T4IS2026 Strategy Dialogue: The Next Compute Paradigm - Quantum and AI and the Discussion on 'Q-Day' and Post-Quantum Cryptography Migration

Points of this release:

- The timeline for quantum computing has compressed from the long-standing '10-15 years' to a credible '5 years'. Practical algorithms are now positioned for the era of Fault-Tolerant Quantum Computing (FTQC) in 5-8 years.

- Japan has ample funds but lacks researchers. Investment in quantum has grown about 10-fold year-on-year to the multi-billion dollar scale, but the layer of researchers and operators capable of absorbing those funds is thin. Japan's opportunity was identified in 'component infrastructure'.

- Investment attractiveness lies in the 'selling shovels to miners' layer. Networking is essential for connecting multiple quantum devices for FTQC, and quantum networking and sensing companies were positioned as the commercial layer that will emerge before full-stack quantum platforms.

- 'Q-Day'—the day quantum computers break classical encryption—was treated as the most disruptive near-future event. Attacks like 'collect now, decrypt later' are already underway, and the US is calling for a transition to post-quantum cryptography by 2035.

- The dilemma of existing large corporations (incumbents) was discussed. The risk of making the wrong bet for market leaders is asymmetric, and 'it is easier to be a great second customer than a great first customer'.

Session Overview

Sothius Inc. (Headquarters: Chuo-ku, Tokyo; Representative Director: Sera Yun) conducted a closed-door session, 'Strategy Dialogue,' at the invitation-only executive summit 'Tech for Impact Summit 2026' (https://tech4impactsummit.com/ja) held at the Tokyo Garden Terrace Kioicho Kioi Conference on Sunday, April 26, 2026. This release summarizes the discussion from one of the sessions: 'The Next Compute Paradigm: Where Quantum Meets AI.'

This session was conducted under Chatham House Rules. Therefore, this release records the themes, points, and proposals discussed, and does not attribute specific remarks to individuals or organizations.

Participants included practitioners versed in the near future of quantum computation and its applications, such as early-stage deep-tech VCs, strategy and operations departments of automotive OEMs, US cybersecurity operators with backgrounds in defense and intelligence, founders of Japan-based quantum software companies, and global communication infrastructure personnel pursuing PhDs in Quantum x AI. As a small technical roundtable, the timing of quantum arrival, connection with AI and post-quantum cryptography, and the positioning of Japan's component infrastructure were discussed.

Highlights of the Discussion

1. Quantum Timeline Compressed The consistent reading of the session was that quantum computing has moved from the position of '10-15 years away', repeated every year, to a credible 5-year horizon.

The era of Noisy Intermediate-Scale Quantum (NISC) was effectively treated as over, because the common understanding around 2023 was that NISC devices could not outperform classical supercomputers on commercially meaningful problems. The new horizon is the era of Fault-Tolerant Quantum Computing (FTQC), and practical algorithms were organized as being 5-8 years away. The forecast announced by major US hyperscalers in late March, moving up the timeline for practical quantum to around 2029, was received as a substantial signal rather than marketing. While the regulatory deadline premised on transition is 2035, there was a consensus that practical operational preparations are being accelerated ahead of that. Its arrival will be like the form when AI reached public awareness—tension rising slowly and then arriving suddenly.

2. Japan—Ample Funds, Lacking Researchers Frank evaluations were presented regarding Japan's positioning in quantum. As one published figure, Japan's investment in quantum has grown about 10-fold year-on-year, reaching a multi-billion dollar scale, and new national infrastructure (large R&D centers, imported dilution refrigerators, hardware from US and domestic vendors) is being developed.

On the other hand, there was also honest criticism that the layer of researchers and operators capable of absorbing those funds is thin. A significant portion of expenditure is going toward hardware procurement rather than building domestic capacity. (The text ends abruptly here.)