Markets move whenassumptions break.

Fulcrius maps the assumptions beneath your investments and watches for the moment they start to break.

AI research agents follow the filings, the supply chains, and the news, then surface the changes heading toward your portfolio.

Every investment thesis contains assumptions.

Demand keeps growing.
A technology succeeds.
A regulation remains in place.
A supplier keeps delivering.

Most investors never write them down.

They discover them only after they break.

What does your portfolio actually depend on?

For example: what happens to your holdings if the Strait of Hormuz stays disrupted for six months?

Here is what an investment thesis looks like inside Fulcrius.

A fictional example follows.

An inspectable research example

Illustrative example

Two fictional companies with different risk and opportunity profiles.

Krevnor DynamicsDembrack Materials

Krevnor Dynamics (fictional company)

Advanced semiconductor packaging equipment

Krevnor makes packaging and test equipment for high-end AI processors. It is profitable, tied to the datacenter buildout, and positioned between a small group of materials suppliers and several very large customers.

Three scores, kept separate

Opportunity 7 high Demand remains strong, but a rival's supplier has narrowed the moat.

This is not an average. The synthesis weighs each opportunity dimension, its confidence, and the recent changes. Krevnor's 7 was an 8 until last week. Competitive position fell from 7 to 5 after Fulcrius tied an interconnect-yield jump at a supplier two tiers below Vossen Systems, Krevnor's closest rival. Structural demand (8) still carries the score, but the competitive lead is narrower than previously assessed.

Based on

Derived from the 11 dimension scores
Recent changes, last 10

The 4 inputs behind this score

Score Confidence

Structural demand 8 very high
Demand is growing faster than the industry can add supply. Backlog rose 22% quarter over quarter, and two hyperscalers extended multi-year contracts. Confidence is very high because the evidence comes from signed contracts and reported backlog, not management commentary.

Sources
- Q3 earnings release
- Customer backlog disclosure
- Signed contract filings
Competitive position 5 high
This is about durability, not size. The score was 7 until Helix Optics, a supplier on Vossen Systems' new packaging line, reported a jump in high-bandwidth interconnect yield from 62% to 89% in one quarter. That closes most of the throughput gap Krevnor had relied on. The moat remains real, visible in gross margins above the sector, but the lead now appears measured in quarters, not years.

Sources
- Helix Optics process certification
- Vossen customer qualification filing
- Market share report
Execution 6 high
The recent record is solid, though not without exceptions. Revenue met guidance, but one manufacturing milestone slipped by a quarter. One delay is manageable; repeated delays would change the score. Confidence is high because the evidence comes from the earnings call, guidance history, and insider transaction filings.

Sources
- Q3 earnings call
- Guidance history
- Insider transaction filings
Valuation 7 medium
A higher valuation score means cheaper relative to fundamentals. Krevnor trades below the median EV/Sales of its peers while growing faster, so the market may be underpricing its earnings power. Confidence is medium because the multiple can move quickly when the share price or analyst estimates shift.

Sources
- Financial data feed
- Peer EV/Sales comp set

Risk 6 medium Cyclicality and a new supply-chain exposure drive most of the risk.

Risk is deliberately not averaged. The worst exposures are allowed to lead instead of being smoothed away. Cyclicality (7) and the newly raised geopolitical risk (6) set Krevnor's level, even though most other risks are low. Confidence stays medium because those two inputs are not yet high confidence.

Based on

Derived from the 11 dimension scores
Recent changes, last 10

The 7 inputs behind this score

Score Confidence

Technology risk 3 very high
Technology risk is low because the packaging process is already running in volume. Recent filings show no unresolved yield or physics issues. Certification records and reported manufacturing yields make the confidence very high.

Sources
- Product certification records
- Manufacturing yield data
Platform risk 4 medium
A competing architecture could make part of Krevnor's product line less relevant. The company sells into several AI architectures, which helps, but a meaningful slice of revenue depends on one interconnect standard prevailing. Customer mix and the wider architecture race keep confidence at medium.

Sources
- Customer mix disclosure
- Architecture dependency map
Financing risk 2 very high
Financing risk is very low. Krevnor has generated positive free cash flow for three straight quarters and has neither announced nor implied an equity raise. The cash flow statement and share count history leave little room for interpretation.

Sources
- Cash flow statement
- Share count history
Geopolitical risk 6 low
This score recently moved to 6. A critical input appears to come from one constrained region, three tiers upstream in the supply chain, an exposure that was not visible at the company level. More of Krevnor's supply, and therefore its cost base, now depends on that region staying stable. Confidence is low because the dependency is newly mapped and rests on softer evidence than audited financials.

Sources
- Supplier 10-K, capacity section
- Supply chain map
- Export control bulletin
Regulatory / political risk 3 high
Regulatory risk is low. Krevnor's products are approved across its major markets, and no pending decision would block shipments. The approval registry and the company's legal disclosures support the high confidence.

Sources
- Regulatory approval registry
- Legal disclosures
Cyclicality risk 7 medium
Cyclicality is elevated because most of Krevnor's revenue rides on datacenter capex. A sharp slowdown in AI infrastructure spending would reach the order book quickly. Confidence is medium because hyperscaler capex plans are estimates, not locked contracts.

Sources
- Revenue by end market
- Hyperscaler capex tracker
Liquidity / volatility risk 4 high
Krevnor is a mid-cap with adequate daily volume. Its beta runs above the broad market, but that is normal for the sector. The risk is moderate, and confidence is high because average volume and beta are directly measurable.

Sources
- Average daily volume
- Beta vs. sector index

Fragility 2 sturdy A bad year would bend Krevnor, not break it.

Fragility asks whether a shock is survivable. At 2, Krevnor looks sturdy: a bad year would bend it, not break it.

Fragility 2 of 11, from these factors

Cash runway under 18 months Not triggered max +3
Cash on hand covers more than three years at the current burn. A downturn would not automatically force a raise.

Sources
- Balance sheet
- Cash flow statement
Refinancing dependent Not triggered max +2
Positive free cash flow means Krevnor does not need access to capital markets to keep operating.

Sources
- Cash flow statement
Customer concentration over 50% Not triggered max +2
The largest customer is about a quarter of revenue. Losing it would hurt, but it should not cripple the business.

Sources
- Revenue by customer
Binary regulatory outcome pending Not triggered max +2
No single approval or ruling could end the business; products are already cleared in major markets.

Sources
- Regulatory disclosures
Single-product dependency over 80% Adds 1 max +1
Most revenue comes from one equipment line. A generational shift in packaging would concentrate the damage.

Sources
- Revenue by product line
Manufacturing single point of failure Adds 1 max +1
Final assembly runs through one facility, which a localized disruption could halt.

Sources
- Facility disclosures

Explore the dependency network

This example is fictional, so linked companies are not available here.

What changed

Opportunity Rival's optics supplier hit 89% yield

Helix Optics, two tiers below Krevnor's closest rival Vossen Systems, reported a jump in interconnect yield that was not visible in either company's filings. Fulcrius tied the link back through the competitor network, lowered competitive position from 7 to 5, and pulled the opportunity synthesis from 8 to 7. Nothing in Krevnor's own backlog changed. The score moved because a moat that looked years wide now looks like quarters.

Rival's optics supplier hits 89% yield
Vossen closes throughput gap on Krevnor
Competitive position lowered to 5
Opportunity synthesis drops from 8 to 7

Sources

Helix Optics process certification
Vossen customer qualification filing
Competitor dependency map

Risk Deep supplier cut left one region in control

A supplier three tiers upstream of Krevnor cut capacity, leaving a critical input effectively single source from one region. The dependency was invisible at the company level. Fulcrius followed the path upstream, raised geopolitical risk from 5 to 6, and flagged the new exposure. The score moved because more of Krevnor's supply now depends on one region staying stable, not because its own results changed.

Deep supplier cuts capacity
Critical input turns single source
Supply now hinges on one region
Geopolitical risk raised to 6

Sources

Supplier 10-K, capacity section
Industry capacity report
Krevnor supply chain map

What you said matters

You told Fulcrius you care about privacy. In Krevnor's customer network, that points to Orsalen Compute, one of its largest buyers. Orsalen builds the inference servers behind a national face-recognition program that human-rights groups have tied to surveillance of a minority population, and Krevnor's equipment packages the chips that make those servers run.

This is shown because you named the principle, not because Fulcrius judges it. You decide whether it changes your view.

Dembrack Materials (fictional company)

Critical-materials processing

Dembrack processes critical materials for high-performance magnets and power electronics. It is profitable and growing quickly, but concentration is material: one main site, one jurisdiction, and one customer that accounts for more than half of revenue.

Three scores, kept separate

Opportunity 7 high Contracted demand remains strong, but one customer has found a way to use less of its input.

This is not an average. The synthesis weighs each opportunity dimension, its confidence, and the recent changes. Dembrack's 7 was an 8 until last month. Structural demand fell from 9 to 7 after Fulcrius connected Torva Motors' rare-earth-free magnet breakthrough to qualification filings and patent data. Competitive position (8) still holds. The long-run demand ceiling looks lower.

Based on

Derived from the 11 dimension scores
Recent changes, last 10

The 4 inputs behind this score

Score Confidence

Structural demand 7 high
Demand was a 9 until Torva Motors' rare-earth-free permanent magnet hit production targets two years ahead of the industry roadmap. Torva is more than half of Dembrack's revenue, and the new design needs far less of Dembrack's primary processed material. Near-term demand still grows under existing contracts, but the long-run ceiling on this input has declined.

Sources
- Torva Motors patent filing
- Third-party magnet teardown
- Customer requalification notices
Competitive position 8 high
Few companies can process these materials at scale outside the dominant producing nation. Requalifying a new processor is slow and costly for customers. That gives Dembrack a strong position that is difficult to replicate.

Sources
- Processor market share
- Customer qualification records
Execution 7 high
Dembrack is profitable and has scaled output on schedule for several quarters, which is rare for a materials company at this stage. Management is delivering, with only normal operational lumpiness. Reported output, margins, and the earnings call support the high confidence.

Sources
- Quarterly output
- Earnings call
Valuation 4 low
A higher valuation score means cheaper relative to fundamentals. Dembrack looks expensive: the shares rerated sharply on the supply-shortage narrative, so plenty of good news is already in the price. Confidence is low because commodity prices and sentiment can move the multiple quickly.

Sources
- Financial data feed
- Commodity price series

Risk 8 high One exposure dominates the rest: geopolitical risk at 9.

Risk is deliberately not averaged. One catastrophic exposure should not disappear because the other risks look calm. Here, geopolitical risk at 9 does most of the work. A straight average of the seven risks would suggest a 5, which would badly understate a company that one policy decision could materially impair.

Based on

Derived from the 11 dimension scores
Recent changes, last 10

The 7 inputs behind this score

Score Confidence

Technology risk 2 very high
Technology risk is low. The processing chemistry is mature and has been proven over years of production. Long-run production records make the confidence very high.

Sources
- Production records
- Process certifications
Platform risk 3 high
Platform risk is low because Dembrack sells a refined input used across many end products. It does not need one specific technology to win. The input is already qualified across multiple industries.

Sources
- Customer mix by industry
- Product qualifications
Financing risk 4 high
Dembrack is profitable, but processing capacity is expensive to expand and some of that expansion uses debt. A downturn would strain the expansion, though not the core business. The balance sheet and capex plans make this a high-confidence view.

Sources
- Balance sheet
- Capex guidance
Geopolitical risk 9 high
Almost all processing runs through one jurisdiction. A newly proposed export rule would require a license to keep serving Dembrack's largest market. This is near-existential: one policy decision could materially impair the business. The concentration is documented, and the proposed rule is public.

Sources
- Facility disclosures
- Proposed export rule
- Supply chain map
Regulatory / political risk 6 low
Beyond export policy, Dembrack depends on operating permits and subsidies that can shift with local politics. The risk is meaningful but not central. Confidence is low because the pending decisions are genuinely hard to predict.

Sources
- Permit filings
- Policy trackers
Cyclicality risk 6 medium
Dembrack's revenue rides on commodity prices and industrial capex, both cyclical by nature. A downturn would dent revenue without ending the business. Confidence is medium because commodity-price paths are estimated, not fixed.

Sources
- Commodity price series
- End-market capex
Liquidity / volatility risk 5 medium
Dembrack is a mid-cap whose shares move sharply with commodity headlines. Exits during a panic could be costly. Beta and volume are measurable, but headline-driven moves are less predictable, so confidence stays medium.

Sources
- Average daily volume
- Beta vs. sector

Fragility 6 fragile Enough single points of failure stack up that one bad break could be permanent.

At 6, Dembrack is fragile. The business is profitable, but the single points of failure accumulate. One adverse event could become permanent, regardless of how strong the opportunity appears.

Fragility 6 of 11, from these factors

Cash runway under 18 months Not triggered max +3
Dembrack is profitable, so it is not drawing down a finite cash pile to survive.

Sources
- Cash flow statement
Refinancing dependent Not triggered max +2
It funds operations from its own cash flow, though expansion does lean on some debt.

Sources
- Balance sheet
Customer concentration over 50% Adds 2 max +2
One customer is more than half of revenue, so a single lost contract would be severe.

Sources
- Revenue by customer
Binary regulatory outcome pending Adds 2 max +2
A pending export-license decision could cut Dembrack off from its largest market in one stroke.

Sources
- Proposed export rule
- Legal disclosures
Single-product dependency over 80% Adds 1 max +1
Almost all revenue comes from one processed material.

Sources
- Revenue by product
Manufacturing single point of failure Adds 1 max +1
Processing is concentrated in one site, with no qualified backup yet.

Sources
- Facility disclosures

Explore the dependency network

This example is fictional, so linked companies are not available here.

What changed

Opportunity Torva's magnet needs far less processed input

Torva Motors, Dembrack's largest customer, filed patents showing a rare-earth-free permanent magnet hitting production spec two years ahead of the industry roadmap. The shift was invisible in Dembrack's own filings. Fulcrius followed it through customer qualifications and teardown data, lowered structural demand from 9 to 7, and pulled the opportunity synthesis from 8 to 7. Near-term offtake contracts still stand. The score moved because the long-run demand this business was priced for may no longer arrive.

Largest customer files rare-earth-free magnet patent
New design needs less of Dembrack's primary input
Structural demand lowered to 7
Opportunity synthesis drops from 8 to 7

Sources

Torva Motors patent filing
Third-party magnet teardown
Customer requalification notices

Risk Export license may be required for largest market

A government proposed new export-licensing rules covering Dembrack's processing jurisdiction. Overnight, serving its largest market could require a license that may not be granted. Dembrack's geopolitical risk moved from 7 to 9, with single-jurisdiction exposure becoming the dominant driver of the overall risk score.

New export rule proposed
License required for main market
Single-jurisdiction exposure
Geopolitical risk raised to 9

Sources

Proposed export rule
Facility disclosures
Supply chain map

What you said matters

You told Fulcrius you will not back severe environmental harm. In Dembrack's supplier network, that points to a mine feeding its main site, in a country that rarely enforces its own rules. The mine's tailings dam failed last year, sending arsenic and lead into the river two downstream villages drink from. Fish died for miles, the water is still unsafe, and no cleanup has begun.

This is shown because you named the principle, not because Fulcrius judges it. You decide whether it changes your view.

Every score should be explainable. Every change should be traceable.

Click a question to see how Fulcrius traces the answer.

You asked

What happens to my portfolio if the Strait of Hormuz stays disrupted for another month? Six months? A year?

Affected dependency network

Research findings

You own an AI-hardware supplier and a small industrial-gases position. The first-order exposure is negative. Advanced fabs use ultra-pure helium for EUV lithography cooling and leak detection, and Qatar is a major supplier shipping through Hormuz.

The answer is not simply "everything falls." If roughly a third of global helium supply is constrained, remaining supply becomes more valuable. U.S., Algerian, Canadian, and other non-Gulf producers cannot replace Qatar overnight, but they may gain pricing power and allocation value.

The same shock separates into three effects: chip-dependent holdings become more fragile, Qatar-exposed gas distributors carry higher operating risk, and non-Gulf helium producers or storage assets may see opportunity scores rise.

How scores move

Advanced-chip holdings Fragility increases The same upstream gas shortage reaches multiple foundries at once.
Qatar-exposed gas distributors Risk increases Supply allocation, customer penalties, and spot-price volatility become central.
Non-Gulf helium producers Opportunity increases The constraint can raise pricing power, though confidence stays medium because spare capacity is limited.

Portfolio relevance

The useful answer is not only what breaks. It is also whether any holding benefits from the disruption, and whether that offset is material at the portfolio level.

You asked

Which companies become most exposed if tensions around Taiwan escalate?

Affected dependency network

Research findings

You hold Nvidia, Apple, and an S&P 500 index fund. On a brokerage screen, those look like separate exposures. The manufacturing dependency converges on one company: TSMC in Taiwan, maker of more than 90% of the world's most advanced chips. Nvidia and Apple are its two largest customers, with AMD, Qualcomm, and Broadcom tied in as well.

A Taiwan escalation would affect TSMC-dependent designers first. It would also raise the strategic value of alternatives: non-Taiwan foundry capacity, U.S. and Japanese fab construction, advanced packaging outside Taiwan, and the equipment suppliers needed to build them.

Those alternatives are not direct substitutes. A leading-edge fab takes years, billions, and rare technical expertise. Opportunity scores may rise, but confidence should remain measured. Holdings with no credible fallback become more fragile.

How scores move

TSMC-dependent chip designers Fragility increases sharply A design win is not enough if the only viable manufacturing path is disrupted.
Alternative foundries Opportunity increases Customer diversification becomes urgent, but execution risk keeps confidence below high.
Reshoring infrastructure Opportunity increases Policy support and customer urgency can pull forward fab, packaging, and equipment spend.

Portfolio relevance

The first question is concentration. The second is substitution: which holdings rely on Taiwan, and which holdings gain strategic value because so many others do.

You asked

Which companies benefit most if nuclear energy expands to meet AI and grid demand?

Affected dependency network

Research findings

Data-center power demand and electrification are converging on a problem wind and solar alone cannot solve: reliable, high-density baseload. Governments in the U.S., Europe, Japan, and South Korea have shifted policy toward nuclear restarts, lifetime extensions, and new small modular reactors. That policy shift feeds through a supply chain that has been underinvested for over a decade.

Uranium supply is structurally tight. Mine output has not matched reactor demand for decades, with the gap filled by drawdowns from inventories and secondary supply that are themselves declining. Spot prices roughly tripled between 2020 and early 2024, briefly exceeding $100 per pound, and long-term contract prices have continued to reach multi-year highs.

The opportunity does not stop at uranium. Fuel fabrication, enrichment services, reactor maintenance, and specialized engineering benefit from the same expansion. If you hold energy or infrastructure positions, some of them may carry more nuclear exposure than a sector label suggests.

How scores move

Uranium miners and explorers Opportunity increases Structural supply deficit meets rising demand from restarts, extensions, and new builds.
Nuclear utilities Opportunity increases License extensions and policy support improve long-term revenue visibility.
SMR developers Opportunity increases, risk remains elevated First-of-a-kind execution risk is real, but customer commitments and policy tailwinds are growing.

Portfolio relevance

A nuclear expansion is not a single trade. It runs through miners, fuel services, utilities, and equipment makers. Fulcrius traces which of your existing holdings already carry nuclear exposure and where the opportunity chain extends beyond the obvious names.

Sources

Global uranium mine production has not matched reactor requirements for decades; the deficit is filled by inventory drawdowns, reprocessing, and other secondary supply that is itself declining. World Nuclear Association, Supply of Uranium
U3O8 spot prices rose from roughly $25/lb in early 2020 to above $100/lb in January 2024, with long-term contract prices reaching multi-year highs through 2025. Cameco uranium price data
38 countries have endorsed the Declaration to Triple Nuclear Energy by 2050, launched at COP28; the U.S., Japan, and South Korea have announced restart and extension programs. U.S. Department of Energy / World Nuclear Association
Microsoft, Amazon, Google, and Meta signed nuclear power agreements in 2024-2025 to secure baseload electricity for AI data centers, representing over 6 GW of potential capacity. Carnegie Endowment for International Peace, 2026

You asked

Can I build a portfolio around the companies shaping the future I believe in, without giving up returns?

Affected dependency network

Research findings

You hold TAN, an Invesco solar ETF, and you care about clean energy. The supply chain adds a material consideration: China's share in every key solar manufacturing stage exceeds 80%. Xinjiang has accounted for roughly a third of global solar-grade polysilicon output in recent analyses, and forced-labor concerns in the region are well documented.

The fund label therefore tells only part of the story. The ETF is not doing anything wrong by existing. But if forced labor in the supply chain matters to you, there is a conflict several layers below the headline exposure.

First Solar shows up as a different path through the same theme. It uses cadmium-telluride thin-film technology and says its modules do not rely on Chinese crystalline-silicon supply chains. Same clean-energy belief, different dependency map.

How scores move

TAN (solar ETF) Alignment decreases Alignment measures fit with your stated values. The clean-energy label hides a supply chain with documented human-rights concerns.
First Solar Alignment increases Same theme, no polysilicon dependency. Different supply chain, different risks.
TAN (solar ETF) Fragility increases 80%+ concentration in one country's manufacturing also makes it a supply-chain risk.

Portfolio relevance

Values become more useful when they are tied to the actual supply chain. Fulcrius does not judge the answer, and it does not forecast returns. It identifies where a holding conflicts with a stated preference, then shows alternative ways to express the same investment belief and how their dependencies and concentration differ.

Sources

China's share in all key solar manufacturing stages exceeds 80%; for polysilicon and wafers it was set to rise above 95%. International Energy Agency, Solar PV Global Supply Chains
First Solar says its CdTe thin-film modules do not rely on Chinese crystalline-silicon supply chains. First Solar SEC filing, 2026
Xinjiang accounted for about 35% of global solar-grade polysilicon supply in 2022, down from about 45% in 2020; forced-labor concerns remain documented. Foreign Policy, citing Sheffield Hallam and Bernreuter Research
TAN is Invesco's Solar ETF, based on an index of companies in the solar energy industry. Invesco TAN fund materials
DOE reports Xinjiang hosts a large share of Chinese polysilicon manufacturing and discusses forced-labor uncertainty in the solar supply chain. U.S. Department of Energy Solar Supply Chain Report

You asked

Which assumptions support the largest portion of my portfolio, and what would break them?

Affected dependency network

Research findings

You hold Nvidia, Equinix, and a private-credit fund exposed to AI infrastructure borrowers. They were bought at different times for different reasons: growth, income, real estate diversification. A brokerage screen places them in different categories.

Underneath, they share the same assumption: AI compute demand keeps growing enough to justify the buildout. Nvidia needs it for data-center chip orders. Equinix is expanding into AI-driven demand. The credit exposure needs its AI-linked borrowers to keep raising revenue or capital.

Together, 42% of this hypothetical portfolio rides on one assumption that no single holding states plainly. If model-efficiency gains or hyperscaler leasing pullbacks reduce expected compute demand, the positions can become vulnerable together. Sector allocation would miss that.

How scores move

Nvidia Risk increases Revenue concentration in data-center GPUs makes it first to feel a capex pullback.
Equinix Fragility increases Expansion plans become more sensitive to slower AI leasing demand.
AI credit exposure Risk increases If borrowers lose their growth story, floating-rate income does not help enough.

Portfolio relevance

What appears to be three separate trades may be one underlying exposure with three entry points. Fulcrius ranks the assumptions the portfolio relies on most heavily, then watches for the first evidence that they are weakening.

Fulcrius is still in development. These examples are illustrative, not live product output, and real answers may differ. Portfolio holdings are hypothetical. Real companies, supply-chain facts, and market figures are sourced where they appear.

Why now

The information already exists.

It is just scattered across filings, disclosures, news, trade data, and supplier records.

Keeping that picture current used to require large analyst teams. AI makes a different research workflow possible: faster reading, linking, verification, and ongoing refresh of the map.

Built for

Individual investors who do their own equity research and want the relationships beneath a thesis surfaced automatically.
Portfolio managers who need dependency and assumption tracking across holdings without adding headcount.
Market analysts who want to automate the repetitive parts of supply-chain and scenario research.

We're building Fulcrius because we believe investment reasoning should be explicit, inspectable, and adaptable.

From the Fulcrius team

Stop monitoring prices. Start monitoring assumptions.

Fulcrius maps the connections beneath your investments and keeps the reasoning tied to the evidence.

Fulcrius is built for research, not execution. It helps you understand what your investments depend on. You decide what to do next.

Join the waitlist

Leave your email. We will reach out when there is something to show you.

Be the first to know when Fulcrius is ready, plus a say in which sectors we map first. We email only when there is something worth sending.