Waymo will run Tokyo's first driverless taxis in 2027
Waymo just committed to a commercial launch date in its second international market, Microsoft quietly let a rival lab's model into Word and Excel, and Oracle started another round of cuts to fund its data-centre build.
Waymo, GO and Nihon Kotsu have formally agreed to prepare Japan's first fully driverless commercial taxi service, targeting a public launch in Tokyo during 2027. The three companies announced Tuesday that the service will run vehicles at autonomous Level 4 — no safety driver on board — starting with an initial fleet and expanding to roughly 100 cars across key Tokyo districts, hailable from both the GO and Waymo apps. Waymo has been testing in the city since 2025 across seven wards, most of that time with a Nihon Kotsu crew member riding along, and the companies say the commercial timeline is contingent on approvals from the land ministry, the National Police Agency, the Tokyo metropolitan government and the relevant wards. The framing from GO chairman Ichiro Kawanabe — "the main act has finally arrived" — is the tell: this is an industry with a driver shortage and an aging workforce volunteering for automation rather than resisting it, and Waymo's second non-US market (after London, with Munich pending) lands in a country where Nissan and Uber are still in pre-revenue Tokyo testing. Waymo says it delivers more than half a million trips a week and is "over 15 times safer than human drivers" — the company's own metric, on the company's own definition, and the number Tokyo regulators will have to interrogate on the company's own paper. We tracked the European half of the same expansion in August — Waymo picks Munich for its first continental Europe robotaxi push.
Microsoft has added Grok models from SpaceXAI as an option inside Copilot for Word, Excel and PowerPoint. The change, announced on the Copilot blog Friday and rolling out through the tenant-level Frontier early-access programme, makes xAI's models the newest outside brain in an Office suite that already carries OpenAI and Anthropic — but the mechanics matter more than the logo on the picker. Choosing Grok sends prompts and document content to servers outside Microsoft-managed environments, Microsoft's own data-residency, audit, compliance and service-level commitments do not apply to that traffic (xAI's enterprise terms do), and SpaceXAI has been added to Microsoft's Online Services Subprocessor List, which is the legal register of outside companies allowed to touch customer data. Access is off by default, requires an admin to accept those terms, and is blocked outright in the EU, EFTA and the UK for the preview; Microsoft also has not said which Grok version is running. A model picker where the different tiers carry different lawyers is a procurement decision dressed as a productivity feature — it sits oddly next to Microsoft's new AI rulebook says its models are not conscious, published this week for consultation.
Oracle began a new round of layoffs this week, with affected employees told their last working day was the day they were told. Business Insider reviewed a notification email reading "we have made the decision to eliminate your role as part of a broader organizational change... today is your last working day," and spoke to three people impacted; posts from affected staff appeared on LinkedIn, Reddit and Blind from Monday, with an internal document reportedly indicating reductions above 10% on some teams and unverified estimates of 7,000 to 10,000 roles. The context is the balance sheet: first-quarter capital expenditure of $28.5 billion against $8.5 billion a year earlier, a maintained fiscal 2027 capex forecast of $90 billion to $95 billion, and a restructuring bill raised by about $700 million to roughly $2.8 billion on top of the 21,000 jobs (about 13% of headcount) Oracle already shed in the year to May 31. Oracle's own filings say AI adoption across its operations "have resulted, and may continue to result, in reductions to our workforce" — we read that ledger last week in Oracle just proved the AI cloud trade is real — the bill for it is coming due anyway.
What to watch: whether the Tokyo launch date survives regulatory review, since every party in that deal has an incentive to announce before the permits are in hand.
If your company's AI feature came with a separate vendor's privacy terms attached, would you read them — or just click the better-looking model? Tell us in the comments.
