IBM Research Zurich turned 70 this year, and the lab celebrated the only way anniversary labs know how: by publishing a blog post about itself. The post, announcing a new research agreement with ETH Zurich — Switzerland's federal institute of technology and one of Europe's most decorated universities — runs through the history, the Nobel prizes, the Gordon Bell awards, and the intern who grew up to run the place. What it does not do is say what any of this is actually for.
The new collaboration aims to build "the algorithmic basis for the future of computing where quantum information theory and classical information theory come together," according to Alessandro Curioni, the head of IBM Research Zurich who arrived as a PhD intern in 2006 and never quite left. That sentence is the entirety of the stated research agenda, per the same blog post. There are no milestones, no deliverables, no hardware targets, and no timelines. IBM's investment will fund students and projects in classical and quantum algorithms. How many students, what kind of algorithms, and over what horizon are questions the post does not answer.
The 70-year backstory is the strongest material on offer. The lab started in 1956 because Ambrose Speiser, a young ETH professor, wrote to Thomas Watson Jr. asking for a job. Watson built him a lab instead. IBM Research Zurich went on to collect two Nobel Prizes and help lay the groundwork for nanoscale semiconductors and high-performance scientific computing. The 2011 nanotechnology center it built with ETH spun out of the scanning tunneling microscope, invented in that same lab by Gerd Binnig and Heinrich Rohrer, who won the 1986 Nobel Prize in physics for it. Atom-by-atom imaging: a factoid that sounds like a trivia question and is also a thing that happened in a building in Rüschlikon.
Curioni's own trajectory from intern to lab head gets substantial ink. He spent his early research years running classical simulations of molecules, then larger molecules, then molecules in solvent on interfaces, describing the work as "fighting the windmill" against computational complexity walls. The windmill eventually retreated: his teams won the 2013 and 2015 ACM Gordon Bell prizes for supercomputing simulations of collapsing bubbles and mantle dynamics. He became an IBM Fellow, then VP of Europe and Africa, then head of the lab where he started two decades earlier.
The quantum section is where the post runs out of runway. Curioni says quantum "will reduce the computational complexity of simulating nature enormously" and "take away the bottleneck." That is a claim the field has been making for decades, and the post offers no updated evidence that the bottleneck is closer to removal than it was when he was fighting the windmill in 2011. The blog does not say which problems quantum will address first, which qubit count or error rate would constitute a useful system, or how "AI times quantum" produces anything a classical algorithm cannot already do faster on existing hardware.
IBM has not been shy about quantum announcements. The roadmap exists, the hardware exists, the developer ecosystem exists. What this post reveals is that the research agenda for the AI-quantum algorithmic direction has not yet been written down in any document IBM has made public. Seventy years is an impressive run. The next chapter will need a better outline.
