants to observe it seems to force quantum to be shown as either heads or tails. This combination of things makes quantum computing potentially many times more potent than conventional chips, but there are many challenges. For example, every effort must be made to combat noise that interferes with computations. Refrigerators that can reach near 0 Kelvin are required for cooling.
Investment, full stack
Despite Intel not suddenly coming out next year with the Intel Core Quantum or something similar, this sector has substantial monetary support. Among governments, China, the EU and the U.S. are the top three in terms of investments. At any rate, Intel has invested $49 billion in the D1-R&D Fab in Oregon, which has 21,000 employees. The company appears to be particularly proud of this facility. Not all that surprising, given the breakthroughs it is making possible and will presumably do so in the future.
In addition, Intel is working with the U.S. Laboratory for Physical Sciences (LPS) at the University of Maryland, among others. Several universities in the U.S. will be the first to receive the chip.
Those who think there may be a hybrid of classical computing and quantum may eventually turn to Intel. It expects to connect HPC (high-performance computing) with quantum computers to process computations and make them manageable.
Also read: IBM to build its first quantum data center in Europe