Try it on real examples: pick a sector, a question and what you need to walk away with. Every path ends in a real Janus answer, edited for length.
Example questions, real Janus answers, as of November 2026
Power & utilities
Capital investment analysis
Should we build this gas plant?
For: A European utility board · running /analyze-capital-investment
We're deciding in 2027 whether to commit about €700 million to a new 800 MW gas plant, online in 2030 and planned to run 25 years for evening peaks and low-wind weeks. Where does the stranding risk sit, and what would be a stronger use of the same capital?
Positional Intelligence ReportExcerptJanus · November 2026
The €700 Million Gas Plant Decision: Stranding Before Payback, and a Stronger Use of the Capital
Summary
The board should not commit €700 million to a merchant 800 MW combined-cycle gas plant with a 25-year life.
Batteries take the evening peaks before the plant opens in 2030. On my reading of the cost curves, the mid-merit hours go between 2032 and 2036, about a decade before an assumed payback in the early to mid 2040s.
The plant strands before payback if it runs merchant.
Batteries already do the evening-peak job more cheaply: US$144.9/kWh installed in 2025, roughly US$54.5/kWh by 2030 (NavigateX forecast).
New gas in Europe now earns its return from public capacity contracts, not from selling electricity.
The same capital works harder as staged batteries, solar and wind on the utility's own grid connections.
BESS (Utility-Scale) NavigateX projection, 80% band
NavigateX Technologies Library. NavigateX projection, 80% band, real 2026 US$.
When each of the plant's jobs is lost
Plant job
What replaces it
Crossing estimate
Evening peaks
Solar shifted by 4-hour batteries
Europe 2027 to 2030
Overnight and winter-evening hours
8-hour-plus batteries with overbuilt solar and wind
2032 to 2036
Multi-day low-wind weeks
Overbuilt wind and solar, storage, flexible demand
Late 2030s to 2040s
Excerpt: about 190 of 4,400 words
Sources in Janus: 16 · Written for an outside audience · Saved as a document in the Space's library
Board paper · 2027 investment decisionExcerptJanus · November 2026
Proposed 800 MW Combined-Cycle Gas Plant: Decline the Merchant Case, Redeploy the Capital
Purpose. The board is asked to decide whether to commit about €700 million to a new 800 MW combined-cycle gas plant, online in 2030 and running for 25 years. This paper recommends declining the plant as proposed and approving a staged alternative.
Summary for directors
If the plant sells its power on the open market, it loses its paying hours about a decade before payback.
Batteries already do the evening-peak job more cheaply, and their cost is still falling fast.
New gas in Europe now earns its return from public capacity contracts, not from selling electricity.
Recommendation
Decline the merchant CCGT in 2027 and approve a staged battery, solar and wind program on the utility's existing connections.
2027: decline the 800 MW merchant CCGT as proposed.
2027 to 2028: approve a first tranche of about 3 GWh of 4-hour batteries at existing thermal sites.
Before 2030: commission whole-system modeling of the zone's residual low-wind gap.
Risks and mitigations (excerpt)
Risk
Mitigation
Battery costs fall more slowly than forecast
Stage procurement. Even the upper band beats the 2025 cost the first tranche is priced at.
A severe low-wind event before 2030 triggers backlash against declining gas
Contract firm cover from existing plants and demand response now.
Grid connection delays hold up the battery tranches
Build first at the utility's own thermal connection points.
Excerpt: about 200 of 2,600 words
Sources in Janus: 2 · Written for an outside audience · Saved as a document in the Space's library
Leadership cheat sheetExcerptJanus · November 2026
The 800 MW gas plant decision on one page
The one thing to remember
If it sells on the open market, the plant loses its paying hours in 2032 to 2036, about a decade before its early to mid 2040s payback.
What strands, and when
Evening peaks: lost to solar shifted by 4-hour batteries, 2027 to 2030, before the plant opens
Overnight and winter-evening hours: lost to 8-hour-plus batteries and overbuilt solar and wind, 2032 to 2036
The utility's existing gas fleet, by the same mechanism, still unmeasured
Where new value forms
Grid connections at thermal sites, the scarce input
Capacity revenue for storage and flexible demand as auction rules credit them
Surplus solar and wind power sold under long-term contracts to flexible loads
Where Janus differs from consensus
Consensus
Janus's view
Standard valuation holds a plant's capacity factor high and constant (US EIA assumes 85%)
Mid-merit hours are gone between 2032 and 2036.
Capacity contracts make new gas investable (German tender, Nov 2026: all gas, 15-year terms)
Returns hold, but as policy risk. A contract changes who pays, not what the asset is worth.
Excerpt: about 170 of 1,600 words
Sources in Janus: 3 · Written for an outside audience
Technology cost comparison
Which technology wins on cost?
For: A utility planning team
Where does new nuclear stand against solar, wind and batteries at the costs of the year it comes online?
Brief answerExcerptJanus · November 2026
Where does new nuclear stand against solar, wind and batteries at the costs of the year it comes online?
New nuclear loses on the cost of the year it comes online. It doesn't come close.
A plant approved today connects around 2035. By then its capital cost will be about what it is now. Over the same years, solar, wind and storage are projected to fall to a fraction of today's prices. Nuclear is priced at the decision. Its competitors get priced at commissioning.
The two cost curves run in opposite directions.
Firming does not rescue the comparison. Cost per kilowatt of capacity is not cost per unit of electricity delivered. Nuclear runs most of the time, while solar and wind need overbuilding plus storage.
The overbuild is an asset, and that is the second loss for nuclear.
Capital cost: nuclear against solar and windUS$/kW
Nuclear Solar PV (Utility-Scale) Onshore Wind
NavigateX Technologies Library.
Excerpt: about 120 of 700 words
Sources in Janus: 6 · Written for an outside audience
Follow-up questionExcerptJanus · November 2026
Earlier in this thread
Where does new nuclear stand against solar, wind and batteries at the costs of the year it comes online?
New nuclear loses on the cost of the year it comes online. It doesn't come close.
Do small modular reactors change this comparison, or follow the same cost curve?
Small modular reactors (SMRs) do not change the comparison. The first real projects cost more per kilowatt than large reactors, not less.
Their case rests on a bet that factory production will eventually bring costs down the way it did for solar panels and batteries. That bet has not paid off anywhere yet. It would also need to pay off faster than solar, wind and batteries keep getting cheaper.
The first reference project prices above large nuclear.
The leading US design failed its own cost test before it was built.
The modular argument is a promise about the experience curve, and the volumes are missing.
Excerpt: about 110 of 550 words
Sources in Janus: 6 · Written for an outside audience
Leadership cheat sheetExcerptJanus · November 2026
New nuclear against solar, wind and batteries
The one thing to remember
Nuclear is priced when the decision is made, while its competitors keep getting cheaper until the day it switches on.
What strands
Commitments to new reactors, large or small, made today against 2035 to 2045 alternatives
Utility capital plans built around state-guaranteed revenue for nuclear.
Equity in SMR developers that is priced on fleet learning and has no measured unit cost behind it.
Where new value forms
The surplus a cost-optimized solar, wind and battery grid produces most of the year, almost free at the margin, what we call Superpower
Loads that move to that surplus.
Grid connections and sites at reactors that are retiring or never get built.
Where Janus differs from consensus
Consensus
Janus's view
Nuclear capacity rises more than 50% to ~650 GW by 2050 under today's policies (IEA, Jan 2025)
New build follows state support, not economics. The flat 15-year installed base is a better baseline.
More than 1,000 SMRs and 120 GW by 2050 with stronger policy (IEA, Jan 2025)
First-of-a-kind SMRs cost more per kW than large reactors. The volume needed for learning is missing.
Excerpt: about 180 of 1,600 words
Sources in Janus: 10 · Written for an outside audience
Public infrastructure check
Is this energy-security project a good bet?
For: A government energy ministry
Where does New Zealand’s proposed LNG terminal stand against solar, wind and storage over its payback?
Brief answerExcerptJanus · November 2026
Where does New Zealand’s proposed LNG terminal stand against solar, wind and storage over its payback?
The terminal is a fixed bet on a problem that solar, wind and batteries are already shrinking.
It would be locked in from 2028 for many years, while its competitors get cheaper every year. Its value rests on dry-year insurance, and the size of that insurance premium falls as those technologies scale.
What the terminal is meant to buy. The economic case is option value, not gas volume.
Both savings figures assume the dry-year premium holds steady for the whole contract. I think that is the flaw.
The cost curves widen the gap every year the terminal is in service.
Sources in Janus: 10 · Written for an outside audience
Follow-up questionExcerptJanus · November 2026
Earlier in this thread
Where does New Zealand’s proposed LNG terminal stand against solar, wind and storage over its payback?
The terminal is a fixed bet on a problem that solar, wind and batteries are already shrinking.
Who carries the stranding risk if the LNG terminal contract is take-or-pay and the dry-year premium collapses by the 2030s?
If the contract is take-or-pay, the payments keep coming even when the terminal sits idle, and someone in New Zealand pays them in the end.
Who pays first depends on which funding route survives the election. The current options put the risk on taxpayers, on gentailer shareholders, or on electricity customers. Customers are the most likely to end up paying in all three cases.
Take-or-pay turns a stranded asset into a stranded contract.
Passing the cost on to customers is where the damage builds.
The other side of the ledger. Every dollar locked into fixed LNG payments is a dollar the alternatives no longer have to beat.
Excerpt: about 110 of 600 words
Sources in Janus: 8 · Written for an outside audience
Leadership cheat sheetExcerptJanus · November 2026
New Zealand's LNG terminal
The one thing to remember
New Zealand's LNG terminal is long-term insurance against dry years, bought just as solar, wind and batteries make dry years smaller. The thing to remember: under a take-or-pay contract, the contract strands, not the ship.
What strands
Take-or-pay contract: fixed charges stay the same while terminal use falls.
Crown underwriting route: taxpayers take the first loss.
Gentailer "fair funding" route: shareholders take the first loss, and the Crown is one of those shareholders.
Where new value forms
Flexible hydro: its value rises as wind and solar let it act as the seasonal store.
Grid batteries: costs fall about 18% a year, and they win daily and weekly firming.
Overbuilt solar and wind: they produce cheap surplus power most of the year, mainly in summer, what we call Superpower.
Where Janus differs from consensus
Consensus
Janus's view
Benefit of NZ$265m to NZ$400m a year, held flat (MBIE Feb 2026; Treasury via b2bnews Oct 2026)
Largest in the early years, shrinking through the 2030s
The terminal caps downside risk
It swaps a domestic risk for a global price risk
Excerpt: about 170 of 1,300 words
Sources in Janus: 10 · Written for an outside audience
Oil, gas & mining
Company exposure analysis
How exposed is our company?
For: A diversified oil and gas company's strategy team · running /analyze-company-exposure
We earn about half our profit from upstream oil, a quarter from LNG, and the rest from European refining and fuel retail. Map our stranding risk and upside from EVs, solar, wind, batteries and electric trucks, ranked by timing to 2035.
Positional Intelligence ReportExcerptJanus · November 2026
Where electric transport, solar and batteries strand a diversified oil and gas company to 2035, and where they pay it
Summary
The biggest exposure isn't today's barrel. It's the long-lived capital the 2026 price spike will tempt the board to commit.
Projects sanctioned on 2026 cash flows reach first production around 2030, inside every window where these curves collide with the business.
LNG ranks first: demand stops growing in the early 2030s, against a consensus peak in 2036.
Upstream oil ranks second, on size. Road-fuel demand peaks in the late 2020s and declines after.
European refining is first to feel the squeeze, once crisis margins fade.
The upside sits in assets the company already owns: corridor sites, grid connections, fleet customers and trading skill.
Capital cost: solar and wind against gas plantsUS$/kW
Solar PV (Utility-Scale) Onshore Wind Gas (CCGT)
NavigateX Technologies Library. Capital cost per kW, so read for direction, not as a parity date.
The ranking
Rank
Vector
Timing window
1
LNG demand ceiling plus supply wave
Early 2030s
2
Upstream oil: road-fuel ceiling
Decline from late 2020s
3
European refining utilization
Gasoline now, diesel from ~2030
A1
Upside: corridor and truck charging
2026 to 2033
Excerpt: about 170 of 4,000 words
Sources in Janus: 37 · Written for an outside audience · Saved as a document in the Space's library
Board paper · Strategy and capital allocationExcerptJanus · November 2026
Committing capital against the electric transport, solar and battery curves to 2035
Purpose. This paper asks the board to adopt a rule for deploying the 2026 windfall, and to approve a sequence of moves across LNG, upstream oil, refining, retail and renewables.
Summary for directors
The board's decision is how to deploy the 2026 windfall, not how to defend today's barrel.
LNG (about 25% of earnings) is the most mispriced exposure.
The upside sits in assets the company already owns, and the familiar pivots are not hedges.
Recommendation
Adopt the capital rule now, and sequence the remaining moves through 2030.
Approve no new long-cycle oil or LNG project on crisis-era prices. Before the next investment decision.
Map every LNG contract renewal against an early-2030s ceiling. Defer new liquefaction trains. Through 2027.
Split the retail network: convert corridor sites to car and truck charging, divest urban fuel-only sites. 2027 to 2030.
Risks and mitigations (excerpt)
Risk
Mitigation
The disruption arrives later than expected
Recommendations favor short-cycle options and deferral, which keep value if the curves slow.
The renewables arm loses value on its own curve
Move the arm toward storage, trading and flexibility.
Others capture charging and fleet value first
Move on corridor sites and the fleet offer in 2027, while customer relationships are intact.
Excerpt: about 200 of 3,400 words
Sources in Janus: 4 · Written for an outside audience · Saved as a document in the Space's library
Leadership cheat sheetExcerptJanus · November 2026
Electric transport, solar and batteries against a diversified oil and gas company, to 2035
The one thing to remember
The exposure is the long-lived capital the 2026 windfall will tempt the company to commit, more than today's barrel. A project sanctioned on crisis prices starts producing around 2030, inside every window where these cost curves collide with the business.
What strands, ranked
LNG: a demand ceiling in the early 2030s meets a supply wave
Upstream oil: the road-fuel ceiling, felt first through the marginal barrel
European refining utilization: gasoline now, diesel from about 2030
Where new value forms
Corridor retail sites with battery-buffered car and truck charging
Refinery and terminal sites as hosts for loads fed by surplus power
Trading skill, and fleet and fuel-card customers
Where Janus differs from consensus
Consensus
Janus's view
Peak LNG demand of 643 Mt in 2036 (Global LNG Hub, Aug 2026)
Growth stops in the early 2030s. Consensus holds battery costs still.
Oil demand plateaus around 105.5 million barrels a day by 2030 (IEA, June 2025)
Agrees on the timing of the peak. Differs on what follows: a decline, driven by electric trucks.
Excerpt: about 180 of 1,900 words
Sources in Janus: 5 · Written for an outside audience
Where does copper stand as EVs and grids scale through 2035? What strands, where does new value form?
Positional Intelligence ReportExcerptJanus · November 2026
Copper to 2035: A Build-Out Boom With a Deficit That Closes Itself
Summary
Through 2035, the disruptions in energy and transport raise copper demand. But the shortage the market is paying for will not last.
Electric vehicles, solar, wind, batteries, grids and data centers all use more copper per unit of capacity than the systems they replace, so demand rises through the period. The record price sets off the responses that close the gap: switching to aluminum, using less copper per product, and recycling.
On the measure that matters, cost per unit of electrical conductance delivered, copper is losing to aluminum now.
I agree with the IEA that copper demand rises through 2035. I disagree that a 30% supply gap follows.
High-cost greenfield mines approved now carry the stranding risk.
Battery cost: grid storage and EV packsUS$/kWh, log scale
Sources in Janus: 21 · Written for an outside audience · Saved as a document in the Space's library
Follow-up questionExcerptJanus · November 2026
Earlier in this thread
Where does copper stand as EVs and grids scale through 2035? What strands, where does new value form?
Positional Intelligence Report
How much could cheap solar power lower aluminum's cost per unit of conductance by 2035?
Cheap solar power could cut aluminum's metal cost per unit of conductance by roughly a tenth to a fifth by 2035. That widens a gap that is already about tenfold.
Power is the one large aluminum cost that solar touches. The rest of the cost, and the copper price, matter more to the ratio. Cheap power's bigger effect is to fix aluminum's advantage in place and to decide where new smelting goes.
Electricity is about a third of the cost of making aluminum, so cheaper power has a ceiling on what it can save.
On my arithmetic, cutting a smelter's power price from US$40 to 60/MWh down to US$20/MWh saves about US$280 to 560 a tonne.
The cost curve supports a falling power price, but there is a catch.
Excerpt: about 130 of 650 words
Sources in Janus: 6 · Written for an outside audience
Leadership cheat sheetExcerptJanus · November 2026
Copper to 2035
The one thing to remember
Electric vehicles and grids raise copper demand through 2035, but the record price is already setting off the substitution and recycling that close the shortage the market is paying for. Status as of November 2026. Copper prices are not forecast here.
What strands
Copper-cobalt mines, 2026 to 2030
Producer-state budgets that depend on cobalt revenue
Cable and component makers locked into copper
Where new value forms
Recycled copper: a 34% share and rising as cars, coal plants and oil assets retire.
The aluminum conductor chain: smelters, cable makers, and the firms shaping the new standards.
Smelters on firm solar, wind and battery power
Where Janus differs from consensus
Consensus
Janus's view
Rises from 27 Mt (2024) to 33 Mt (2035) (IEA, May 2025)
Agree
Project pipeline 30% short (IEA, May 2025). Refined shortfall 3.6 Mt (McKinsey, Jun 2025)
Differ: 2035 mined copper requirement below the IEA's figure
Excerpt: about 140 of 1,500 words
Sources in Janus: 3 · Written for an outside audience
Quick question
Is this critical-mineral boom real?
For: A critical-minerals investor
Is germanium a critical mineral for solar, or a short window? What replaces it?
Brief answerExcerptJanus · November 2026
Is germanium a critical mineral for solar, or a short window? What replaces it?
Germanium is not a critical mineral for the solar disruption.
The solar panels driving that disruption are crystalline silicon, and they use essentially no germanium. Germanium matters for one niche: high-efficiency multijunction cells (cells built from several stacked layers) for satellites, which are grown on germanium wafers.
The price spike comes from export policy and other buyers, not from solar.
In space solar, the right comparison is cost per watt delivered in orbit, and on that measure germanium is losing.
The stranding risk falls on new germanium supply that is priced off today's peak.
Key numbers
Gallium and germanium prices vs 20239 to 10x
Germanium spot, China28,500 yuan/kg (about US$3,768), a record
Germanium, US warehouseUS$6,405/kg
China's share of primary gallium99% (2025)
Excerpt: about 90 of 550 words
Sources in Janus: 9 · Written for an outside audience
Follow-up questionExcerptJanus · November 2026
Earlier in this thread
Is germanium a critical mineral for solar, or a short window? What replaces it?
Germanium is not a critical mineral for the solar disruption.
Does gallium face the same short window as germanium, or is its demand more durable?
Gallium's price spike is probably a short window, like germanium's. Its demand is not.
Germanium's role in solar sits in a niche that is being engineered away. Gallium's demand comes from technologies that are themselves being disrupted into: power electronics, radio-frequency chips and LEDs. As those get cheaper, they use more gallium, not less.
The supply squeeze has the same cause as germanium's.
Gallium's demand depends on technologies that are still getting cheaper.
The real substitution threat is silicon carbide, and it splits the market rather than taking it.
Excerpt: about 90 of 600 words
Sources in Janus: 7 · Written for an outside audience
Leadership cheat sheetExcerptJanus · November 2026
Germanium and gallium
The one thing to remember
Germanium's solar role is a niche that is being engineered away. Gallium's demand is durable. For both metals, today's price spike is a policy window that will close, and supply priced off the 2026 peak is what strands.
What strands
New germanium supply priced off today's peak
Standalone gallium projects modeled on 2026 prices
The space-cell business model built on thick germanium wafers
Where new value forms
Reusable wafers and thin-layer germanium processes
Gallium recovery at alumina and zinc refineries that already run
The first high-purity gallium approved by chipmakers
Where Janus differs from consensus
Consensus
Janus's view
Scarcity is coming, prices will rise, and future solar demand "can only partially be met" (Sverdrup & Haraldsson, Dec 2024)
Solar demand for germanium shrinks per watt
Satellite launches drive germanium wafer growth; 6,000 to 15,000 wafers per satellite (Fastmarkets, Dec 2024)
More satellites, but less germanium in each
Excerpt: about 140 of 1,000 words
Sources in Janus: 9 · Written for an outside audience
Food & agriculture
Capital investment analysis
Should we build this processing plant?
For: A dairy cooperative's board · running /analyze-capital-investment
We're deciding whether to build a US$400 million milk-powder and whey-protein plant, opening in 2029 with a 25-year life, for pasture-based export milk. Stress-test it against precision fermentation and other protein cost curves. Where does the stranding risk sit, and what's a stronger use of the capital?
Positional Intelligence ReportExcerptJanus · November 2026
A US$400 Million Milk-Powder and Whey Plant Against the Protein Cost Curves, 2029–2054
Summary
The cooperative should not build this plant as specified.
It would earn its best margins on whey protein, and precision fermentation is commercializing that exact molecule now. On my reading, the whey line's price premium erodes from about 2029 to 2033, which is the plant's first five years.
Fermented whey cost has already reached today's record whey prices: about US$25–30/kg in 2025.
The largest export market for milk powder is already shrinking: China's imports halved from 2021 to 2025.
Payback lands after the crossing, and the cooperative structure turns a margin squeeze into a member-supply squeeze.
The cooperative already owns what fermentation companies lack: permeate feedstock, drying capacity, export channels and land.
Fermented whey cost against whey pricesUS$/kg
Fermented whey, 2020~100
Fermented whey, 202525–30
Record whey price, Sep 202625–28
Normal isolate price band8–15
Figures cited in the Janus report (RethinkX 2026 review; trade press, 2026). Ranges shown as midpoints.
The plant's life against the protein curves
Year
Plant
Protein system
2029
Plant opens
Fermented whey entering the isolate price band
2029–2033
Years 1–5
Whey premium erodes
2033–2040
Years 5–12
Powder margins and member milk price under pressure
2041–2044
Assumed payback
Commodity dairy protein in decline
Excerpt: about 170 of 3,500 words
Sources in Janus: 24 · Written for an outside audience · Saved as a document in the Space's library
Board paper · Cooperative investment decisionExcerptJanus · November 2026
The proposed US$400 million milk-powder and whey plant
Purpose. This paper asks the board to decide whether to approve the 25-year greenfield plant as specified, or to use the same capital for a staged position. It recommends the staged position.
Summary for directors
Fermented whey already costs about what whey sells for, and it reaches normal isolate prices in the plant's first five years.
The plant's largest export market for milk powder halved before fermentation arrived.
Payback, assumed at 2041 to 2044, lands after the economics turn.
Recommendation
Decline the greenfield plant as specified, and approve a staged program with capital released at evidence gates.
Now: decline the 25-year greenfield configuration.
2026–2027: approve a smaller brownfield powder expansion with a payback of about 10 years.
By 2027–2028: open partnership talks with one or two fermentation strain owners, offering permeate, processing capacity and Asian customers.
Risks and mitigations (excerpt)
Risk
Mitigation
Fermented whey stalls above US$20/kg, and the cooperative forgoes record whey margins
The brownfield tranche keeps powder capacity, and gates release more capital if the stall is confirmed.
Fermentation arrives faster than expected
Size the brownfield tranche for a payback of about 10 years and review gates every six months.
A fermentation partner fails
Talk to two strain owners. Commit only when contract prices fall below US$15/kg.
Excerpt: about 200 of 2,200 words
Sources in Janus: 1 · Written for an outside audience · Saved as a document in the Space's library
Leadership cheat sheetExcerptJanus · November 2026
The US$400 million milk-powder and whey plant
The one thing to remember
The plant would earn its best margin on whey, and whey is the first product precision fermentation competes for, in the plant's first five years (2029 to 2033).
What strands
The whey isolate line on dairy whey, 2029 to 2033
Export whole-milk-powder capacity, from the mid-2030s
The cooperative's share-capital base, as members exit
Where new value forms
Lactose and permeate as fermentation feedstock
Membranes, ion exchange and dryers built to take fermentation broth
Member land for solar, wind and batteries
Where Janus differs from consensus
Consensus
Janus's view
Global dairy consumption grows with income and population (OECD-FAO Outlook 2026–2035)
Protein demand grows. Demand for exported powder and whey falls from the 2030s.
Producers expect fermented whey at US$8–12/kg by 2027–2028
US$10–15/kg in 2028–2033. Slower than the producers, because purification and capacity build set the pace.
Excerpt: about 130 of 1,500 words
Sources in Janus: 1 · Written for an outside audience
Industry exposure
Will a new technology undercut our product?
For: A dairy exporter's leadership
Where do the largest dairy producers stand against precision fermentation at 2030 costs? What strands first, where does new value form?
Brief answerExcerptJanus · November 2026
Where do the largest dairy producers stand against precision fermentation at 2030 costs? What strands first, where does new value form?
By 2030 the largest dairy groups split in two.
Companies whose earnings depend on commodity milk proteins (whey, casein, milk powders) sit on the cost curve that precision fermentation is crossing first. Those that own brands and recipes can buy the cheaper molecule and keep the margin.
Our cost read puts commodity whey parity within 2027 to 2030, starting with the premium grades.
Strands first: ingredient exporters and the assets behind them.
Sources in Janus: 6 · Written for an outside audience
Follow-up questionExcerptJanus · November 2026
Earlier in this thread
Where do the largest dairy producers stand against precision fermentation at 2030 costs? What strands first, where does new value form?
By 2030 the largest dairy groups split in two.
Which dairy export economies carry the most farmland stranding risk if whey isolate reaches parity by 2028?
New Zealand and Ireland carry the most farmland stranding risk, with Uruguay close behind.
In all three, most milk is exported, dairy covers a large share of farmland, and farmers get the world commodity price with little home market to cushion them.
Whey parity by 2028 starts the squeeze, but it won't strand farmland by itself.
Exposure depends on three things: how much milk is exported, how much land is in dairy, and whether farmers are price-takers.
Pasture-based exporters are the most exposed because their land has few uses that pay as well as dairy.
Excerpt: about 100 of 600 words
Sources in Janus: 8 · Written for an outside audience
Leadership cheat sheetExcerptJanus · November 2026
Dairy against precision fermentation
The one thing to remember
Commodity milk protein is the first part of dairy to be undercut by cheaper proteins made by microbes, with cost parity around 2028 to 2030. Pasture-based export economies carry the land risk. Brand owners and fermentation capacity capture the new value.
What strands
High-value proteins such as lactoferrin (fermentation already wins).
Whey isolate for sports and medical nutrition (the producers' 2027-28 target already falls inside today's isolate price range).
Whey concentrate, casein and milk powders for processed food.
Where new value forms
Recipes and brands: owners swap in the cheaper molecule and keep the saving.
Fermentation capacity near cities: the supply bottleneck moves from herds to bioreactors.
Strain intellectual property: the microbe that makes the protein becomes the scarce asset.
Where Janus differs from consensus
Consensus
Janus's view
World milk output +1.8% a year to 1.146 billion tonnes by 2034, driven by yields (OECD-FAO Outlook, July 2025)
Agree that liters can keep growing in fresh milk in South Asia. Differ on internationally traded protein, whose value erodes from 2028-2030
It offers "small volumes" today (Rabobank, September 2024)
Agree on today's volumes. Differ on what they imply
Excerpt: about 180 of 1,300 words
Sources in Janus: 4 · Written for an outside audience
Commodity exposure
What happens to the crop we grow?
For: A plantation group
Where does palm oil stand when road diesel demand falls away? Who is exposed, where does the land and capacity find new value?
Brief answerExcerptJanus · November 2026
Where does palm oil stand when road diesel demand falls away? Who is exposed, where does the land and capacity find new value?
Palm oil's road-fuel demand is created by policy, and it shrinks with diesel volume regardless of how high the mandates go.
About a quarter of the world's palm oil is now burned as biodiesel. That share depends on blend rules applied to a diesel pool that electric vehicles and trucks will drain.
The near term runs the other way, and that is part of the exposure.
A second disruption reaches the food market, the other three quarters of demand, on a later clock.
The new value is in equatorial land, existing industrial sites, and cheap electricity.
Key numbers
Share of global palm oil burned as biofuel27%
Indonesia's share of world palm supplyabout 58%
Indonesia biodiesel blendB50 (50%)
Palm used in biodiesel under B5016.3 to 17 million tonnes
Excerpt: about 100 of 650 words
Sources in Janus: 8 · Written for an outside audience
Follow-up questionExcerptJanus · November 2026
Earlier in this thread
Where does palm oil stand when road diesel demand falls away? Who is exposed, where does the land and capacity find new value?
Palm oil's road-fuel demand is created by policy, and it shrinks with diesel volume regardless of how high the mandates go.
When does Indonesia's road diesel demand peak once electric trucks and two-wheelers scale?
On my reading, Indonesia's road diesel demand (fossil plus palm biodiesel) peaks in the early to mid 2030s.
Reform of the diesel subsidy could pull that to around 2030. Electric two-wheelers barely affect the date, because Indonesian motorbikes burn gasoline, not diesel. The clock is set by trucks, and by the subsidy that hides their running cost.
Electric two-wheelers cut gasoline imports, not diesel. Diesel is a freight fuel.
The subsidy decides when diesel peaks, more than battery cost does.
Battery costs are falling fast enough that fleets paying full price switch first, and that is already starting.
Commercial-vehicle battery pack costUS$/kWh
Commercial-vehicle battery pack cost NavigateX projection, 80% band
Sources in Janus: 10 · Written for an outside audience
Leadership cheat sheetExcerptJanus · November 2026
Palm oil after road diesel
The one thing to remember
Palm biodiesel demand is set by a blend rate applied to a diesel volume that is about to start falling, so today's tightness is a floor with an expiry date.
What strands
Plantations replanted now
Biodiesel refining capacity
Blending subsidy budgets (Indonesia, Malaysia)
Where new value forms
Equatorial sun: plantation regions have strong solar all year.
Mill and refinery sites: they already have grid connections, water, heat handling and export logistics, so they suit co-located plants that run on that surplus.
Fermentation on site: solar feeds a data center, and its waste heat runs precision fermentation (a Killer Combo).
Where Janus differs from consensus
Consensus
Janus's view
Structural: "the end of cheap palm oil" (Reuters, Mar 2025)
Temporary and policy-pinned. A 50% blend of a shrinking pool is still a shrinking number of tonnes
Down 3.7% (BMI, Dec 2024)
Same direction for fossil diesel, driven by blending, not electrification
Excerpt: about 150 of 1,400 words
Sources in Janus: 3 · Written for an outside audience
Technology & AI
Strategy question
How should we power our data centers?
For: A data center developer
Does building data centers on clean energy change their cost, speed and strategic position? Who loses if it does?
Brief answerExcerptJanus · November 2026
Does building data centers on clean energy change their cost, speed and strategic position? Who loses if it does?
Yes, on all three counts, but only if the data center is designed around its own solar, wind and battery system rather than plugged into the grid with clean-power contracts on top.
Cost moves from a fuel bill to equipment that keeps getting cheaper. Speed improves because the grid connection is now the slowest step. Strategically, the operator ends up owning cheap surplus power it can sell or use.
Cost: the power bill becomes equipment that gets cheaper every year.
Speed: building your own power is now faster than waiting for the grid.
Strategic position: a site sized for its worst weeks has surplus power most of the year, and that surplus is an asset.
Sources in Janus: 9 · Written for an outside audience
Follow-up questionExcerptJanus · November 2026
Earlier in this thread
Does building data centers on clean energy change their cost, speed and strategic position? Who loses if it does?
Yes, on all three counts, but only if the data center is designed around its own solar, wind and battery system rather than plugged into the grid with clean-power contracts on top.
How big is the surplus from a solar-powered data center campus, and what could it run?
A campus run entirely on its own solar, wind and batteries makes roughly as much surplus electricity as it consumes, and about twice that with a modest extra build.
If the campus uses that surplus instead of discarding it, its own power costs roughly half as much. The best uses are flexible loads that can follow the surplus: batch AI compute, desalination, precision fermentation and storable products like ammonia.
The surplus is about one to two times the campus's own demand.
Using the surplus is what makes the economics.
The surplus favors loads that can wait for it.
Excerpt: about 100 of 600 words
Sources in Janus: 3 · Written for an outside audience
Leadership cheat sheetExcerptJanus · November 2026
Clean-powered data centers
The one thing to remember
Data centers designed around their own solar, wind and batteries win on speed now and on cost increasingly over time. They also gain surplus power to sell or use, while the market is still betting on onsite gas to bridge the wait.
What strands
Onsite gas turbines at data centers
New utility gas plants justified by data center load forecasts
Hubs whose value rests on scarce grid connections
Where new value forms
Speed to market: Owning your own power lets you skip a 4 to 10 year queue. That is worth more than any tariff saving.
The surplus: We call this surplus power Superpower.
Flexible loads that can follow the surplus
Where Janus differs from consensus
Consensus
Janus's view
Onsite gas is the main approach (SemiAnalysis, Sep 2026)
Gas solves this year's speed problem. Solar plus batteries becomes the cheaper path to speed within the life of the turbines (Library cost curves)
Demand rises to ~945 TWh by 2030; renewables meet about half of the growth, with gas, coal and nuclear the rest (IEA, Apr 2025)
Agree on the demand surge. Clean power's share of new self-supplied capacity runs above half wherever land and resource allow
Excerpt: about 190 of 1,800 words
Sources in Janus: 9 · Written for an outside audience
Company question
Is power the limit on our chip growth?
For: A chipmaker's strategy team
AI chip demand on a gas and coal grid: where does a leading Asian chipmaker stand, and which side decides its position by 2030?
Brief answerExcerptJanus · November 2026
AI chip demand on a gas and coal grid: where does a leading Asian chipmaker stand, and which side decides its position by 2030?
The chipmaker sits on the winning side of AI chip demand, but it builds that capacity on a fuel-priced grid that cannot keep up.
By 2030, the chip side decides how much it sells. The energy side decides whether it can build the fabs to sell it, and at what cost. The market is not yet pricing that energy side.
AI demand is outrunning its clean power supply: only about a quarter of its chip division's electricity came from renewables in 2025.
Power is now the constraint that sets the pace. Its new fab clusters need more than 14 GW by 2041, over new long-distance transmission lines.
Solar costs in its home market are falling fast enough to change where the next fabs should go.
Utility-scale solar PV in the chipmaker's home marketUS$/kW
Utility-scale solar PV in the chipmaker's home market NavigateX projection, 80% band
Sources in Janus: 12 · Written for an outside audience
Follow-up questionExcerptJanus · November 2026
Earlier in this thread
AI chip demand on a gas and coal grid: where does a leading Asian chipmaker stand, and which side decides its position by 2030?
The chipmaker sits on the winning side of AI chip demand, but it builds that capacity on a fuel-priced grid that cannot keep up.
Is the gas plant planned for its new chip cluster at risk of stranding before its contract ends?
Yes, but only one of its two businesses is at risk, and it is the one without a contract.
The plant sells steam to the fabs under a direct supply arrangement and sells its electricity into the wholesale market. The steam is the protected core. The electricity faces lower prices and fewer running hours from the day it opens, around 2030.
The electricity half is exposed before the plant has paid itself off.
The steam half holds longer, but it is not safe for the plant's full life.
Policy support will slow the losses but not change the plant's value.
Excerpt: about 100 of 750 words
Sources in Janus: 6 · Written for an outside audience
Leadership cheat sheetExcerptJanus · November 2026
A chipmaker's power problem
The one thing to remember
The national power plan already has gas plants running less than half as often by 2038, and the new chip cluster is adding gas capacity anyway. Battery and solar cost curves pull that decline earlier.
What strands
Wholesale electricity revenue from the cluster's gas plants, exposed from the day they start running
Government gas plants relocated to serve the chip complex, built for a running time they are unlikely to see
The state utility's balance sheet, where capacity payments and price floors end up
Where new value forms
Whoever owns cheap firm clean power next to a fab
Fabs sited in the region with the best solar and wind resource
Battery and solar assets that compete with gas plants for the evening peak
Where Janus differs from consensus
Consensus
Janus's view
Semiconductor power demand reaches 15.4 GW by 2038 (national power plan)
Agree. Demand is large and real.
Gas plants run about 31% of the time in 2030, falling to about 13% by 2038 (implied by the national plan)
The decline starts earlier than 2035. Batteries take over the evening peak role through the 2030s.
Excerpt: about 180 of 1,500 words
Sources in Janus: 7 · Written for an outside audience
Company question
Is our business model exposed to AI agents?
For: A consumer-device company's leadership
As AI moves off the device and off search, where does a global consumer-device and app-store company stand? What is exposed, and where is its new value?
Brief answerExcerptJanus · November 2026
As AI moves off the device and off search, where does a global consumer-device and app-store company stand? What is exposed, and where is its new value?
The cost of machine intelligence per task is falling fast, and that turns the model into a commodity.
Renting a model whose price keeps falling is a sound call. It works only if the company keeps the relationship with the user.
The default-search payment is the most exposed income the company has, and the threat comes from users more than from the courts.
App-store fees are exposed too: an agent that books, orders and pays across services never opens the app.
It owns things a cloud model cannot rent: billions of devices with private data on them, its own chips, payments and identity.
Key numbers
Default search payment, share of services gross profit~21%
Services share of gross profitabout 40%
Services share of salesabout a quarter
Excerpt: about 100 of 600 words
Sources in Janus: 9 · Written for an outside audience
Follow-up questionExcerptJanus · November 2026
Earlier in this thread
As AI moves off the device and off search, where does a global consumer-device and app-store company stand? What is exposed, and where is its new value?
The cost of machine intelligence per task is falling fast, and that turns the model into a commodity.
How much of the company's services margin depends on owning the default, and how fast could agents erode it?
The default search payment is roughly a fifth of the company's services profit.
It is also the profit stream AI agents can wear down fastest, because the price is renegotiated every year, so repricing does not have to wait for search volumes to collapse.
Services earn about 40% of gross profit from about a quarter of sales, and the search payment is the highest-margin part of that.
The payment can fall while search still looks healthy: it is a price per query times the number of queries, and both are under pressure.
Excerpt: about 90 of 500 words
Sources in Janus: 6 · Written for an outside audience
Leadership cheat sheetExcerptJanus · November 2026
A device maker's AI exposure
The one thing to remember
The company's AI exposure is a toll problem. AI agents route around the two places where it charges for what a user wants: the default search slot and the app store. Markets still value its services as a steady, growing income stream.
What strands
The default search payment: close to pure profit and repriced every year
App-store commission on purchases agents make
The valuation premium that steady recurring revenue earns
Where new value forms
Renting a frontier model rather than owning one, which is a wasting asset
The permission layer: identity, payments, private on-device data and the device itself
A toll on outside agents that could replace the search toll
Where Janus differs from consensus
Consensus
Janus's view
Services grow more than 10% a year for five years (analyst consensus, Nov 2026)
Below 10% unless the company builds a toll on agents.
A 1 to 2 point slowdown with little effect on earnings, mainly a valuation issue (broker research, Nov 2026)
Agree for next year. Disagree that it stops there.
Excerpt: about 170 of 1,300 words
Sources in Janus: 10 · Written for an outside audience
Transport & logistics
Supply chain analysis
Where is our supply chain exposed?
For: An auto-parts supplier's executive team · running /analyze-supply-chain
We make fuel injection, exhaust after-treatment and transmission parts for combustion cars and trucks, with plants in Germany, Mexico and China. Break our products into components and materials, trace the EV, battery and autonomy disruptions, and tell us which assets strand, when, and where new value forms through 2035.
Positional Intelligence ReportExcerptJanus · November 2026
Stranding and new value for a combustion-engine parts supplier, 2026 to 2035
Summary
Your China plants are already losing volume. Germany's start losing money from 2027, and Mexico keeps a shrinking tail into the mid-2030s.
A battery-electric vehicle contains none of your fuel-injection or after-treatment content and only a small part of your transmission content, so the shift removes all three product lines together. What survives is the capability inside your plants.
China is past the midpoint of the curve: electric trucks took 50.3% of September heavy-truck sales.
Germany is the most expensive plant to keep open. Losses start from 2027.
Mexico has the longest tail, because its US customers are buying hybrids.
Your capabilities transfer, but your products don't: precision gearing, thermal valves, robot actuators, catalyst-metal recovery.
China heavy trucks: electric share of sales% of sales
202414
202529
2026 Jan–Sep38
Sep 202650
Janus answer, citing Electrek (2 Nov 2026).
Stranding windows by plant
Revenue line
Plant
Most volume gone
Heavy-truck diesel injection
China
2027 to 2030
Passenger-car injection and catalysts
China
2028 to 2031
Multi-speed transmission parts
Germany
2029 to 2031
US-bound injection and catalysts
Mexico
2033 to 2036
Excerpt: about 170 of 4,400 words
Sources in Janus: 24 · Written for an outside audience · Saved as a document in the Space's library
Board paper · Executive team and boardExcerptJanus · November 2026
Running down engine content and redeploying precision capability, 2026 to 2035
Purpose. The board is asked to approve a capital freeze on engine-content lines in China and Germany, moving the remaining combustion volume into Mexico from 2028, reopening customer capacity terms in the 2027 contract cycle, and three staged new-value programs.
Summary for directors
Electric vehicles remove all three product lines at once, so diversifying across them gives no protection.
Each plant strands on its own clock: China now, Germany from 2027, Mexico from about 2029 to 2031.
Our capability transfers even though our products do not.
Recommendation
Stop new capital going into engine content, move the remaining volume to the plant with the longest tail, and redeploy capability before its resale value falls.
Stop growth and replacement capital spending on all China engine-content lines. Freeze now; convert-or-close decisions by 2028.
Freeze German passenger-car engine-content investment and move grinding and heat treatment to electric-drive gearing. From 2027.
Make Mexico the consolidation site for remaining combustion volume, run lean. From 2028.
Risks and mitigations (excerpt)
Risk
Mitigation
Single-source sub-suppliers cut capacity early
Map them in 2027; arrange last-time buys, second sources or service-parts stock.
Orders run below contracted lifetime volumes
Tie capacity to reservation fees or shortfall compensation in the 2027 cycle.
Used machine-tool values collapse as every supplier sells at once
Redeploy machines internally as lines close.
Excerpt: about 200 of 2,800 words
Sources in Janus: 4 · Written for an outside audience · Saved as a document in the Space's library
Leadership cheat sheetExcerptJanus · November 2026
Engine-content rundown and redeployment, 2026 to 2035
The one thing to remember
Electric vehicles remove injection, after-treatment and most transmission content at the same time, so our plants strand on three different clocks while our precision capability carries forward.
What strands, where, and when
China heavy-truck diesel injection: most volume gone 2027 to 2030
Germany multi-speed transmission parts: losses from 2027
Mexico US-bound injection and catalysts: most volume gone 2033 to 2036
Where new value forms
Electric-drive and heavy-truck gearing
Thermal-management valves and pumps
Plant sites and their grid connections
Where Janus differs from consensus
Consensus
Janus's view
China's government target: 40% electric heavy-truck sales by 2030 (June 2026)
Already obsolete: monthly share passed 50% in September 2026.
The EU's softer 2035 rule means a longer combustion life (committee vote, Nov 2026)
No reprieve for German volume. Buyers already register more battery-electric cars than combustion ones.
Excerpt: about 130 of 1,800 words
Sources in Janus: 5 · Written for an outside audience
Industry exposure
What happens to the cargo we carry?
For: A shipping and ports group
When oil, coal and gas cargo falls, where does shipping stand? Which fleets and ports strand, where does new demand form?
Brief answerExcerptJanus · November 2026
When oil, coal and gas cargo falls, where does shipping stand? Which fleets and ports strand, where does new demand form?
Shipping is among the most exposed industries on the incumbent side of the energy disruption, and it barely appears on most energy-risk maps.
Fossil fuels and petrochemicals make up about 40% of seaborne trade by weight: 4,500 million tons of roughly 11,000 million, per IENE's September 2025 presentation citing UNCTAD figures.
Why the loss is structural. The useful measure is the cost of each kilowatt-hour delivered. For imported fuel, freight is a recurring part of that cost.
Single-cargo assets with decades of life ahead are the most exposed, and LNG is the most mispriced.
New value goes to ports that become places of production, not to ships that replace the lost tons.
Key numbers
Fossil fuels and petrochemicals in seaborne trade4,500 Mt of ~11,000 Mt (~40%)
Seaborne coal1,333 Mt (EURACOAL: nearer 1,200 Mt)
Seaborne crude2,224 Mt
Global LNG demand422 Mt
Excerpt: about 110 of 750 words
Sources in Janus: 3 · Written for an outside audience
Follow-up questionExcerptJanus · November 2026
Earlier in this thread
When oil, coal and gas cargo falls, where does shipping stand? Which fleets and ports strand, where does new demand form?
Shipping is among the most exposed industries on the incumbent side of the energy disruption, and it barely appears on most energy-risk maps.
Do new mineral cargoes come close to replacing lost coal and oil tonnage?
No. The new mineral cargoes from the energy disruption are one to two orders of magnitude too small to replace lost coal and oil tonnage.
Even the full material needs of the new energy system, spread over three decades, ship fewer tons each year than a small slice of today's coal trade. For dry bulk carriers and tankers, minerals are a niche, not a replacement.
About 3.5 billion tons of coal and crude moved by sea in 2025, against roughly 0.2 billion tons a year of materials for the whole new energy system.
Two caveats point in opposite directions. Ore weighs more than refined metal. The IEA's Global EV Outlook 2026 counts about 8 billion tons of critical-mineral ore mined in 2024, about 85% of it copper.
Mineral demand comes as a one-time build-out that recycling later shrinks, while fuel is shipped again every year.
Fossil cargo against new mineral cargobillion tons a year
Coal and crude moved by sea, 20253.5
Materials for the whole new energy system, per year0.2
Figures from the Janus follow-up answer (2025).
Excerpt: about 150 of 600 words
Sources in Janus: 6 · Written for an outside audience
Leadership cheat sheetExcerptJanus · November 2026
Shipping after fossil cargo
The one thing to remember
Fossil fuels are about 40% of seaborne tonnage, and the solar, wind and battery disruption removes most of that cargo for good. Minerals replace the value, not the tonnage.
What strands
Coal carriers and coal export terminals
Crude and product tankers: The direction is certain and the date is not. A ship ordered in 2026 trades into the 2050s.
LNG carriers and import terminals
Where new value forms
Ports with strong sun and wind
Mineral export and processing ports
Specialized and project-cargo shipping: Turbines, modules, batteries and refined metals.
Where Janus differs from consensus
Consensus
Janus's view
IEA Current Policies Scenario: rises to 113 mb/d by 2050 (IEA WEO 2025)
Road fuel peaks and declines as EVs climb the adoption S-curve. I give no date.
Shell: +54–68% by 2040 and +45–85% by 2050, from 422 Mt (Business Recorder, Mar 2026)
Volumes plateau well below this. The growth markets in Asia are where solar and batteries are cheapest.
Excerpt: about 150 of 1,000 words
Sources in Janus: 5 · Written for an outside audience
Also asked here: “What do self-driving fleets do to our assets?” (answered under Finance & investment).
Finance & investment
Convergence cascade analysis
Where will new markets appear, and what strands?
For: An infrastructure investor's investment committee · running /analyze-convergence-cascade
As the cost curves of batteries, electric drivetrains, solar and self-driving converge, trace the new category that emerges, which of our real assets it strands (fuel retail, parking, car finance, toll and transit), and the second- and third-order categories that cascade, to 2040.
Positional Intelligence ReportExcerptJanus · November 2026
Robotaxi Fleets and Real Assets to 2040: What Strands, What Gains
Summary
Robotaxi service is already commercial in a few dozen cities. Cost no longer sets its pace; regulatory approval does.
Battery packs crossed their cost threshold in 2024 to 2025. For an infrastructure portfolio, value is lost well before adoption makes headlines: assets priced on private car ownership start to lose value once robotaxis carry a few percent of urban miles.
Ownership-priced assets lose value early. Airport parking goes first, from the late 2020s.
The contract decides exposure more than the asset type. Availability-payment roads are largely insulated.
The value moves to what fleets need: depots, grid connections, charging sites, priced curb access and fleet finance.
Consensus sizes robotaxis against today's ride-hailing market. Expect continued upward revisions.
Lithium-ion battery costUS$/kWh, log scale
Lithium-Ion Battery Li-Ion Pack (Passenger BEV)
NavigateX Technologies Library. Recorded data points.
When ownership-priced assets start losing value
Asset
What fleets remove
Loss starts
Airport parking
The trip becomes a curbside drop-off
2027 to 2031
Downtown and event parking
Storing cars near the destination
2029 to 2034
Car finance and lease residuals
Private ownership
Late 2020s to early 2030s
Demand-risk transit concessions
Off-peak riders move to robotaxis
Early to mid 2030s
Excerpt: about 190 of 3,800 words
Sources in Janus: 26 · Written for an outside audience · Saved as a document in the Space's library
Board paper · Investment committeeExcerptJanus · November 2026
Robotaxi Fleets and Real Assets to 2040
Purpose. This paper asks the committee to approve six portfolio actions in response to self-driving electric fleets that sell transport by the mile, and to the assets those fleets weaken or strengthen to 2040.
Summary for directors
Self-driving electric fleets are commercial, and regulatory approval now sets their pace, not cost.
Assets priced on private car ownership lose value once robotaxis carry a few percent of urban miles.
Consensus forecasts are sized against ride-hailing, so they will keep rising.
Recommendation
Approve six actions, staged so the portfolio reprices ownership-priced assets now and builds fleet-facing positions as approvals arrive.
Within 12 months, re-underwrite airport and downtown parking on a robotaxi scenario, and negotiate curb-access fees.
Before the 2027 investment cycle, sort road and transit holdings by contract type and reprice demand-risk concessions.
From 2027 to 2030, secure depot sites, grid connections and charging sites in metros approaching approval.
Risks and mitigations (excerpt)
Risk
Mitigation
A serious driverless crash leads to national suspensions
Act through re-underwriting and options, not forced sales.
Waymo stays near 500,000 weekly rides into 2027
Trigger: a lasting plateau pushes parking and forecourt timing out one to two years.
Selling ownership-priced assets too early, or overpaying for depot land
Convert rather than exit where possible; secure depot land through options.
Excerpt: about 200 of 2,300 words
Sources in Janus: 4 · Written for an outside audience · Saved as a document in the Space's library
Leadership cheat sheetExcerptJanus · November 2026
Robotaxi fleets and real assets to 2040
The one thing to remember
Cost no longer decides when self-driving electric fleets arrive; regulatory approval does. Assets priced on private car ownership start losing value once fleets carry a few percent of urban miles.
What strands, and when
Airport parking: leading metros 2027 to 2031
Downtown and event parking: leading metros 2029 to 2034
Car finance, leases and car-loan bonds: from the late 2020s
Where new value forms
Fleet depots, charging and battery-swap sites, where grid connections are the scarce input
Priced curb access for robotaxi pick-ups and drop-offs
Fleet finance secured on per-mile revenue rather than resale value
Where Janus differs from consensus
Consensus
Janus's view
US robotaxi revenue of US$19 billion in 2030, up from US$7 billion in under a year (Goldman Sachs Research)
We agree: 2030 is still the early part of the curve.
A US$415 billion global market in 2035, sized against ride-hailing (Goldman Sachs Research, Apr 2026)
Higher in leading metros. Once a robotaxi mile costs less than an owned-car mile, the market is all private car travel.
Excerpt: about 170 of 1,800 words
Sources in Janus: 8 · Written for an outside audience
Company question
Is this company's order book safe?
For: An analyst covering a gas-turbine maker
Gas turbine orders are being filled by data centers. Where does a leading gas-turbine maker's backlog stand against solar, wind and storage over the turbines' life?
Brief answerExcerptJanus · November 2026
Gas turbine orders are being filled by data centers. Where does a leading gas-turbine maker's backlog stand against solar, wind and storage over the turbines' life?
The gas backlog is safe as revenue through about 2032. The risk sits in the second half of its life.
Turbines delivered in 2029 to 2032 are built to run into the 2050s. Long before then, I expect solar plus batteries to deliver firm power more cheaply than they do. Expect the damage to show up first in how many hours these plants run, then in services revenue, and only later in new orders.
The order book is real, and it pays early.
Data centers buy turbines for speed and firm power, not low cost, and batteries are eroding both reasons.
For the turbine maker, the exposed part is services, not equipment.
Capital cost: gas plants against solarUS$/kW
Gas (CCGT) Solar PV (Utility-Scale)
NavigateX Technologies Library.
Excerpt: about 110 of 800 words
Sources in Janus: 8 · Written for an outside audience
Follow-up questionExcerptJanus · November 2026
Earlier in this thread
Gas turbine orders are being filled by data centers. Where does a leading gas-turbine maker's backlog stand against solar, wind and storage over the turbines' life?
The gas backlog is safe as revenue through about 2032. The risk sits in the second half of its life.
How fast would falling utilization erode the company's services backlog under different battery cost paths?
The services backlog wears down slowly, then faster. I don't expect material damage before the early 2030s on the central battery path.
The cost path for batteries mainly changes when the damage starts, by about five years either way. It barely changes where it starts.
Services revenue follows hours run, so fewer hours means less money from the same contract.
The first plants hit will be the older gas turbines that sell into the grid, not the new units ordered by data centers.
The battery cost path sets the timing.
Excerpt: about 90 of 750 words
Sources in Janus: 6 · Written for an outside audience
Leadership cheat sheetExcerptJanus · November 2026
A gas-turbine maker's backlog
The one thing to remember
The gas backlog is safe as deliveries through about 2032 but exposed in how many hours the turbines run over their lives. The risk is not in the orders but in the run hours, and the services revenue that depends on them.
What strands
Hour-based gas service revenue, from the early 2030s on the central battery path
Slot reservations, the softest part of the order book
Gas plants bought on 2020s assumptions about running hours
Where new value forms
Grid equipment, which grows whatever generation gets built
Solar plus batteries for data centers, which wins on speed to power before it wins on cost
Fixed-fee capacity and backup contracts for gas units moved to standby duty
Where Janus differs from consensus
Consensus
Janus's view
The contracted gas backlog keeps growing through year-end (company guidance, Oct 2026)
Agree. Demand is real through 2032.
Most analysts rate the stock a Buy and expect earnings to more than double (Oct 2026)
Differ beyond 2030. Near-term earnings hold. Services erosion begins in the early 2030s on the central battery path.
Excerpt: about 170 of 1,500 words
Sources in Janus: 9 · Written for an outside audience
Research update
Where do the new weight-loss drugs stand?
For: A healthcare investor
What has moved since the RethinkX healthcare report? Where do its technologies stand now?
Brief answerExcerptJanus · November 2026
What has moved since the RethinkX healthcare report? Where do its technologies stand now?
Since RethinkX published its health report in March 2026, the cost curve and the pill format have moved faster than it assumed. The muscle-preserving combination it saw as the "iPhone moment" has not.
In RethinkX's Rethinking Health report, Health Optimizer Therapies (HOTs) are drugs that cut weight and metabolic disease. The pressure on the old system is now coming from price and access, not from a new breakthrough.
Prices are now falling the way the report said they would, through generics and new formats.
Pills arrived on schedule, which widens who can use these drugs.
Next-generation drugs keep improving with each generation. Retatrutide, which acts on three hormone targets at once, has now succeeded in six late-stage (Phase 3) trials.
What a month of semaglutide costs in the USUS$ per month
List price from Jan 2027675
Lowest US cash price, Oct 2026149
Eligible Medicare patients~50
Figures cited in the Janus answer (Oct 2026 tracker; Jan 2027 list price; Medicare).
Excerpt: about 120 of 700 words
Sources in Janus: 17 · Written for an outside audience
Follow-up questionExcerptJanus · November 2026
Earlier in this thread
What has moved since the RethinkX healthcare report? Where do its technologies stand now?
Since RethinkX published its health report in March 2026, the cost curve and the pill format have moved faster than it assumed. The muscle-preserving combination it saw as the "iPhone moment" has not.
How fast could generic semaglutide pricing in India and Canada spread to the US and EU markets?
Generic prices themselves can't legally reach the US or EU until about 2031. Price pressure will arrive well before that, through the brand-name drugs and through competing products.
The patent wall sets the date for generics. It does not set the date for falling prices, and those two dates are years apart.
India and Canada show where prices go once patents expire.
The US and EU generic dates are fixed by patents, with little room to move earlier.
In the meantime, the US price is falling through government negotiation and brand price cuts.
Excerpt: about 90 of 650 words
Sources in Janus: 10 · Written for an outside audience
Leadership cheat sheetExcerptJanus · November 2026
Weight-loss and metabolic drugs
The one thing to remember
A month of weight-loss and metabolic drug treatment is getting cheaper through generics, pills, government price negotiation and brand price cuts at the same time.
What strands
Weight-loss services: already stranded. The RethinkX report records the collapse of WeightWatchers and Jenny Craig.
Brand pricing power on semaglutide
Snacks, fast food and other discretionary food
Where new value forms
Generic ingredient and finished-drug makers
Oral small-molecule manufacturing: cheap to make, needs no refrigerated supply chain. This is the route to global volume.
Peptide production by precision fermentation and continuous manufacturing
Where Janus differs from consensus
Consensus
Janus's view
US$130 to 135B by 2030 (UBS, Jun 2026); US$200B by 2030 (J.P. Morgan, Feb 2026)
Differs. Revenue understates the disruption. Patient numbers grow faster, and price per month falls further, than these forecasts assume.
about 25M Americans on treatment by 2030 (J.P. Morgan, Feb 2026)
Differs, toward higher. Bain (Nov 2026) expects nearly one in three US adults to have used one by 2030.
Excerpt: about 160 of 1,600 words
Sources in Janus: 7 · Written for an outside audience
These are real Janus answers written in November 2026 for an outside audience, edited for length. Company examples are composites or have their names removed. Projections are NavigateX model forecasts, not published prices.