Strategic Acquisition Brief · Confidential
Apex is the enabling technology under hundreds of brands and 22 million accounts. AscendOS gives the ecosystem real-time rails and an agentic AI layer. The one thing it does not own — and is currently renting — is a planning engine that computes the correct lifetime answer and lets anyone “drill into any assumption and verify outputs.” That is MaxiFi: thirty years of Kotlikoff’s deterministic optimization engine. One integration into AscendOS distributes it fleet-wide — and insures the embedded-advice strategy you’ve already, publicly, committed to.
On March 4, 2026 Apex announced a partnership with Wavvest — “the AI engine automating professional financial planning” — to deliver AI-powered planning to Apex clients through the AI Suite. Wavvest pulls client data from AscendOS APIs and, “in minutes,” generates comprehensive plans. The roadmap is explicit: read-only today, read-write next, so AI recommendations auto-execute through Apex’s APIs across millions of accounts.
Two phrases in Apex’s own release are the whole argument. First, the buying criterion: advisors must be able to “drill into any assumption and verify outputs” before presenting to clients. That is a demand for determinism — and it is MaxiFi’s native language, not an LLM’s. Second, Bill Capuzzi frames planning as something Apex acquires “through strategic partnerships.” Translation: Apex does not own a planning-and-advice computation engine. It is renting a probabilistic one — from a firm founded in 2024.
Apex has retirement rails (IRAs, rollovers), an advice workflow (“Augmented Advice” — portfolio automation), and an AI prototyping layer (the AI Suite on Vertex). What it does not have is the planning-and-advice computation — that is outsourced, probabilistic, and not its own IP. MaxiFi supplies exactly that primitive: owned, computed, decades-validated. Not a feature on the P&L — the missing computational core of the embedded-advice strategy Apex has already committed to.
Apex defines itself as the layer that powers other companies’ financial products — the infrastructure under hundreds of clients and tens of millions of end investors. Coinbase chose Apex to power stock and ETF trading inside its “Everything Exchange” (Feb 2026). State Street is both partner and investor on joint global digital-wealth custody. That is the distribution surface a planning engine would ride.
This is the fact a revenue multiple cannot see. A planning-and-advice computation embedded in AscendOS APIs does not reach one seat — it propagates instantly across every brand Apex enables. One integration, fleet-wide distribution across 22M accounts and 200+ brands. The value is the fleet, not the seat. A SaaS comp prices one vendor’s ARR; the keystone reality is computed, verifiable planning inherited by the entire ecosystem the moment it lands in the OS.
Apex’s whole thesis is leverage: build once, distribute across the ecosystem. Apply that thesis to MaxiFi and the math inverts the tuck-in frame. A planning primitive that costs one integration and distributes to every brand is not a line item — it is infrastructure. The right comparison is not “what does a planning SaaS earn,” it is “what is correct, verifiable advice worth across 22 million accounts whose advice carries Apex’s name.”
A revenue multiple on MaxiFi’s standalone ARR is the floor — the number a deal team starts from. It is the wrong unit. MaxiFi is not a SaaS line item to bolt onto the P&L; it is the missing computational primitive that completes and insures the embedded-advice strategy Apex has already declared. The keystone value is the product of four things a comp cannot price:
One integration into AscendOS reaches 22M accounts and 200+ brands — not one seat. The asset prices against the ecosystem, not against itself.
A 30-year economist-built engine no competitor can replicate — and placed here, denied to Pershing/BNY, Fidelity, Schwab, DriveWealth racing for the same ground.
Deterministic, auditable outputs convert the biggest scaling risk — wrong advice at auto-execute scale — into a differentiator. A comp prices a feature; this prices the insurance policy.
State Street stake + a live IPO track = a window where a proprietary AI-correctness moat — “AI you can trust with money” — supports a richer multiple.
Move the unit of value from “MaxiFi’s ARR × a multiple” to “the risk-adjusted value of correct advice across 22M accounts, plus the moat denied to every rival reaching for the same embedded-advice ground.” The first is a tuck-in. The second is a keystone to the trajectory Apex is already pricing into its own story.
The threat, stated soberly. Apex is in the advice-manufacture path now — plans generated on its data, through its AI Suite and a rented engine, with its name on the rails, heading toward read-write auto-execution across millions of accounts. Apex is a regulated entity, not a neutral pipe: SEC Reg BI, FINRA Rule 2111 (suitability), 3110 (supervision), 2210 (communications) all reach it. Probabilistic advice at scale, after a regulator has put the industry on notice, is supervision liability — and the franchise/equity damage lands on the provider whose name is on the rails.
FINRA fined Apex Clearing $3.2 million on February 4, 2025 — for the fully-paid securities-lending program — for supervision (Rule 3110) and communications (2210) failures; more than 5 million retail investors received documents containing misrepresentations. Apex has already been penalized for failing to supervise a program that reached millions of downstream investors. Wrong-AI-advice at auto-execute scale is the same failure mode — larger.
FINRA’s 2026 Regulatory Oversight Report (Dec 2025) treats GenAI as “a supervised technology” and specifically scrutinizes AI agents that “can act or transact,” demanding audit trails, reproducibility, and human checkpoints before execution. That maps onto Apex’s agentic AI Suite and the Wavvest read-write roadmap exactly — and forecloses the “novel technology” defense.
The antidote — the cybersecurity-vendor posture. A security vendor lets its clients stand behind what they ship. MaxiFi does the same for advice: it offers the ecosystem computed, verifiable, reproducible planning instead of probabilistic estimates. Outputs are deterministic and auditable — you can “drill into any assumption,” which is Apex’s own stated criterion — delivering precisely the audit-trail, reproducibility, and human-checkpoint posture the 2026 report asks for. The fiduciary kicker: MaxiFi computes “the most a household can sustainably spend with what it has” — defensible by construction — rather than the aspirational “how much will you need” that manufactures a wrong, litigable number.
Net: a comp prices a feature. The keystone frame prices the insurance policy on the whole embedded-advice strategy — computed, not guessed.
MaxiFi is the financial-planning platform of Economic Security Planning, Inc., built over more than three decades by Professor Laurence Kotlikoff of Boston University. It uses consumption smoothing and dynamic programming to compute the single, mathematically optimal lifetime plan — solving simultaneously across Social Security strategy, Roth-conversion sequencing, withdrawal order, and the full current tax code. The answer is computed, not generated; it cannot hallucinate the lifetime math.
Goals-based tools answer “What is the chance you hit your number?” MaxiFi answers “What is the optimal path, and how much can I spend today without jeopardizing tomorrow?” It is not a better simulator. It is a different class of engine — the “verify outputs” engine Apex is already asking for.
Prof. Laurence Kotlikoff — William Fairfield Warren Professor at Boston University; Harvard Ph.D.; former Senior Economist on the President’s Council of Economic Advisers; Fellow of the American Academy of Arts & Sciences and the Econometric Society; named by The Economist among the 25 most influential economists.
Taught by Nobel Laureate Robert Merton at MIT Sloan as “an outstanding science-based lifecycle and retirement management platform”; featured in Bankrate’s “Best financial planning software of 2025” roundup (May 5, 2025), cited as best for near- and long-term tax planning and the decumulation phase. (Prof. Merton teaches it; he does not endorse products.)
Algorithms refined over 30+ years, built from economic theory rather than scraped text — exactly the IP a large language model cannot reverse-engineer and a clearing firm cannot buy off the shelf. Vertex AI is the general layer; MaxiFi is the content of correctness it can’t generate.
Apex added Chip Roame (Tiburon) to its board to signal advisory seriousness — i.e., it buys credibility. Kotlikoff is a category-defining name on the IP itself, available as continuing contributor and spokesperson — helping integrate, and standing behind the engine in public. A Wavvest-style vendor cannot match that.
The integration is direct and uses the architecture Apex already built. MaxiFi’s engine is exposed through AscendOS APIs as the planning-and-advice computation layer. Wavvest’s LLM and the AI Suite remain the conversational and prototyping front end; MaxiFi becomes the verified backbone they call for the numbers — so “drill into any assumption and verify outputs” stops being a promise and becomes the system’s default. Read-write auto-execution then runs on a deterministic, auditable core instead of a probabilistic one.
Beyond the API embed, Apex can train its AI Suite on MaxiFi-generated cases — perturb the Federal Reserve’s Survey of Consumer Finances into billions of synthetic households, run each through MaxiFi, and train on the verified input–output pairs. The correct economics then lives in the weights — no separate runtime layer to stand up, operate, or carry as a compute line item. The fastest path to “verify outputs” at fleet scale is the engine that computes the answers the AI Suite is meant to deliver.
Pershing/BNY, Fidelity Institutional, Schwab, and DriveWealth are racing for the same embedded-advice ground. Each can build an LLM wrapper, license a probabilistic planner, or stand up an AI assistant. None of them has — and none of them can quickly build — a 30-year economist-built lifetime-consumption engine. That asymmetry is the whole point.
There is one MaxiFi. Placed inside AscendOS, it is denied to every rival reaching for the same embedded-advice channel. The moat is not the feature — it is the exclusivity of computed correctness across the infrastructure layer of the investing ecosystem.
It is the one planning primitive a clearing firm cannot buy off the shelf, the “verify outputs” engine that turns Apex’s declared advice strategy from a rented promise into owned, defensible infrastructure — and the proof point that re-rates the IPO story.
Over the past ten weeks Larry has published a six-post sequence on his Substack, running named LLMs — Claude, ChatGPT, Gemini, Perplexity — against MaxiFi on real household problems. Findings are dated, reproducible, and dollar-specific — precisely the “verify outputs” standard Apex is advertising. These are the exact failure modes a rented probabilistic engine carries into auto-execution.
“Claude recommended discretionary spending that was far too low initially and far too high later in life. It also told John to purchase $1 million less in life insurance than he needed.”
Read the head-to-head →“Large language models are trained on text, not on solving optimal household financial problems. They have no internal model of taxes, Social Security, mortality risk, or lifetime budget constraints.”
Read the structural argument →“ChatGPT gave a confident Social Security claiming recommendation that would cost this household tens of thousands of dollars over their lifetimes.”
Read the Social Security test →“Such training would be straightforward based on billions of cases we could easily construct by perturbating observations in the Federal Reserve’s Survey of Consumer Finances — observations we could then run through MaxiFi.”
Read the training architecture →“Claude concluded that John’s real sustainable discretionary spending was $167,000 per year — 72.7 percent more than John can afford. If John spent at that level, he’d run out of money mid-retirement.”
Read the Roth test →“Claude understates John’s base plan’s final estate by 31 percent. On a re-prompt, Claude now says the final plan reduces John’s terminal estate by over $1 million.”
Read the estate test →Larry’s Substack has 137,000+ subscribers as of May 2026 and is growing. Acquiring MaxiFi acquires the engine and the megaphone — pointed, with credibility no one in the category can match, at the exact “verify outputs” standard Apex has put in writing.
MaxiFi is being offered through a focused strategic process. For Apex the integration path is direct — the computational core embedded in AscendOS, distributed fleet-wide — the strategic payoff is immediate, and the structure is straightforward. The ask is a full acquisition of MaxiFi, placing the keystone where it completes and insures the embedded-advice trajectory Apex has already declared. Larry Kotlikoff continues as contributor and spokesperson — helping integrate and standing behind the engine in public — which de-risks the integration.