Strategic Acquisition Brief · Confidential
Cortex is “the interface for all of Robinhood,” and its first declared pillar is Accuracy. MaxiFi is the deterministic lifetime-planning engine that makes Accuracy literal — the computed, verifiable, reproducible answer to every financial question a life actually asks. Not sampled. Not approximated. Built over 30 years by BU economist Laurence Kotlikoff.
Robinhood has built every financial endpoint of a modern life — trading, banking, credit, retirement, managed money, and now accounts for 60 million children. Cortex is rolling out across that base. Announced at SYNERGY26 (June 3, 2026), Cortex for Advisors is rolling out to RIAs on TradePMR (available in Fusion at no additional cost — AI portfolio analysis, tax insights, meeting preparation). The same assistant reaches the consumer app and the advisor desk. The base is aging into the questions that test that assistant hardest.
Robinhood framed Cortex on trusted inputs “to ensure accuracy.” On the one domain where a wrong number does lasting harm — how much a household can safely spend across a lifetime — a model that samples tokens estimates. MaxiFi computes. The gap between those two words is the entire thesis. Cortex can explain a portfolio. It cannot yet compute a life.
That gap is now visible on two surfaces Robinhood operates: 4.3M Gold consumers asking planning questions and RIA advisors on TradePMR who operate under the fiduciary standard. MaxiFi closes it — the validated, deterministic engine that produces the mathematically correct lifetime plan, computed and not generated.
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 tax code.
Goals-based tools answer “What is the chance you hit your number?” MaxiFi answers “What is the optimal path, and how much can I safely spend today without jeopardizing tomorrow?” It is not a better simulator. It is a different class of engine — the deterministic, computed answer Cortex needs under it.
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. He intends to keep contributing to the product, help integrate, and stay on as spokesperson.
Taught by Nobel Laureate Robert Merton at MIT Sloan as an “outstanding science-based lifecycle and retirement management platform.” The economics build on Nobel-laureate lifecycle and optimization work. Named Bankrate’s “Best Financial Planning Software of 2025” — cited as best for near- and long-term tax planning and the decumulation phase.
Patent-winning algorithms refined over 30+ years, built from economic theory rather than scraped text — exactly the kind of intellectual property a probabilistic model cannot reverse-engineer by sampling tokens. Not reproducible by prompt-engineering.
MaxiFi is in the wealth-advisor channel today via its Pro subscription — fee-only planners and RIAs use it as top-of-funnel lure and client-retention glue. The acquirer inherits a paying advisor base that extends naturally across the TradePMR desk Cortex just entered, and the 137,000+ subscriber megaphone of Kotlikoff’s Substack.
Cortex’s three pillars are Accuracy, Personalization, Action. Robinhood does not train its own models — it deploys leading foundation models (via Amazon Bedrock) behind Cortex. So the integration is not “train the weights”; it is to own the deterministic engine Cortex calls. Cortex converses on the front end, MaxiFi computes the lifetime answer on the back end, and every dollar Cortex returns is computed, traceable, and defensible — one correctness core under the three surfaces Robinhood already operates.
Prof. Kotlikoff has described the architecture publicly: the LLM is the conversational front end that guides data entry; MaxiFi is the back end that produces the precisely correct result. “AI’s best hope of providing accurate economics-based planning is by serving as a front end guiding data entry and using MaxiFi as the back end to produce precisely correct, not clearly pretend, results.”
Cortex keeps the interface, the reach, and the brand; MaxiFi supplies the one thing a sampled model cannot — the computed, correct number. (MaxiFi-generated cases can also fine-tune the foundation models Robinhood already deploys, where that helps.)
Most engines start from the aspirational question — “how much will you need?” — which manufactures a target number that is easy to state and hard to defend. MaxiFi alone starts from “what is the most I can safely spend with what I have?” — sustainable by construction. It is the question a household actually has, and the answer a platform serving a $30B retirement base can stand behind.
The regulatory direction on AI and financial advice is clear and consistent across multiple authorities: the deployment of AI does not modify existing fiduciary duty, suitability obligations, or anti-fraud rules — and is not a liability shield. The organization that places an AI model in its advice layer remains responsible for the output that customer receives. Courts have begun to treat AI synthesis as the deployer’s own statement rather than a neutral pass-through.
For a regulated brokerage, that means personalized money answers at subscriber scale sit inside FINRA suitability and, for Strategies and TradePMR customers, under the fiduciary standard. SEC 2026 examination priorities flag AI and technology risk explicitly: “if AI affects investor decision-making, it becomes an exam priority.” The FINRA 2026 Annual Regulatory Oversight Report dedicates a section to generative AI in client-facing agents — naming hallucinations, auditability, and domain-knowledge gaps as standalone exam topics. The rules are technology-neutral; they apply to embedded third-party tools as well as proprietary ones.
AI engines trained on pre-July-2025 data widely told users the estate-tax exemption would “sunset” January 1, 2026, when the 2025 law had in fact permanently raised it to $15M per person. A household that acted on the wrong number — restructuring assets, gifting early, changing beneficiary designations — on the basis of a confident-but-incorrect AI answer had a real financial loss. MaxiFi computes the correct answer from current law. The engine updates; the answer follows.
MaxiFi is not a compliance tool and we do not frame it as one. It is a deterministic planning engine whose outputs are computed, reproducible, and traceable — qualities a fiduciary and an examiner can verify. Answers that are correct by construction are not a liability shield; they are an absence of the underlying liability. That is the antidote: not a better disclaimer, but an engine whose math is right.
The clearest datapoint is not a vendor claim. In May 2026, CBS MoneyWatch ran an identical retirement prompt — a 50-year-old single woman, retiring at 65 — through three leading AI engines. The verdicts diverged. Not directionally. On the same question, with the same inputs, three engines produced three different answers. Prof. Kotlikoff’s assessment: AI tools “may do more harm than good,” particularly on Social Security optimization and longevity assumptions.
A second finding from the same period: AI engines — trained on data predating the July 2025 law — widely stated that the estate-tax exemption would “sunset” on January 1, 2026, when the One Big Beautiful Bill Act had in fact permanently raised it to $15M per person. A confident, specific, plannable-sounding answer — and the wrong one. The harm is not theoretical: a household that restructured assets on the basis of that answer had a real and irreversible cost.
“Asked whether a 50-year-old single woman could retire at 65, three AI engines diverged. Kotlikoff: AI ‘may do more harm than good,’ mishandling Social Security and wrongly averaging longevity instead of using maximum life expectancy.”
The divergent-verdict story →AI engines trained before the July 2025 law widely told users the estate-tax exemption would expire January 1, 2026, when it was permanently raised to $15M per person. A household acting on the wrong number — gifting early, restructuring assets — bore a real and irreversible cost. MaxiFi computes from current law.
Estate-tax head-to-head →Neither finding is directed at any single platform. They describe the sector-level condition: general-purpose language models making confident, specific financial statements that diverge from each other and from law. The question for any regulated brokerage fielding planning questions at subscriber scale is what its deployed AI is doing on problems like these. MaxiFi is the answer that is computed, reproducible, and current — not because it is a smarter language model, but because it does not use language model inference for the lifetime math.
Over the past several months Larry has published a six-post sequence on his Substack, Economics Matters (137,000+ subscribers), running named models — by name, on real household problems — against MaxiFi. Each post is dated; each finding is reproducible; each dollar figure is verifiable. These are the seller’s principal’s own public experiments, before any conversation with a buyer.
“The AI said John and Jane can spend approximately $52,000 per year in discretionary spending. MaxiFi’s demonstrably correct answer — verifiable by inspecting its reports — is $63,382.”
Read the head-to-head →“AI’s best hope of providing accurate economics-based planning is by serving as a front end guiding data entry and using MaxiFi as the back end to produce precisely correct, not clearly pretend, results.”
Read the structural argument →“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 →“The median household leaves $182,370 of lifetime Social Security on the table. AI tells Jane a job change adds at most $35K in lifetime benefits when the right answer is $168K.”
Read the Social Security 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 →Acquiring MaxiFi acquires the megaphone these pieces ship from — pointed, with credibility no one in the category can match, at exactly the planning surface Cortex is now reaching across consumers and, soon, advisors. Larry intends to keep contributing to the product and to stay on as spokesperson. The Substack series is the dated, dollar-specific record; the CBS finding is the neutral national press corroboration.
Robinhood frames Cortex on trusted inputs “to ensure accuracy.” On lifetime money questions, accuracy means computed, not sampled — what a probabilistic model cannot guarantee and MaxiFi guarantees by construction. The Accuracy pillar is either a technical fact or a brand position; MaxiFi makes it the former.
Cortex now answers planning questions for 4.3M Gold consumers and for advisors on TradePMR. The questions a $30B-and-growing retirement base will ask are exactly the ones that require a computed answer. The surface is open; the correct engine is not yet under it.
Personalized money answers at subscriber scale sit inside FINRA suitability. TradePMR advisors and Strategies operate under the fiduciary standard. Regulators are explicit: AI does not change fiduciary duty and is not a liability shield — the deployer owns the output. A confident-wrong number at scale is an unpriced liability that grows with the subscriber base.
MaxiFi’s answers are computed, reproducible, and auditable — turning Cortex’s planning answers from an examination exposure into an examination exhibit. And they start from “the most you can safely spend,” sustainable by construction, not the aspirational number that manufactures the wrong, litigable answer.
There is exactly one MaxiFi. It is the Gold-and-advisor feature no rival can replicate, and placed elsewhere it is gone permanently from every other platform competing for the same planning-decade wallet. Kotlikoff stays on to integrate and as spokesperson — the megaphone and the mathematician, owned.
A 30-minute briefing with a live demonstration: MaxiFi solves a real household’s lifetime plan while the leading frontier models are asked to match it. The gap between the computed answer and the sampled one is the entire thesis. Larry Kotlikoff would like to show Robinhood leadership, personally.
MaxiFi is being offered through a focused strategic process. The preference is an acquisition — that is where the strategic value sits. Founder continuity de-risks it: Larry Kotlikoff intends to keep contributing to the product, help integrate, and stay on as spokesperson. The integration path is short, and the strategic payoff — Cortex’s Accuracy pillar made literal, competitive denial, and wallet share across the planning decade — is immediate.