Phased Research Roadmap

The timelines are approximate targets, not commitments — and the right way to read them is conditional: "if substantial sustained investment occurs and current trajectories continue without major surprises, these are reasonable horizon estimates." Reading them as delivery promises leads to over-commitment; reading them as ungrounded guesses leads to under-investment. The honest middle is "well-considered estimate, subject to revision at each gate."

What would compress them: a breakthrough in one of the named bottlenecks — an unexpected B3 result could pull Phase II forward by years. A major step-change in funding (the timelines assume gradual scaling; a concentrated national-scale commitment would accelerate). A shift of industry investment into AIHS-relevant areas, so the programme rides commercial momentum rather than carrying the full cost.

What would extend them: an unsuccessful decision gate — a Phase II finding that B3 will not yield to current approaches extends the whole programme substantially. A public-trust event — a serious adverse incident (the confidently-wrong-intervention failure mode) could set the field back years. Fragmentation — geopolitical or commercial siloing that reduces the data-sharing and collaboration the harder advances depend on.

The asymmetry matters for a funder: the downside risks (gate failures, trust events, fragmentation) are more numerous and more likely than the upside accelerators, so a prudent plan treats the stated horizons as optimistic-to-central rather than conservative, and builds in the decision gates precisely so that ambition can be recalibrated to evidence rather than held fixed against it.

The comparison is instructive mainly for how it fails, because AIHS is structurally unlike the canonical big-science programmes.

The Human Genome Project had a single definite endpoint (sequence the genome), a known technology trajectory (sequencing capability scaling predictably), and a clear completion criterion — you knew when you were done. AIHS has none of these: multiple capabilities rather than one goal, no single endpoint, and a success criterion defined by disease coverage rather than a binary finish line. You cannot "complete" AIHS the way you complete a genome sequence.

The BRAIN Initiative is closer in spirit but far smaller in scope — it maps roughly onto a single AIHS advance (C4 neural reconnection draws on exactly this kind of work) rather than the whole programme.

In sheer scale, AIHS is larger than any single named programme — comparable to the entire NIH research portfolio plus DARPA medical research plus equivalent international investment, sustained over decades. The closest honest analogy is not a programme at all but the post-war scaling of biomedical research as a whole: a sustained national-and-international priority maintained across generations, not a project with a charter and an end date.

For a funder, the implication is that AIHS should be thought of and structured at the scale of a sustained research priority with periodic strategic review, not as a project with a fixed budget and deliverable. This is precisely why the roadmap is built around decision gates rather than a critical-path schedule — the programme is too long, too uncertain, and too dependent on unpredictable breakthroughs to plan as a single integrated delivery effort.

AIHS does not have a sharp conclusion, and pretending otherwise would misrepresent it. The Phase IV maturity state continues indefinitely: once mature systems are deployed, refinement and extension continue, and the frontier work in any maturity state simply focuses on whichever bottleneck problems remain partial.

The more useful question is not "when does AIHS finish" but "when does the AIHS programme as a distinct entity dissolve into mainstream medicine." The likely trajectory: by mid-Phase IV, AIHS systems are routine clinical infrastructure rather than a research programme — the way continuous glucose monitoring or MRI are now simply part of medicine rather than active frontier programmes. The programme succeeds by ceasing to be a programme.

What comes after depends on which problems remain interesting once the core capability is mature. Plausible successors, each its own programme building on AIHS infrastructure: brain-computer interfaces and richer neural connectivity work; longevity-focused medicine extending beyond disease treatment into healthy-lifespan extension; population-scale preventive medicine through continuous monitoring; and specialised AIHS deployments for environments with no clinician access — trauma medicine, military medicine, space medicine.

For a funder, the takeaway is that AIHS investment builds durable infrastructure — datasets, standards, regulatory frameworks, manufacturing capability, trained workforce — that outlasts the programme itself and seeds whatever comes next. The infrastructure is arguably a more certain return than any specific capability milestone, because it retains value across every scenario, including the partial-success scenario where the hardest bottlenecks never fully yield.

The gates are the operational heart of the roadmap for a funder, and the key thing to understand is that they are recalibration points, not pass-fail exams. None of the three gates is a continue-or-stop verdict; each is a moment to adjust ambition to the evidence then available.

The end-of-Phase-II gate (~year 15) asks whether causal disease modelling (B3) has reached the level needed for meaningful counterfactual reasoning in care. A "no" does not end the programme — it pivots funding toward maximising value within mechanistically-understood domains rather than continuing to chase universal capability. The money keeps flowing; its target changes.

The end-of-Phase-III gate (~year 25) asks whether integrating the three buckets produced clinical value beyond what specialised systems alone would deliver. A "no" here redirects investment away from integration infrastructure and toward the specialised systems that are demonstrably working — again a redirection, not a termination.

The mid-Phase-IV gate (~year 40) asks whether the named bottlenecks (B3, C4, A2, C3) are showing breakthrough progress. A "no" means accepting that AIHS will be a mature partial system — explicitly framed as a successful outcome, since most of the clinical value lives in the partial system.

The discipline the gates impose is that the criteria should be set in advance — pre-registered, so that the assessment is honest rather than retrofitted to justify continuation. The decision-gates detail page covers what evidence each gate examines and who the decision-makers are. The underlying philosophy is that a programme spanning decades with this much irreducible uncertainty must be steerable: the gates are the steering mechanism, and a funder's most important role is ensuring they are taken seriously rather than waved through.