Prediction 1: Small, Discretionary Investors Make a Comeback
Small fundamental managers have been losing ground for a long time. They face stiff competition from large multi-strategy funds, ETFs, and quantitative/systematic strategies. Traditional value investing has also taken a hit, particularly as retail investors drive a larger share of trading volume.
In 2026, however, that dynamic may begin to change. Two tailwinds are working in favor of small funds:
1) Small and nimble organizations are reaping immediate benefits from AI, while large funds and asset managers struggle with clunky internal builds and low user adoption. AI collapses some—though not all—of the advantages of scale.
(See “Small Players Benefit Disproportionately” in our previous post for a full discussion.)
2) AI makes it easier for fundamental investors to adopt elements of systematic investing. Fundamental teams can now cover many more names without sacrificing depth, while integrating automated, systematic checks. (Find a list of practical ways investors are cutting time here.) Meanwhile, quantitative funds are struggling to extract the full value of LLMs.
Most people assume quantitative strategies are the primary beneficiaries of AI. This is not true. Point-in-time data requirements, long backtests, and frequent model retraining create strong disincentives to change. Even small updates can invalidate years of work.
Backtesting constraints also push quantitative funds toward smaller models, fewer reasoning steps or chains, and less computationally expensive approaches. The result is a slower rate of change and lower adoption of some of the most powerful AI applications.
Prediction 2: All-in-One Platforms Go Extinct
There is an increasingly common narrative that software is dead. This is a silly claim. The narrative assumes that enterprises will build software from scratch using tools like Claude Code and, therefore, never need to purchase software of any kind.
At Hudson Labs, we automate as much as we can. But partially automating code generation does not mean we can walk away from our laptops and return the next morning to fully functional features. Building software requires functional design, data strategy, security, testing, quality assurance, etc. It turns out those parts still require a lot of human effort.
We have an extraordinary amount of technical capacity in-house, yet we still pay—and will continue to pay—for many software products. It is difficult to imagine a world in which we would choose to build any of the following from scratch: QuickBooks, Notion, Gmail, CustomerIO, Lemlist, Slack, Postgres, or S3.
That said, we do believe a certain category of software will die. All-in-one platforms are headed toward extinction. Examples include SAP, Salesforce, HubSpot, and Microsoft Office. None of these tools is best-in-class across most of what they offer.
Historically, their value came from centralization: one vendor, one system of record, and fewer integrations to manage.
AI has dramatically reduced the cost of integration. Narrow tools that excel at a single task are proliferating, inexpensive, and improving quickly. As interoperability improves, the premium for “everything in one place” declines.
Users will increasingly notice that many features inside large platforms are inferior to standalone alternatives. By combining narrower tools and integrating them internally, enterprises can now achieve semi-custom solutions without the inevitable setbacks of building everything from scratch.
Large platform solutions also often lack strong developer tooling, making it difficult to adapt software to internal workflows. As demand for automation grows, selling software without developer tools will become increasingly untenable.
In an era where integration is cheap and code is abundant, mediocrity at scale is no longer defensible.
Side Prediction on Microsoft Office
Spreadsheets are powerful, but Excel will eventually die (hopefully taking Teams with it). The decline may be slow, but it is inevitable.
Spreadsheet-centric workflows are already giving way to notebooks and web-based tools with built-in coding, versioning, and automation. Over time, Excel’s dominance will weaken in favor of cloud-native and AI-native alternatives.
Are you an investor looking to integrate narrow, best-of-breed AI software in 2026? Try Hudson Labs. Start your trial here.
