Solution · Portfolio Transitions

Move the money without paying for it twice.

Onboarding a legacy book, shifting a client into a new strategy, or putting new cash to work all share one hard truth: the path from the old portfolio to the target runs through realized gains, transaction costs, and tracking error — and the naive path is expensive. PRISM finds the route that respects all three, for every account, at once.

In plain English. When a client moves from one investment strategy to another — or joins with a portfolio they already own — you can't just sell everything and start again. Selling triggers a tax bill, every trade costs money, and while you're mid-move the portfolio doesn't match either the old plan or the new one. There are many possible routes from here to there, and they differ by real amounts. This page is about finding a good one. Unfamiliar words? See the glossary.

Current book Target model
Realized gainwithin budget
Trade cost12 bps
Tracking err8 bps
75,257
Real US-equity assets carried on the transition evidence lane
1
Objective — gains, cost and tracking priced together, not in sequence
3
Levers — realized gains, cost, tracking — balanced in one objective
100,004
Assets in the largest structured benchmark, solved deterministically
A large portfolio moving from a legacy book to its target along an optimized path
PRISM · Portfolio Transitions Move a book to its target without paying for it twice — realized gains, cost and tracking error balanced in one solve. A transition an exact commercial solver takes 179 s to run returns in ~1.4 s.
01 · The problem

The straight line is the expensive one.

Selling the old book and buying the target realizes every embedded gain at once and pays full freight on costs. The good path threads three trade-offs simultaneously — which is what makes it an optimization, not a checklist.

Realized gains

Defer gains where the tax saved beats the tracking it costs; harvest losses on the way. Stay inside a gain budget you set.

Transaction costs

Spread, commissions, and market impact sit inside the objective — so the chosen plan is the one that actually nets out best.

Tracking error

Every dollar left un-transitioned is tracking risk against the target. PRISM moves exactly as far as the budget allows — no further.

↓ all three, balanced in a single solve, for every account in the wave ↓
02 · The evidence

Faster, by margins that change the workflow.

Repeated-run p50 on real US-equity transition workloads, PRISM (GPU) vs an exact commercial baseline (CPU). Every quality gap held under 0.01%.

Speedup by transition workflow · withdrawn pending re-measurement
The per-workflow speedup figures previously shown here are being re-measured under a contention guard, with machine load recorded alongside each run, and will be republished per lane once clean. The quality result is unchanged: transition plans track an exact reference to within the tolerance the workflow is run at. Losing cases kept in.
03 · How PRISM fits

One objective, three levers, every account.

The transition is a single optimization balancing realized tax, cost, and tracking toward your target — run across the whole wave deterministically.

Inputs
  • Current holdings & lots
  • Target model / index
  • Gain budget & constraints
  • Cost & ADV assumptions
PRISM
Tax-aware optimization core

Finds the cost- and tax-aware path to the target for every account at once.

Outputs
  • Transition trade list
  • Realized-gain summary
  • Cost & tracking estimate
  • Audit log
Finance workflow · Illustrative model

What a slow transition costs.

Set the universe and latency budget to see the decision value protected by re-solving the transition in time, before the moment to act has passed.

Bring a real transition. We'll show you the route.

In a matched-workload pilot, run a representative transition on your data and see the realized-gain, cost, and tracking trade-offs PRISM finds versus your current process — with every number traceable.

Request a pilot →