Skip to content
PortFrameStart free
Your first portfolio is free — no card

Turn into a portfolio.

Describe a theme you believe in. PortFrame researches it, picks the companies, weights them and shows how it would have done.

  • No account to start
  • Nothing to link
  • Ready in a couple of minutes

Browse all 26 portfolios in the gallery

Power and energySimulatedGrid Under Load+26.9%a yearbacktested 2005–2026Worst fall −45%13 holdingsGrid EquipmentTransmission ConstructionOn-Site GenerationPower Generation OwnersVSTVRTMTZBEGEVCMI+7Data through Sep 29, 2026Open it

Simulated. Hypothetical results, not a recommendation to invest. Disclosures

How it works

One sentence in. A researched, backtested portfolio out.

The same five steps run for every idea, and you watch each one happen on screen. Here is one, sped up: the idea behind Grid Under Load, from the gallery.

Recreated from the app and sped up. It ends on Grid Under Load, a real portfolio in the gallery.

Simulated. Hypothetical results, not a recommendation to invest. Disclosures

  1. 01It reads the idea

    Splits what you wrote into the parts that can actually be owned.

  2. 02It researches

    Recent news, filings and what each company actually earns from the theme.

  3. 03It picks

    Only companies with real exposure, not a passing mention — and every listing is checked.

  4. 04It weights

    Market-cap weighted, with a cap on each position so no single company runs the portfolio.

  5. 05It backtests

    Over the full available history, against a broad US market benchmark — the good years and the bad.

A couple of minutes from start to finish. Leave the page if you like — it keeps building, and you'll find it in your portfolios. Or open a finished one in the gallery.

What you get

A portfolio you can read, not a list of tickers.

Every portfolio opens on a page that explains itself: why each company is in it, what it's really betting on, and what it would have been like to hold.

A reason for every company

Each holding comes with one line on how it earns from the idea, grouped under the sub-themes the research found.

  • Grid Equipment · 5
  • Transmission Construction · 3
  • On-Site Generation · 3
  • Power Generation Owners · 2
VSTVistra CorpVistra owns a large portfolio of nuclear, gas and other generation assets and can monetize rising demand for firm electricity and capacity from data centers. The linkage is economically direct, though the company is exposed to wholesale power prices, fuel costs, market design and non-data-center customers.7.7%
VRTVertiv Holdings CoVertiv supplies UPS systems, power distribution, switchgear-related systems and other critical-power infrastructure used inside data centers. Its economic linkage is especially direct to AI data-center power density and uptime requirements, although it also earns substantial revenue from thermal-management products.7.7%
MTZMastec IncMasTec's power segment builds transmission, distribution, substations and other utility infrastructure that enables new generation and large-load interconnections. The linkage is direct, but the company remains diversified across communications, energy and other infrastructure markets.7.7%

The good years and the bad ones

Growth against the market, the worst fall and how much it swings — risk sits right next to return, never behind it.

Data through Sep 29, 2026
A year+26.9%backtested
Worst fall−45%
Swing a year30.2%

Alerts when its companies do something Essentials

Filings, news and big daily moves in what it holds, summarised in the app — with a short brief on your Home each morning.

  • Filings8-K, 10-Q and 10-K from its companies, summarised
  • NewsWhen its companies make the headlines
  • Daily moves of 8% or moreIn any of its 10 largest holdings

Ask about it Essentials

Chat opens with the portfolio attached, so “is it too concentrated?” knows what “it” is.

Is this just a bet on one story?

Read the backtest4s

Largely, yes. All 13 holdings earn from the same build-out of power for data centers, so they tend to move together: in the backtest it swung about twice as much as the US Cap 500.

Simulated. Hypothetical results, not a recommendation to invest. Chosen with hindsight. Disclosures

Ways to build

Start from a story, a rule or a mix.

An idea is the quickest way in. When you want more control, build from rules, factors or a whole allocation — every one backtested the same way.

  • Thematic Free

    Describe an idea. It researches the theme, picks the companies that really earn from it, and balances them.

    One input · a couple of minutes

  • Screened Core

    Pick companies by rules instead of a story: low debt, high cash, the top 40 by whatever you care about.

    Ask for it, or set it up yourself

  • Factor Core

    Traits long linked to returns — quality, value, momentum. Choose the tilt and it picks the names.

    Mix any of seven

  • Asset allocation Core

    Stocks, bonds and the rest in one portfolio: pick the mix, how it's split, and whether it gets safer over time.

    Start from one of five risk profiles

  • Replicate an ETF Core

    Name an ETF you own. It builds a portfolio that follows it for less, and shows how closely.

    One ticker · see the saving first

  • Remix One free

    Blend two ideas into one new thesis. It re-picks the companies from scratch and backtests the result like any other.

    Two ideas · one portfolio

Or start from the gallery. Every portfolio there opens without an account, and any of them can be remixed with your own idea. Open the gallery.

Or just say it. Every method is also a sentence in chat: “Screen for low-debt companies with high free cash flow” builds the same portfolio as the form, while you watch. Essentials

Up close

The gallery, drawn from its own numbers.

Every portfolio in the gallery on one chart. Pick any of them — on the chart or from the list — to see every holding at its share and its worst months beside its benchmark’s.

All 26 on one chart

What each returned a year in its backtest, against the worst fall along the way.

Theme · Power and energy

Grid Under Load

+26.9% a year backtested 2005–2026

Worst fall −45% · 13 holdings

The switchgear, transmission builders and power supply, on the grid and on site, that AI data centers need to run.

Open Grid Under Load

Themes, backtested: +12.3% to +32.2% a year, with worst falls of −29% to −68%.

Allocations, backtested: +6.3% to +9.6% a year, with worst falls of −21% to −41%.

See all 26 in the gallery

Every holding, at its share

Grid Under Load: 13 companies under 4 sub-themes, each drawn at its weight. Pick one to see why it’s in.

VRTVertiv Holdings Co7.7% of the portfolio

Grid Equipment

Vertiv supplies UPS systems, power distribution, switchgear-related systems and other critical-power infrastructure used inside data centers. Its economic linkage is especially direct to AI data-center power density and uptime requirements, although it also earns substantial revenue from thermal-management products.

All 13 carry the same weight, 7.7% each.

Open Grid Under Load

Its worst months

Grid Under Load’s 8 worst months in the backtest, 2005–2026, beside what the US Cap 500 did in the same months.

  • Jan 2008−14.9%Grid Under Load−5.2%US Cap 500
  • Sep 2008−13.5%Grid Under Load−8.9%US Cap 500
  • Oct 2008−32.0%Grid Under Load−18.6%US Cap 500
  • Aug 2010−13.2%Grid Under Load−5.2%US Cap 500
  • Jan 2015−17.1%Grid Under Load−3.2%US Cap 500
  • Oct 2018−14.3%Grid Under Load−6.6%US Cap 500
  • Mar 2020−23.5%Grid Under Load−12.6%US Cap 500
  • Jul 2026−18.0%Grid Under Load+0.6%US Cap 500

It fell further than the US Cap 500 in every one of them. The worst, Oct 2008: −32.0% against −18.6%.

Simulated. Hypothetical results, not a recommendation to invest. Themes are picked with hindsight. Data through Sep 29, 2026. Disclosures

Read it honestly

Every number says what it is.

A backtest is a simulation, and PortFrame treats it like one. The page tells you how the figures were made, right where you read them.

  • “Backtested,” at the number

    Every simulated figure carries the word right beside it, and every block of figures the note: simulated, hypothetical, not a recommendation.

  • Hindsight, labelled

    A theme is chosen knowing how its story turned out, and companies that failed along the way were never candidates. Its page says so, right under the figures.

  • Risk next to return

    The worst fall and the yearly swing sit beside every return. A return is never shown on its own.

  • Research, not advice

    PortFrame doesn't hold money, connect to your accounts or tell you what to buy.

“Read it as an illustration of what the theme would have held, not as evidence of what it will do.”
The note on every thematic portfolio

See it on any portfolio in the galleryRead the disclosures

Pricing

Free to start. Pay when you want everything underneath.

Your first portfolio and the whole gallery cost nothing. Plans add the full analysis, more ways to build, and AI credits for chat and builds.

Essentials Most popular

See the whole of every portfolio, and build as many as you like.

$10/ month

  • Unlimited thematic portfolios
  • Risk, factor attribution and monthly returns
  • Chat with your portfolios attached
  • Alerts on holdings, filings and large daily moves
  • Remix anything in the gallery
  • 900 credits of AI a month for chat and builds
Choose Essentials

Core

For rules instead of stories — and a better-built portfolio underneath them.

$29/ month

  • Everything in Essentials
  • Screened and factor portfolios
  • Asset allocations, and a diversified allocation around any idea
  • Cheaper replicas of ETFs you own
  • 2,610 credits of AI a month
Choose Core
  • Cancel any time — your portfolios stay
  • Credits never charge automatically
  • No brokerage link

Questions

What people ask before they start.

Are these real returns?

No. Every figure on this page is a backtest: a simulation of how a portfolio's rules would have done over past prices. The results are hypothetical — they don't reflect actual trading or any investor's experience, and real results may differ materially. Past performance, actual or simulated, is not indicative of future results.

Is this investment advice?

No. PortFrame is research software, for education and information. Nothing in it is a recommendation to buy, sell or hold anything; it doesn't hold money or connect to your accounts, and nothing is tailored to your circumstances.

Is a PortFrame portfolio an ETF?

No. A thematic ETF is a fund you buy as one ticker. A PortFrame portfolio is research: the companies, their weights and the reason for each, and a backtest of how they would have done together. Nothing is bought for you — you can export the holdings and decide for yourself.

How is it different from a thematic ETF?

A thematic ETF holds what its index rules or managers choose for the theme. PortFrame starts from your idea in your own words, researches which companies really earn from it, and shows every holding with its weight and its reason. A thematic ETF's returns are a real track record; a PortFrame portfolio's are a hypothetical backtest, chosen with hindsight.

How are the companies picked and weighted?

The research reads your idea and finds the companies that actually earn from it — in their filings and recent news, not a passing mention — and checks every listing. They're weighted by market value with a cap on each position, so no single company runs the portfolio; with only a few companies, they end up evenly weighted.

What is each backtest compared against?

Themes against a broad US market benchmark (US Cap 500), allocations against a 60/40 stocks-and-bonds benchmark, over the same years. Both lines are drawn on every portfolio's page, with the worst fall of each.

Can it follow an ETF you own?

Yes, on the Core plan: name it, and PortFrame builds a portfolio that follows its exposures with fewer holdings, and shows how closely it would have tracked. A replica follows the fund's exposures, not its exact holdings or a manager's picks.

Can I see some before I build my own?

Yes. The gallery is open to everyone, no account needed: finished portfolios with their holdings, the reason for each company, and their backtests. Open any of them, or remix one with your own idea. Open the gallery

Do I need an account to try it?

No. Your first portfolio starts building without one. While it builds, you're asked once to save it to a free account — no card, no bank connection, nothing to link. You can say not now and keep it in your browser.

How long does a build take?

Researching an idea takes a couple of minutes; a factor portfolio about one. You can leave the page — it keeps building, and it will be waiting in your portfolios.

What does “backtested” mean?

It's a simulation: the portfolio's rules run over past prices and rebalance through history, so you can see how it would have behaved — the good years and the bad. It doesn't reflect actual trading, and it doesn't say what will happen next.

Why does a theme's backtest look so good?

Because it's chosen with hindsight. These are today's companies, picked knowing how the idea turned out, and companies that failed along the way were never candidates. Each thematic page says so, right under its figures.

Can I buy the portfolio through PortFrame?

No. PortFrame doesn't trade, take custody or connect to a brokerage. You can download a portfolio's holdings as a spreadsheet.

What are credits?

Paid plans come with a monthly allowance of AI credits for chat and builds. Nothing is charged automatically: when you run low, you choose whether to buy more, and bought credits never expire.

Can I cancel?

Any time. Your portfolios stay either way.

Disclosures

Simulated results. Every performance figure on this page is backtested or simulated, not an actual result. Backtested and simulated results are hypothetical, do not reflect actual trading or any investor's experience, and may differ materially from real-world outcomes. Past performance, actual or simulated, is not indicative of future results.

Chosen with hindsight. A thematic portfolio holds today's companies, picked with what is known now and applied to the past, and companies that failed along the way were never candidates. Its backtest illustrates the idea; it is not a strategy anyone could have followed at the time, and not a forecast.

Not a recommendation. PortFrame is provided by MGI Machine Guided Investments LLC for educational and informational purposes only. MGI is not registered as an investment adviser, is not a broker-dealer and does not provide personalized investment advice. Nothing here is an offer or a solicitation, and the inclusion of any security is not a recommendation to buy, sell or hold it.

Risk. All investing involves risk, including the possible loss of principal. Outputs are generated with AI and quantitative models and may contain errors. You are solely responsible for your investment decisions and should consult a licensed financial professional before acting.

The figures are the gallery's own, as the app shows them, with data through Sep 29, 2026. Read the full Key Disclosures

What do you believe in?

Write it the way you'd say it. The research, the picks and the backtest take a couple of minutes.

Research and simulation only. Backtested results are hypothetical and are not a recommendation to buy or sell anything.

Or start from one of these

Not sure yet? Look through the gallery first