Strategic Acquisition Brief · Confidential
MaxiFi is a hybrid financial intelligence engine: its deterministic core computes the optimal household plan — taxes, Social Security, longevity, lifetime consumption — and its stochastic layer evaluates upside and risk around it. The ground-truth economic model beneath a conversational interface. Built over 30 years by BU economist Laurence Kotlikoff. The computational backbone the May 5 FS agent slate needs — and the math layer behind any consumer-direct surface Claude builds in answer to OpenAI’s May 15 launch.
In the last thirty days Anthropic publicly committed to leading in financial services — and a competitor moved on the consumer flank of that surface ten days later. The May 5 FS Briefing put ten agent templates and a Moody’s partnership in market, headlined by Dario Amodei and Jamie Dimon. On May 15, OpenAI shipped ChatGPT Personal Finance — Plaid across 12,000+ institutions, Pro tier at $100/month, $30B reportedly raised by Anthropic days earlier to fund the fight. The contest is no longer about who builds a chatbot, but which company can deliver financial advice and engagement households can audit and trust.
A general-purpose language model is a probabilistic next-token predictor. It approximates. On the one question households and the advisors who serve them care about most — how much can I safely spend, and how do I make it last? — an approximation that is confidently wrong is worse than no answer at all. The May 5 agent slate is a public commitment; without a deterministic math layer underneath it, that commitment is a brand-and-fiduciary exposure every quarter it isn’t resolved.
MaxiFi resolves it — the validated, deterministic engine that produces the mathematically correct lifetime plan. It is the substance the May 5 agent slate needs to live up to the commitment, and the engine that lets Anthropic answer OpenAI’s consumer move with something OpenAI cannot replicate from its own stack.
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 post-OBBBA tax code.
Three planning modes are user-selectable: Safe Investing (pure deterministic — a unique solution against a guaranteed safe return); Upside Investing (deterministic floor from safe assets, Monte Carlo on risky upside); Full-Risk Investing (Monte Carlo across the full asset stack). Crucially, every Monte Carlo step runs the deterministic code — which is what lets MaxiFi properly incorporate taxes and path-specific cash flow in stochastic mode. No other Monte Carlo platform gets this right.
Goals-based planners 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.
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”; named Bankrate’s “Best Financial Planning Software of 2025.” The economics trace to Nobel-recognized work on lifecycle consumption and optimization.
Patented algorithms refined over 30+ years, built from economic theory rather than scraped text — exactly the kind of intellectual property a large language model cannot reverse-engineer.
MaxiFi is already in the wealth-advisor channel via its Pro subscription — fee-only planners use it as top-of-funnel lure and client-retention glue. The acquirer inherits a paying B2B/B2B2C base that extends naturally to Claude’s fast-growing consumer reach on the other side. OBBBA-current.
Most financial-AI tools force a tradeoff between precision and realism. MaxiFi does not. It is a hybrid engine: its deterministic core computes the optimized household plan against real-world rules — the tax code, Social Security, household structure, lifetime consumption smoothing — and its stochastic layer evaluates upside and risk around that optimal baseline. The sequencing is the unfair advantage: optimize first, simulate second.
LLMs can explain financial concepts, but they cannot — on their own — compute optimal household decisions across taxes, Social Security, longevity, and consumption smoothing. MaxiFi supplies the missing computation that turns AI from a financial explainer into a financial decision system — and it can be trained directly into Claude.
The fastest, cleanest integration is to train Claude on MaxiFi-generated cases — so the correct economics lives inside the model, not in a separate runtime layer to stand up, operate, and maintain. Kotlikoff has laid out the method publicly: perturb the Federal Reserve’s Survey of Consumer Finances into billions of synthetic households, run each through MaxiFi’s 30-year engine, and train on the verified input–output pairs. Claude stops improvising on money and starts returning answers that are correct by construction.
No new compute line item, no orchestration layer, no bolt-on verifier — ease and speed of integration is the point. MaxiFi supplies the ground-truth training signal; Claude keeps the interface, the reach, and now the math — the financial co-pilot a next-token model and a Monte-Carlo simulator each fail to deliver on their own.
This isn’t sky-watching — it’s the ordinary path from a known-defective product, plus notice, to aggregated liability. FINRA put it in writing in 2024 (Reg Notice 24-09) that existing rules bind AI output regardless of the technology; in 2026 it made generative AI a standalone examination focus and named the exact failure modes.
“General-purpose AI agents may lack the necessary domain knowledge to effectively and consistently carry out complex and industry-specific tasks.”
“Complicated, multi-step agent reasoning tasks can make outcomes difficult to trace or explain, complicating auditability.”
The threat. Read together, the two documents are the notice. Wrong advice at scale, plus a regulator’s written notice that the rules still apply, is the precondition for a class action — one whose damages run across the entire advised population, not a single account. The notice forecloses the only defense that would otherwise work: “the technology was novel” is unavailable once the rules are in print.
The cost. The direct money loss — restitution and penalties across the whole user base — is the visible part. The larger one is the equity re-rating: multiple compression as the market re-tags the platform “the AI you can’t trust with money,” and reputational damage that bleeds into every other product line. For the banks and insurers adopting Anthropic’s FS agents, correct-by-construction math has moved from a nice-to-have to an examination line.
The antidote. MaxiFi is the control FINRA is asking for — and it is specific to the threat. Own the deterministic engine and you can assure customers, advisors, and investors that every number is computed, verifiable, and reproducible — correct by construction, with an audit trail a fiduciary and an examiner can stand behind. Not better disclaimers — a different mechanism. And it starts from the fiduciarily correct question: not the aspirational “how much will you need,” which manufactures the wrong, litigable number, but “what is the most you can safely spend with what you have.” Sustainable by construction is also defensible by construction.
Over the past ten weeks Larry has published a six-post sequence on his Substack, Economics Matters, running named LLMs — Claude, ChatGPT, Gemini, Perplexity — against MaxiFi on real household problems. Findings are dated, reproducible, and dollar-specific. Two of the six posts test Claude by name; this is the seller’s principal, in public, before the conversation Anthropic and Kane & Company are now beginning.
“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 →“No one in any of the AI companies has asked me to train their LLMs on MaxiFi’s results. 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 understates John’s base plan’s final estate by 31 percent and his final plan’s final estate by 28 percent. On a re-prompt, Claude now says the final plan reduces John’s terminal estate by over $1 million.”
Read the estate test →“I fed Claude all of John’s data. It concluded that John’s real sustainable discretionary spending was $167,000 per year or 72.7 percent more than John can afford! If John were to spend at that level, he’d run out of money mid-retirement.”
Read the Roth test →Larry’s Substack has 137,000+ subscribers as of May 2026 and is growing. Acquiring MaxiFi acquires the megaphone these pieces ship from — pointed, with credibility no one in the category can match, at the FS narrative Anthropic is building. Larry’s own architectural recommendation: use MaxiFi and the Federal Reserve’s Survey of Consumer Finances to train Claude on the correct deterministic math. The math becomes native to the model rather than an external dependency. That is the integration path an acquirer owns on day one.
Anthropic publicly committed on May 5 to leading the professional surface — ten agent templates aimed at advisors, asset managers, banks, wealth platforms. Ten days later OpenAI claimed the consumer surface with ChatGPT Personal Finance. MaxiFi is the deterministic engine underneath both. One asset. Two surfaces. The $30B reported May 12 makes build-vs-buy a live capital-allocation question this quarter.
The May 5 agent slate — pitch builders, KYC, audit, household-advisory copilots — is sold into the wealth-advisor channel. MaxiFi is in that channel today via its Pro subscription: fee-only advisors use it as top-of-funnel lure to win prospects and as the recurring lifecycle math that keeps client relationships annual. The May 5 slate gets a real upgrade when MaxiFi sits beneath it as substrate — substantive delivery on the public commitment.
OpenAI’s May 15 launch puts the consumer-FS surface in motion under a competitor’s brand. Whatever Anthropic mounts in response will need a deterministic engine that returns correct answers under regulatory scrutiny. The math under that surface is also MaxiFi — the credibility bar that lets a Claude-fronted consumer FS product compete with what OpenAI just shipped, and built on Constitutional AI alignment principles that map cleanly onto MaxiFi’s objective function (maximize the household’s lifetime consumption; never product fees or AUM).
With $30B reportedly raised at a $900B valuation, the equity story needs defensible high-margin assets that live outside the compute line item. 30-year trade-secret IP, Bankrate’s #1 planning software for 2025, OBBBA-current, defensible under SEC Reg BI, DOL fiduciary, FINRA AI, and EU AI Act — the kind of vertical that underwrites a premium multiple in the public narrative.
There is one MaxiFi. If it lands at OpenAI, Google DeepMind, Meta, or a large fintech, Anthropic’s May 5 commitment to FS leadership weakens permanently — the most defensible single piece of AI-FS infrastructure is now owned by a competitor while Anthropic is still in agent-template territory. The denial case sits alongside the offensive one.
A 30-minute briefing with a live demonstration: MaxiFi solves a household’s lifetime plan while the leading frontier models are asked to match it. The gap is the entire thesis.
MaxiFi is being offered through a focused strategic process. For Anthropic the integration path is short, the talent is complementary to the teams already on hand, and the strategic payoff — substantive delivery on the May 5 FS commitment, an IPO-grade moat, and competitive denial against OpenAI — is immediate.