The most useful thing you can do with a foundry roadmap is timestamp it. When a transistor architecture is first described in future tense in a filing, that date becomes the baseline for judging whether 'will deliver' ever turned into 'shipping.' For Intel (INTC), that baseline is the fiscal 2021 10-K.
In its Form 10-K for the period ended December 25, 2021, filed January 27, 2022, Intel describes two breakthrough technologies - RibbonFET and PowerVia - and identifies RibbonFET as Intel's implementation of a gate-all-around transistor, framed as its first new transistor architecture in years. The filing was surfaced via EdgarBeast, the SEC filing data API and evidence index, and the language is quoted from the document.
Note the tense. In this early filing the technologies are forthcoming - a stated intention to build a gate-all-around transistor and a backside power scheme. Gate-all-around wraps the gate fully around the channel for tighter control than a FinFET; backside power moves the power-delivery network beneath the device. At this point both were architecture commitments, not volume claims.
That is exactly why the origin matters. By later fiscal years Intel's annual reports shifted to saying 18A introduced these technologies, and one even called 18A the first high-volume commercial implementation. The distance between the fiscal 2021 future tense and the later past tense is the real measure of execution.
For node-watchers, the discipline is to read the earliest filing that names an architecture, fix the date, and then track the verb. Intel's fiscal 2021 10-K is the RibbonFET baseline; read its breakthrough-technology language directly and measure every later claim against it.