PhysicsX

June 15, 2026

Public disclaimer: This is an independent public teardown based solely on publicly available information retrieved on 2026-06-15. It is research support, not investment, legal, tax, or financial advice. Company-provided statements are labeled as claims, not verified facts. PhysicsX did not participate in or review this report. Errors and omissions are possible; final decisions remain with the reader.

Temasek just valued PhysicsX near $2.4 billion. At that price a seed-sized check buys about one-tenth of one percent — and that math is the whole problem. The round itself is not in doubt: Bloomberg, Yahoo Finance, and Sifted all report the roughly $300 million Series C, with NVIDIA and Siemens among the backers. The product is the kind that makes investors lean in — Deep Physics Models that compress industrial simulation from months to seconds, built by founders with rare Formula 1 and frontier-simulation pedigree. We read 66 public sources across five partner lenses in a day, every claim cited, no founder spin. What the headline can't tell you is whether a genuinely impressive company is also a workable deal here, and whether the way it earns its money holds up under the price. The teardown separates the company from the deal. Read it before you let a $2.4B round answer for both.

What The Company Does

PhysicsX describes itself as an AI-native engineering and simulation software company for industrial engineering teams, with a platform spanning simulation workbench, artificial-intelligence workbench, Deep Physics Models, engineering applications, a model catalog, data unification, and platform services (PhysicsX homepage; PhysicsX platform). The named buyer is the engineering, research and development, simulation, and industrial program owner in semiconductors, aerospace and defense, materials, energy, renewables, and automotive markets (PhysicsX industries). Those product, buyer, and platform-module descriptions remain company claim evidence because they come from PhysicsX-controlled pages.

The public financing record is much stronger than the public commercial record. Bloomberg, Yahoo Finance, and Sifted independently support a roughly $300 million Series C at about $2.4 billion, led by Temasek with NVIDIA, Siemens, Applied Materials, Atomico, and General Catalyst among continuing or named backers (Bloomberg; Yahoo Finance; Sifted). The same public record does not independently confirm named paid production customers, recurring software revenue, annual contract value, gross margin, or renewal; partnership announcements with Siemens, NVIDIA, and Microsoft are not customer proof (Siemens; PhysicsX NVIDIA announcement; Microsoft UK Stories).

For this fund, the company and the deal separate cleanly. PhysicsX may merit a first call as a high-quality deep-tech company with rare founder-market fit; the deal does not clear a seed-stage portfolio-fit bar at roughly $2.4 billion unless a meaningfully sized, lower-basis, pro-rata-bearing entry exists.

Key Takeaways

Investor reaction
PhysicsX is genuinely impressive: the team, category, and financing are strong enough to avoid a company-quality pass, but the current entry looks structurally wrong for a seed fund.
Current verdict
pass for this fund, with tracking warranted and medium confidence.
Why investors might lean in
The Series C and valuation are independently corroborated, Corbo and Tuluie have unusually relevant public track records, and the company is attacking a difficult industrial engineering category where legacy simulation tools are being disrupted.
Why investors might pull back
At roughly $2.4 billion, a seed-sized check buys about 0.10 percent before further dilution, while revenue mix, customer proof, and NVIDIA or Siemens value-capture terms remain unavailable from public sources.
Highest-leverage fix
Disclose audited or management revenue mix, including recurring software annual recurring revenue, gross margin, and services split, and put one customer-controlled production reference with renewal or expansion evidence in front of investors.
Best next move
Most seed funds should log and track rather than open a full diligence process; if optionality on the company matters, spend one founder call on revenue quality, one production reference, and whether any pro-rata-bearing lower-basis allocation exists.

What Makes This Potentially Fundable

The fundable version of PhysicsX is not merely faster computer-aided engineering. It is a new physics-AI layer for high-complexity industrial research and development: the company claims Deep Physics Models and future Large Physics Models can compress simulations, unify engineering data, and become reusable model assets across difficult domains (PhysicsX platform; PhysicsX Fourier neural operator blog). If that model catalog compounds across programs and turns frontier engineering work into high-margin recurring software, the ceiling moves above a narrow simulation-software market.

The reason investors care is that two founder profiles fit the problem unusually well. Harvard SEAS describes Jacomo Corbo's PhD, Formula 1 race-strategy role, and QuantumBlack history (Harvard SEAS). Motor Sport Magazine profiles Robin Tuluie's Formula 1, Mercedes, Ducati, and simulation-science background (Motor Sport Magazine). That is a rare combination of physics simulation, performance engineering, and enterprise artificial intelligence productization.

The hard truth is that the public evidence does not yet prove durable value capture. It proves the company raised a major round, claims an ambitious platform, and has credible partnerships. It does not prove paid production customers, software-like margins, off-NVIDIA portability, or a fund-relevant ownership path. That is why the report lands at pass for this fund rather than pursue, while still recommending that the company be tracked.

The Four Holes To Close Before Fundraising

Hole	Investor Fear	What To Bring
Revenue quality is not public.	A company-claimed roughly $50 million 2026 revenue base could be high-margin recurring software, or it could be forward-deployed services priced as software.	Audited or management accounts showing annual recurring revenue, annual contract value, recurring share, gross margin, services share, and cohort retention.
Customer proof is not customer-controlled.	Named industrials and partnerships may be pilots, investor-overlap logos, or ecosystem programs rather than paid production accounts.	One customer-controlled production reference with deployment status, metrics, order form context, renewal, and expansion.
NVIDIA and Siemens are both validation and capture risk.	The most powerful partners also ship overlapping or adjacent stacks, so PhysicsX may create value that partners can capture or absorb.	Counterparty-validated summaries of intellectual-property, data, distribution, exclusivity, revenue-share, roadmap-control, and portability terms.
The deal may be structurally uninvestable for seed funds.	A small check at roughly $2.4 billion buys a token stake and may dilute passively through capital-intensive follow-on rounds.	Exact allocation, effective basis, pro-rata rights, information rights, founder vesting, and cap-table summary.

Decision Snapshot

One-sentence company description: PhysicsX is a growth-stage physics-AI company building AI-native engineering and simulation software for industrial research and development teams.
Screen: pass for this fund, with tracking warranted.
Confidence: medium.
Suggested next action: Do not open a full diligence process at this entry unless a founder call can quickly produce revenue-mix proof, one customer-controlled production reference, and a fund-fit allocation path.
Why this matters now: The company has just raised at a price where seed-fund ownership math becomes the binding question, not merely the company's technical quality.
Investor-readiness diagnosis: PhysicsX has strong company-quality signals but a thin public proof packet for revenue quality, production customers, partner economics, and fund-fit ownership.
Best founder use of this report: Separate the high-quality company story from the deal-structure problem, then prepare the narrow proof set that can reopen the verdict.

IC Disagreement Map

Five partners reviewed the same body of public evidence independently, each through a distinct lens: asymmetric upside, category timing and market structure, founder quality and execution, critical flaws and unsupported claims, and ownership, incentives, and long-horizon compounding. The final vote split across pass, hold, and pursue, which is why the verdict preserves dissent rather than averaging it into a vague middle position.

Power-law partner
passhigh confidence
The entry price and ownership math fail the fund-returning test. A seed-sized position at roughly $2.4 billion cannot matter enough unless PhysicsX proves a much larger platform ceiling and a real ownership-building entry.
Would flip on
High-margin recurring software revenue, off-NVIDIA portability, multi-sector production expansion, and a workable ownership position would move them toward pursue; services-heavy revenue or only token ownership would keep them at pass.
Prepared-mind partner
holdmedium confidence
The category shift is real, but market structure is unsettled because simulation incumbents and NVIDIA are moving into overlapping physics-AI capabilities.
Would flip on
Demonstrated cross-program model reuse, software-like margins, and production wins against incumbent artificial-intelligence simulation tools would move them to pursue; per-customer resets or hard NVIDIA dependence would move them to pass.
Founder-jockey partner
pursuemedium confidence
The Corbo and Tuluie founder-market-fit evidence is unusually strong and justifies a first call even though commercial proof is still missing.
Would flip on
Verified paid production customers, a deep engineering bench, and clarified Haag evidence would raise confidence; key-person concentration or pilot-only traction would move them toward pass.
Risk-reduction partner
holdmedium confidence
No fatal flaw is confirmed, but every value-capture-deciding claim remains unverified: recurring software revenue, paid production customers, partner economics, off-NVIDIA portability, and export posture.
Would flip on
Counterparty-validated PhysicsX-favorable partnership terms, a paid production customer, and a high-margin recurring revenue base would move them to pursue; partner-subordinated economics or services-heavy margins would move them to pass.
Long-horizon partner
passhigh confidence
Even a great company is not a fund-fit entry if a seed-sized check buys roughly 0.10 percent and cannot be defended through large future rounds.
Would flip on
A meaningfully sized lower-basis allocation with pro-rata, aligned founder economics, and durable independent value capture would move them toward hold or pursue; a token no-pro-rata position confirms pass.

After follow-up evidence review, the power-law partner raised confidence from medium to high while keeping the pass vote; the other four partners confirmed their votes and confidence unchanged. The new UK accounts evidence strengthened the view that revenue quality remains unresolved, while partnership economics, customer status, and allocation terms stayed unavailable from public sources.

Scenario Range

Scenario	What The Company Looks Like In 3-5 Years	Falsifiable Trigger To Watch	Earliest Evidence
Strikeout	PhysicsX becomes a respected but capped, high-burn forward-deployed engineering business whose per-customer models do not compound into durable software economics, and it is absorbed by a strategic backer at or below the current valuation.	Revenue per head stays flat or falls as headcount scales, pilots do not convert to multi-year production at improving gross margin, and partner economics favor NVIDIA or Siemens.	Within 6 to 12 months, management accounts and references show services-heavy revenue, no production-retention proof, and no cross-program model reuse.
Base	PhysicsX becomes a well-funded category brand with real frontier accounts and a mix of recurring software and services, but the outcome is more likely strategic than independent public-company scale.	The company shows some production accounts and margin improvement, but the model catalog remains only partly reusable and incumbents keep the broad distribution advantage.	Within 12 to 18 months, the company can produce one or two references and partial recurring software metrics, but not a majority high-margin recurring base.
Home run	PhysicsX becomes a horizontal physics-AI platform for industrial research and development, with reusable model assets, high-margin recurring revenue, multi-sector production expansion, and independence from any single compute or simulation platform.	Deep Physics Model or model-catalog components are reused across multiple programs at improving accuracy or cost, recurring software gross margin becomes software-like, and the company demonstrates off-NVIDIA deployment.	Within 12 to 18 months, audited or management financials, customer-side references, and technical artifacts show cross-program reuse, net expansion, and partner terms that preserve PhysicsX's value capture.

What Investors Will Test

What We Looked For	Current Read	Investor Implication	Founder Prep Priority
A fund-returning path from the current price	The company may be large, but a seed-sized check at roughly $2.4 billion does not produce fund-relevant ownership unless a special entry exists.	The investor will separate company admiration from portfolio construction.	Bring allocation, effective basis, pro-rata, and information-rights terms before asking a seed fund to engage deeply.
Outside-the-room proof of traction	The round and valuation are corroborated; named paid production customers are not.	Investors will discount logos, partnerships, and pilots until a customer controls the proof.	Bring one referenceable production customer with renewal, expansion, and deployment metrics.
Durable value capture	The software-versus-services split, gross margin, and partner economics are not public.	Investors cannot underwrite a software multiple or moat from the public record alone.	Prepare segmented revenue, gross margin, recurring share, partner terms, and portability evidence.
Founder-market fit	Corbo and Tuluie are strongly evidenced; Haag and the broader engineering bench are thin publicly.	The team case supports a first call but does not override the price and ownership problem.	Provide Haag's prior systems record and the named senior engineering leadership map.
Regulated-market readiness	Aerospace and defense and semiconductor exposure is company-claimed, while export-control posture is not public.	Investors will treat compliance as a serviceable-market and sales-readiness question.	Bring export-control counsel memo, security attestations, and named compliance ownership.

Claim Reconciliation: Inconsistencies Investors Will Catch

Claim Or Metric	Where It Appears	Conflicting / Unreconciled Versions	Why Investors Flag It	How To Reconcile
Series B size	Tech.eu PhysicsX Greenhouse job board	Tech.eu reports $135 million, while the company-linked jobs page states $155 million. The UK accounts confirm June 2025 timing but not the amount (Companies House accounts).	A funding-size mismatch on a company-controlled surface is a cheap accuracy flag and affects dilution math.	Provide closing documents, cap table, share-allotment detail, and a reconciled financing-history schedule.
Revenue and growth	Yahoo Finance PhysicsX Series C announcement	The roughly $50 million 2026 figure, doubled recognized revenue, tripled bookings, and more-than-doubled customer count have no public base or audited reconciliation.	Investors cannot know whether the number is annual recurring revenue, total revenue, projected revenue, bookings, backlog, or services-inclusive revenue.	Provide recognized-revenue bridge, bookings-to-revenue reconciliation, annual recurring revenue, services split, gross margin, and definitions.
Headcount	PhysicsX Series C announcement Sifted LinkedIn Crunchbase Companies House accounts	Company says more than 300 employees; Sifted reports 150 to 350 growth; LinkedIn showed 282 employees and a 51 to 200 band; Crunchbase showed 101 to 250; FY2024 accounts show average monthly group headcount of 74.	Headcount is both a momentum signal and a services-risk signal, so stale or inconsistent figures can mislead revenue-per-head analysis.	Provide current full-time employee count, contractor count, role mix, and month-end history.
Customer status	Yahoo Finance General Catalyst Siemens	Applied Materials, Siemens, Stellantis, and Velo3D are named in press or investor narratives, but no customer-controlled paid production case study was found. Applied Materials and Siemens also appear as investors or partners.	Logo walls and partner names get discounted when paid status, production status, renewal, and customer control are not clear.	Build a customer-status sheet with paid status, production status, scope, renewal, and reference permission.
NVIDIA relationship	PhysicsX NVIDIA announcement PhysicsX Deutsche Telekom and NVIDIA announcement NVIDIA PhysicsNeMo GitHub	PhysicsX frames NVIDIA as partner and infrastructure; NVIDIA also maintains PhysicsNeMo, an overlapping physics-AI framework.	The same relationship can be distribution leverage or dependence on a competitor, depending on private terms.	Provide off-NVIDIA deployment evidence and counterparty-validated terms covering intellectual property, data, exclusivity, distribution, and roadmap control.
Security and compliance	PhysicsX ISO announcement PhysicsX Microsoft Discovery article	PhysicsX claims ISO 27001 and SOC 2-related posture, but public checks did not surface a certificate body, scope, trust center, subprocessor list, or SOC 2 Type II report.	Regulated industrial buyers will require proof, not marketing language.	Provide ISO certificate details, SOC 2 Type II report, data processing addendum, subprocessor list, retention policy, and artificial-intelligence governance materials.

Diligence Findings: Issues To Fix Before You Raise

2 deal-breaker3 high3 medium

The fund cannot underwrite a fund-returning position without exact allocation, basis, pro-rata, and follow-on rights.
deal-breaker
Financing and portfolio construction
Fix
Provide cap table, allocation terms, pro-rata rights, and information-rights path before substantive investor time.
Revenue quality is not publicly underwritable.
deal-breaker
Financial and commercial diligence
Fix
Provide audited or management revenue segmentation, recurring share, gross margin, and cohort retention.
No customer-controlled paid production proof is visible.
high
Customer diligence
Fix
Provide at least one customer-controlled production reference with attested metrics and renewal.
NVIDIA and Siemens economics decide whether the company keeps or cedes value.
high
Legal, commercial, and technical diligence
Fix
Provide counterparty-validated term summaries and off-NVIDIA portability evidence.
Model-catalog compounding is asserted but not proven.
high
Product and technical diligence
Fix
Show cross-program reuse at improving accuracy or cost, and reconcile it with customer data isolation.
Series B amount and cumulative capital are unreconciled.
medium
Financing diligence
Fix
Provide closing documents and a financing-history schedule.
Regulated-market readiness is not publicly evidenced.
medium
Legal, security, and procurement diligence
Fix
Provide export-control program materials, ISO certificate scope, SOC 2 Type II report, and security package.
Third-founder and senior bench evidence is thin.
medium
Team diligence
Fix
Provide Haag's prior work evidence and named senior technical leadership references.

Data Room Readiness Checklist

Procurement-readiness score: partial.
Rationale: PhysicsX has public privacy, ISO, and SOC 2-related claims, but public sources did not show a complete buyer trust package, export-control program, data processing addendum, subprocessor list, third-party certificate scope, SOC 2 Type II report, or artificial-intelligence governance materials.

Data Room Area	What Investors Expect	Current Read	Status	Priority
Company overview & corporate	One-pager, incorporation, good standing, structure	PHYSICSX LIMITED is confirmed at Companies House, with London jurisdiction and prior name history.	partial	high
Financials	Monthly P&L for 18 to 24 months, three-year model with assumptions, burn and runway	FY2024 group loss is public, but turnover, annual recurring revenue, margins, burn, cash, and runway are not.	partial	critical
Cap table & funding history	Clean cap table, prior rounds, SAFEs or notes, 409A	Series A, Series B, and Series C are publicly reported, but Series B amount and post-round control remain unreconciled.	partial	critical
Legal & IP	Bylaws, board consents, intellectual property assignments, material contracts	Public record confirms entity and officers; intellectual-property ownership, partner terms, and customer contracts are private.	missing	high
Product & technology	Roadmap, architecture overview, security and compliance docs	Platform claims are public, but model-catalog reuse, portability, and customer-side benchmarks are private.	partial	high
Team	Org chart, key employment and advisor agreements, vesting	Founders and hiring are public; Haag evidence, bench depth, and vesting are not public.	partial	high
Customers & traction	Retention cohorts, ARR/MRR bridge, pipeline, three to five references	No customer-controlled paid production proof is public.	missing	critical
Market & competition	TAM/SAM/SOM, competitive landscape, pricing	Market and competitors can be framed publicly; PhysicsX-specific serviceable market, pricing, and win/loss remain missing.	partial	medium

First-Call Agenda For The Startup

Time	Topic	Founder Goal	Evidence To Bring
0-10 minutes	Deal structure and ownership	Establish whether this fund can own enough to matter.	Allocation, effective basis, pro-rata rights, cap table, and information-rights path.
10-25 minutes	Revenue quality	Show whether the business is recurring software, services, or a mix.	Annual recurring revenue, annual contract value, gross margin, services split, bookings bridge, and recognized revenue.
25-35 minutes	Production customer proof	Convert the traction story from logos to evidence.	One customer-controlled production reference with deployment metrics, paid status, renewal, and expansion.
35-45 minutes	Partner value capture	Explain whether NVIDIA and Siemens aid or capture PhysicsX's value.	Term summaries, off-NVIDIA portability proof, and intellectual-property or data-rights summary.
45-52 minutes	Moat and model reuse	Prove what compounds in the platform.	Cross-program Deep Physics Model or model-catalog reuse artifact with accuracy or cost improvement.
52-60 minutes	Team and regulated readiness	Close the hardest-build and compliance gaps.	Haag evidence, engineering bench map, export-control program, security attestations, and named owners.

Business Model And Revenue Signals

Enterprise, engineering-led sales motion with no public pricing page found.
inferencePhysicsX-controlled pageshigh
SourcePhysicsX platform
Sales motion appears enterprise and technical rather than self-serve.
Forward-deployed engineers embed in customer programs.
PhysicsX-controlled pageshigh
SourcePhysicsX about PhysicsX platform
This can be a moat or a services-margin warning.
Pilots typically run one to three months.
PhysicsX-controlled blogmedium
SourcePhysicsX pilots blog
This helps explain the public customer-proof gap, but pilots are not paid production proof.
Roughly $50 million 2026 revenue is attributed to a CEO quote.
press relaying company statementlow
SourceYahoo Finance
The figure is not independently audited or tied to annual recurring revenue.
FY2024 accounts show group loss and average monthly headcount, but not turnover.
source-backed factUK public registry filinghigh
SourceCompanies House accounts
The accounts neither confirm nor refute the 2026 revenue claim.
Series C investor support is strong.
source-backed factreputable presshigh
SourceBloomberg Yahoo Finance Sifted
Backing supports capital access, not software economics.

Public revenue quality is not available. The roughly $50 million 2026 figure remains a company claim; any revenue scenario in this report is an estimate or diligence question, not a fact.

Revenue Quality Checklist

Before PhysicsX can move from pass for this fund to pursue, the company should be ready to show annual recurring revenue, recognized revenue, bookings, annual contract value by customer, software gross margin, forward-deployed-services gross margin, deployment cost, customer acquisition cost, lifetime value, payback, gross revenue retention, net revenue retention, churn, concentration, and a source-to-model bridge tying contracts, invoices, board metrics, and the financial model together. The company should also define whether the roughly $50 million 2026 figure is contracted, recognized, projected, annualized, or services-inclusive.

Funding And Ownership Context

Item	Public Read	Evidence Label	Diligence Request
Total raised / rounds	Series A was reported at $32 million, Series B at $135 million by Tech.eu, and Series C at roughly $300 million; cumulative equity raised is at least roughly $467 million if those round figures are used.	`source-backed fact` for the reported rounds; `inference` for cumulative sum	Provide financing-history schedule and reconcile the Series B amount.
Instruments (equity / SAFE / notes)	Public sources do not disclose instruments or preference stack.	`unknown`	Provide financing documents, preference stack, and side-letter summary.
Investors	Temasek, NVIDIA, Siemens, Applied Materials, Atomico, General Catalyst, M&G, and Intrepid appear in public sources.	`source-backed fact` for named investor presence in press and database records	Provide full cap table, board composition, and strategic rights.
Post-money valuation	Series C valuation is reported around $2.4 billion; prior valuation is reported around $1 billion.	`source-backed fact`	Provide exact post-money, share price, option pool, and liquidation preferences.
Cap table / ownership	UK persons-with-significant-control filings show no registrable person after prior founder entries ceased, but private cap table is not public.	`source-backed fact` for public registry status; `unknown` for ownership	Provide post-Series C cap table, founder ownership, founder vesting, and governance rights.
Burn & runway	FY2024 accounts show loss; current burn, cash, and runway are not public.	`source-backed fact` for FY2024 loss; `unknown` for current runway	Provide monthly burn, cash balance, hiring plan, and runway sensitivity.
Next-round plan	Not public.	`unknown`	Provide milestones needed before the next financing and whether future rounds will require large follow-on capital.

SEC EDGAR was checked and showed no matching Form D filing for PhysicsX. This carries little weight for a UK-incorporated company that may raise through non-US structures.

Founder And Team

Jacomo Corbo
Chief executive officer and co-founder; independently associated with Harvard SEAS PhD work, Formula 1 strategy, and QuantumBlack.
PhysicsX about LinkedIn Harvard SEAS
Evidencesource-backed fact for identity and independent background
Confidencehigh
Robin Tuluie PhD
Co-founder and chairman; independently profiled for Formula 1, Mercedes, Ducati, and simulation science.
PhysicsX about LinkedIn Motor Sport Magazine
Evidencesource-backed fact for identity and independent background
Confidencehigh
Nicolas Haag
Co-founder and director of simulation engineering; public footprint is thinner than the other two founders.
PhysicsX about LinkedIn
Evidencesource-backed fact for identity; unknown for pre-PhysicsX depth
Confidencemedium for identity; low for track record
Company team scale
LinkedIn showed 282 employees; Crunchbase showed a 101 to 250 band; Sifted reported growth from 150 to 350; FY2024 UK accounts showed average monthly group headcount of 74.
LinkedIn Crunchbase Sifted Companies House accounts
Evidencesource-backed fact for public proxies and filing figures; company claim for 300-plus
Confidencemedium
Open roles
Thirty-six open roles were visible across London, New York, San Francisco, and Singapore, with forward-deployed, machine-learning, simulation, research, and delivery emphasis.
PhysicsX Greenhouse job board
Evidencesource-backed fact
Confidencehigh

Founder Competency Coverage

Domain depth
Jacomo Corbo
evidenced
Robin Tuluie
evidenced
Nicolas Haag
claimed
What The Evidence Is
Corbo and Tuluie have independent Formula 1 and simulation evidence; Haag's domain role is visible on company and LinkedIn surfaces only.
Technical build capability
Jacomo Corbo
evidenced
Robin Tuluie
evidenced
Nicolas Haag
claimed
What The Evidence Is
Corbo and Tuluie's backgrounds support technical depth; Haag's simulation-engineering ownership is plausible but not independently evidenced.
Product
Jacomo Corbo
claimed
Robin Tuluie
claimed
Nicolas Haag
claimed
What The Evidence Is
The product maps to the founders' domain, but public evidence does not show product leadership artifacts or customer adoption details.
GTM / sales
Jacomo Corbo
claimed
Robin Tuluie
unknown
Nicolas Haag
unknown
What The Evidence Is
Corbo's QuantumBlack history supports enterprise artificial-intelligence exposure, but PhysicsX sales repeatability is not public.
Leadership / hiring
Jacomo Corbo
evidenced
Robin Tuluie
claimed
Nicolas Haag
unknown
What The Evidence Is
Funding and hiring velocity support leadership for the company overall; individual leadership scope below the founders is not public.
Fundraising history
Jacomo Corbo
evidenced
Robin Tuluie
evidenced
Nicolas Haag
evidenced
What The Evidence Is
The company raised Series A, Series B, and Series C rounds supported by public sources.
Prior founding outcomes
Jacomo Corbo
evidenced
Robin Tuluie
unknown
Nicolas Haag
unknown
What The Evidence Is
Corbo's QuantumBlack history is independently profiled; prior founding outcomes for Tuluie and Haag were not found.

Building reusable physics-AI models that compound across programs without violating customer data isolation
high
Team coverage today
Corbo and Tuluie provide strong senior domain coverage; Haag and the bench need stronger public or referenceable evidence.
What would close it
Cross-program reuse evidence, named technical owners, and references from prior shipped systems.
Selling into aerospace, semiconductor, energy, and other regulated industrial programs
high
Team coverage today
Founder expertise is strong, but commercial repeatability and compliance leadership are not public.
What would close it
Production customer references, export-control program, security package, and regulated-sales ownership.
Turning forward-deployed work into software-like margins
high
Team coverage today
The delivery model is visible, but margin and recurring share are not.
What would close it
Segmented revenue, gross margin, and delivery-efficiency trend by customer cohort.

Public Professional Footprint

These checks of public technical artifacts, professional social presence, and education or credential traces carry low decision weight by design; they corroborate or weaken the competency table above and sharpen founder-call questions, but none alone changes the verdict.

Technical artifacts (repo level)
Jacomo Corbo
unknown
Robin Tuluie
unknown
What The Public Record Shows
No decision-relevant public repository artifacts were identified for either founder in the evidence reviewed.
Professional social content
Jacomo Corbo
claimed
Robin Tuluie
claimed
What The Public Record Shows
LinkedIn profiles support current association with PhysicsX; the public professional history is stronger for Corbo through Harvard and for Tuluie through Motor Sport Magazine than through social posts.
Education / credentials
Jacomo Corbo
evidenced
Robin Tuluie
claimed
What The Public Record Shows
Corbo's Harvard PhD is independently supported; Tuluie's astrophysicist description and technical background are press-supported, while full credential records were not independently verified.
Publications / patents / certifications
Jacomo Corbo
unknown
Robin Tuluie
unknown
What The Public Record Shows
No decision-relevant public patents, publications, or certifications were surfaced for the founders in this research.

Net read: public founder evidence is strong enough to justify one targeted conversation on team quality, but the verdict still turns on ownership, revenue quality, customer proof, and value capture.

Founder-Market Fit Read

What is promising: Corbo and Tuluie fit the physics-simulation and enterprise artificial-intelligence problem unusually well, and the company has recruited and financed aggressively.
What is missing: Haag's prior track record, senior engineering bench evidence, production customer conversion, export-control leadership, and founder ownership after large growth rounds.
What to prepare: Dated biographies, references, Haag evidence, engineering org chart, senior bench credentials, founder vesting, and customer references tied to commercial execution.

PHYSICSX LIMITED
UK operating entity, company number 12134466, incorporated 1 August 2019; prior name MOTODYNAMICS LTD.
Companies House
Evidencesource-backed fact
Confidencehigh
Follow-upProvide full corporate structure and subsidiaries if any.
PHYSICSX LIMITED officers
Corbo and Tuluie are active UK directors in the public registry.
Companies House officers
Evidencesource-backed fact
Confidencehigh
Follow-upProvide board composition after Series C.
Persons with significant control filing
Public registry shows no registrable person after prior founder entries ceased.
Companies House PSC
Evidencesource-backed fact
Confidencehigh
Follow-upProvide post-round cap table, founder ownership, and governance summary.
QuantumBlack
Corbo is independently profiled as a co-founder of QuantumBlack, which McKinsey acquired in 2015.
Harvard SEAS
Evidencesource-backed fact
Confidencehigh
Follow-upVerify role scope and references if the founder story is load-bearing.

Traction

Customer Status Table

Siemens
Siemens confirms collaboration around LGM-Aero and PhysicsX Ai.rplane; PhysicsX also has company-controlled Siemens collaboration pages.
yes for collaboration existenceMarketplaceUnknownUnknownUnknownhigh for collaboration existence; low for paid customer status
NVIDIA
PhysicsX announces NVIDIA alignment and native NVIDIA infrastructure; NVIDIA maintains PhysicsNeMo, an overlapping physics-AI framework (PhysicsX NVIDIA announcement; NVIDIA PhysicsNeMo).
no customer proofMarketplaceUnknownUnknownUnknownhigh for platform relationship; low for customer status
Microsoft
Microsoft UK describes PhysicsX in the context of transforming engineering with physics AI on Microsoft Discovery and Azure-related narratives.
yes for Microsoft-authored featureMarketplaceUnknownUnknownUnknownmedium for collaboration narrative; low for paid customer status
Applied Materials
Named in Yahoo Finance as a customer and appears as an investor in the funding context.
No proofClaim onlyUnknownUnknownUnknownlow
Stellantis
Named in Yahoo Finance as a customer.
No proofClaim onlyUnknownUnknownUnknownlow
Velo3D
General Catalyst describes a Velo3D deployment reference.
no customer-controlled proofClaim onlyUnknownUnknownUnknownlow

Google Search was checked for customer-controlled paid production case studies and showed no matching listing. Partnership and press mentions cannot substitute for paid-production proof.

Traction Signals

Series C and valuation
source-backed facthigh
SourceBloomberg Yahoo Finance Sifted
Exact terms, preference stack, and ownership.
Prior Series A and Series B
source-backed facthigh
SourceTechCrunch Tech.eu
Series B size reconciliation and cap table.
Company-claimed revenue and growth multiples
low
SourceYahoo Finance PhysicsX Series C announcement
Audited or management accounts, definitions, and margin split.
FY2024 group accounts
source-backed facthigh
SourceCompanies House accounts
Turnover and software-versus-services mix remain missing.
Hiring velocity
source-backed factmedium
SourcePhysicsX Greenhouse job board Sifted Companies House accounts
Confirm current full-time employee count, contractors, and role mix.
Category visibility
inferencelow
SourceGoogle Search (United States, United Kingdom, and Germany)
Search visibility should be tied to pipeline and win rates before it is treated as demand; see Category Visibility Snapshot.

Hiring And Org Momentum

Company careers page
PhysicsX careers routed to a Greenhouse board and describes an organization with customer-facing engineering delivery.
SourcePhysicsX careers
Hiring is active and aligned with a technical enterprise model.
Greenhouse jobs board
source-backed factThirty-six roles were visible across London, New York, San Francisco, and Singapore, including forward-deployed, machine-learning, simulation, research, and delivery roles.
SourcePhysicsX Greenhouse job board
Role mix supports build ambition but also raises the question of services delivery versus software leverage.
LinkedIn Jobs
source-backed factA senior data scientist posting was found on LinkedIn Jobs.
SourceLinkedIn Jobs
Confirms public technical hiring beyond the company board.
UK accounts headcount
source-backed factFY2024 average monthly group headcount was 74, up from 28 in FY2023.
SourceCompanies House accounts
Confirms growth through FY2024 but lags the 2026 company and press headcount claims.

Indeed and Glassdoor were checked and showed no decision-relevant PhysicsX employer listing or employee-sentiment signal. These absences carry little weight because Greenhouse is the primary linked board and LinkedIn Jobs confirmed active technical hiring.

Role-mix read: The role mix is strongly technical and delivery-heavy, which can support a differentiated forward-deployed model but also makes revenue-per-head and margin trend critical.
Momentum read (growth / steady / contraction / unknown): growth, with medium confidence from public hiring, Sifted's headcount report, and UK accounts showing earlier headcount growth.

Traction Quality Read

Traction Dimension	Current Status	Good Enough For First Call?	Needed For Deep Diligence
Financing	Series C, Series B, and Series A are publicly supported.	yes	Exact terms, preference stack, allocation, cap table, and pro-rata rights.
Revenue	The roughly $50 million 2026 figure remains a `company claim`.	yes for a call; no for underwriting	Recognized revenue, annual recurring revenue, services split, gross margin, and retention.
Customer existence	Partnerships and press-named accounts exist; paid production customers are not independently confirmed.	yes for a call; no for underwriting	Customer-controlled references and deployment evidence.
Retention	No public retention, renewal, or expansion data was found.	no	Gross and net revenue retention, cohort tables, and customer references.
Deployment depth	Pilot-led go-to-market is company-described; production depth is unknown.	partially	Deployment scope, systems touched, timeline, margin, and renewal.
Repeatability	The model-catalog compounding thesis is not proven publicly.	no	Cross-program reuse evidence and improving deployment efficiency.

Competitive Landscape

Segment	Examples	Customer Alternative	Pressure On Company
Compute and framework platform	NVIDIA PhysicsNeMo	Build or train physics-AI models on NVIDIA's open framework or partner ecosystem.	PhysicsX must prove it adds proprietary delivered solutions and captures value beyond the framework.
Incumbent simulation and CAE platforms	Siemens Simcenter PhysicsAI, Ansys AI, Synopsys, Dassault SIMULIA, COMSOL, Altair PhysicsAI	Buy artificial-intelligence simulation capabilities from existing simulation or industrial software vendors.	Incumbents control budgets, workflows, and procurement relationships.
AI-native simulation challengers	Neural Concept, Luminary Cloud, Navier AI, Pasteur Labs	Choose another venture-backed AI simulation or computational-fluid-dynamics specialist.	PhysicsX must prove category leadership beyond funding and search visibility.
Cloud simulation and infrastructure substitutes	SimScale, Rescale	Use cloud-native simulation or high-performance-computing infrastructure with artificial-intelligence features.	PhysicsX must show differentiated model outcomes rather than infrastructure access alone.
Internal build or services	In-house machine-learning, simulation, and forward-deployed engineering teams	Build bespoke physics-AI workflows internally using open frameworks and incumbent tools.	PhysicsX must prove faster time to value, better accuracy, and lower total cost of ownership.

Category Visibility Snapshot

Google Search / United Statesphysics AI simulation software

Captured page-one results (2026-06-15): PhysicsX appeared prominently along with NVIDIA PhysicsNeMo, Rescale, Neural Concept, Luminary Cloud, Siemens, and other physics-AI or simulation pages.
Was the company present?: Yes

Google Search / United Kingdomphysics AI simulation software

Captured page-one results (2026-06-15): PhysicsX appeared prominently, with similar physics-AI and simulation competitors visible.
Was the company present?: Yes

Google Search / Germanyphysics AI simulation software

Captured page-one results (2026-06-15): PhysicsX appeared on the first page but not in the top position, with Siemens and other incumbent or specialist pages also visible.
Was the company present?: Yes

Google Search in the United States was checked for "neural operator engineering simulation" and "AI for computational fluid dynamics" queries and did not surface PhysicsX on page one. These absences carry little weight because adjacent technical queries emphasize research sources rather than commercial vendors.

Visibility read (inference, low confidence): PhysicsX has strong visibility on the exact "physics AI simulation software" phrase, but that may reflect category creation rather than durable demand capture; absence from adjacent technical queries limits the weight of this signal.

Pricing And Competitive Benchmark

Alternative	What It Offers	Public Price Signal	Price Vs. This Company	Evidence Label
PhysicsX	AI-native engineering and simulation platform with forward-deployed delivery.	No public pricing found.	Baseline unknown.	`unknown`
SimScale	Cloud-native simulation software.	Public pricing page indicates quote-based and plan-tier packaging.	SimScale provides a packaging proxy, but not a direct enterprise price comparison.	`company claim` via SimScale pricing
SIMUL8 Professional	Simulation software listed on a software marketplace.	Capterra lists per-user pricing signals.	It is a lower-end pricing proxy, not a direct competitor to PhysicsX's enterprise frontier programs.	`source-backed fact` via Capterra
Ansys	Enterprise simulation suite.	Vendr provides transaction-data pricing proxy, while Ansys enterprise pricing is typically quote-based.	The proxy suggests high annual contract values are possible in simulation software.	`inference` via Vendr
Siemens Simcenter PhysicsAI	Incumbent AI-assisted simulation feature inside Siemens ecosystem.	Public dollar price not found.	Cannot compare without PhysicsX annual contract value and Siemens package pricing.	`company claim` via Siemens Simcenter PhysicsAI
NVIDIA PhysicsNeMo	Open physics-AI framework and ecosystem.	Open-source or developer framework signals, not enterprise solution pricing.	It can lower build costs for customers or competitors, but it is not directly priced against PhysicsX.	`source-backed fact` via GitHub and NVIDIA PhysicsNeMo

Price positioning read: PhysicsX cannot be benchmarked publicly because its annual contract value, packaging, and services split are private.
Price claims to correct or substantiate: PhysicsX should avoid any public claim of cost superiority or software-like economics until it can tie pricing to customer total cost of ownership, deployment margin, and retained outcomes.

Competitive Wedge

PhysicsX's defensible wedge would be a forward-deployed physics-AI model catalog that compounds across difficult industrial programs faster than incumbents can reproduce through self-serve tools. The copyable version is a bespoke services layer over NVIDIA and incumbent simulation infrastructure. Current evidence supports the question, not the answer: the company claims the model-catalog platform, while the public record does not prove cross-program reuse, software-like margin, or partner-independent distribution.

Risks And Open Questions

Risk	Severity	Evidence	What To Ask
Fund-fit ownership mismatch	deal-breaker	Public sources show a roughly $2.4 billion valuation, but no allocation or pro-rata path.	What position size, basis, and rights can this fund actually secure?
Revenue quality unknown	deal-breaker	The roughly $50 million 2026 figure is a `company claim`, while FY2024 accounts disclose loss and not turnover.	What share of revenue is recurring software, and what are gross margins?
No customer-controlled production proof	high	Public searches did not find paid production case studies, and named accounts are press or investor narratives.	Which customers are paid and in production, and can they speak?
NVIDIA and Siemens capture risk	high	NVIDIA and Siemens are partners or backers with overlapping or incumbent assets.	Do partner terms preserve PhysicsX's intellectual property, data rights, distribution, and roadmap control?
Model-catalog compounding unproven	high	PhysicsX claims data isolation and model-catalog ambition, but no cross-program reuse evidence is public.	Has any model component been reused across programs with improving accuracy or cost?
Export-control exposure	medium	Aerospace and semiconductor exposure is company-claimed, while export-control posture is not public.	What is the ITAR, Export Administration Regulations, UK dual-use, and European Union dual-use program?
Series B and cap-table opacity	medium	Public sources differ on Series B size and persons-with-significant-control filings do not reveal current control.	Can the company provide financing documents, cap table, and founder vesting?

Pre-Mortem: The Most Likely Obituary

Cause of death (one sentence): PhysicsX became a respected, well-funded forward-deployed engineering-services business whose per-customer models did not compound into a defensible software platform, and it was absorbed by a strategic backer below the independent outcome needed for a seed-fund return.
The causal chain from today to the shutdown:
1. Revenue remains majority services or bespoke delivery, with gross margin below software benchmarks.
2. Customer data isolation prevents a cross-customer data network effect, and model-catalog reuse proves weaker than expected.
3. NVIDIA PhysicsNeMo and incumbent tools from Siemens, Ansys, Synopsys, and Dassault become good enough for many buyers.
4. Growth requires large additional rounds while a small seed-fund stake dilutes passively.
5. A strategic partner or incumbent buys the company for talent, customers, and roadmap control rather than for an independent public-company outcome.
The earliest observable warning sign: Revenue per head stays flat or falls as headcount scales, and pilots fail to convert into multi-year production accounts at improving gross margin.
The question that defuses this chain today: "Can you show recurring-software gross margin, one paid production reference, and a model-catalog component reused across programs at improving accuracy or cost?"

Decision-Critical Unknowns

Unknown	Why It Is Decision-Critical	Best Evidence	Decision Effect
Available ownership and pro-rata rights	The current verdict is a structural pass unless the fund can own enough to matter.	Allocation letter, cap table, pro-rata terms, information rights, and effective basis.	A meaningful lower-basis allocation with pro-rata would move the decision toward hold or pursue; a token stake with no rights confirms pass.
Revenue mix and gross margin	Determines whether the company is software-like or services-like.	Segmented revenue, annual recurring revenue, annual contract value, services share, gross margin, and retention.	Majority high-margin recurring software would reopen headroom; majority low-margin services would move to a stronger pass.
Paid production customers	Separates real traction from logos, pilots, and partnerships.	Customer-controlled reference, deployment date, scope, order-form context, renewal, and expansion.	A strong production reference would justify a first call; no production proof would reinforce pass.
NVIDIA and Siemens economics	Determines whether partners aid or capture value.	Counterparty-validated terms and off-NVIDIA deployment evidence.	Favorable terms and portability would increase conviction; partner control or lock-in would lower it.
Cross-program model reuse	Determines whether the model catalog compounds.	Technical artifact showing reused components across programs with better accuracy or cost.	Reuse would support the platform thesis; per-customer reset would weaken it materially.
Haag and bench depth	Tests hardest-build coverage and key-person risk.	Prior systems evidence, references, and senior engineering leadership map.	Strong evidence would support the founder-jockey case; thin bench would cap it.

Diligence Questions

First Call

Question	Why It Matters	Good Evidence
What position size, basis, and pro-rata rights are available to a sub-$2.5 million check?	It determines whether this fund has an investable deal at all.	Cap table, allocation terms, pro-rata rights, and information-rights path.
What is recurring-software annual recurring revenue, its share of total revenue, and gross margin versus forward-deployed services?	It determines value capture and terminal-value math.	Segmented financials, annual contract value, services split, gross margin, and retention.
Which named accounts are paid production customers and what is net retention?	It converts the traction story from claims to proof.	Customer-controlled references, deployment metrics, renewals, and order-form context.
Do NVIDIA and Siemens terms preserve PhysicsX's intellectual property, data rights, and distribution, and can the platform run off NVIDIA?	It tests whether the strongest partners are also value capturers.	Counterparty-validated terms and non-NVIDIA deployment evidence.

Follow-Up

Question	Why It Matters	Good Evidence
Has any Deep Physics Model or model-catalog component been reused across two or more programs with improving accuracy or cost?	It tests the compounding moat.	Cross-program reuse artifact with baseline and improvement metrics.
Reconcile the Series B amount and provide the post-Series C cap table, shareholder agreement summary, and founder vesting.	It resolves accuracy, dilution, and incentive-alignment concerns.	Closing documents, cap table, shareholder-agreement summary, and founder economics.
What did Nicolas Haag build before PhysicsX, and who leads the senior engineering bench?	It tests hardest-build coverage.	Prior systems evidence, references, and org chart.
What export-control, ISO, SOC 2, and regulated-buyer trust materials are ready?	It tests sales readiness in aerospace, defense, semiconductor, and high-classification settings.	Counsel memo, ISO certificate, SOC 2 Type II, data processing addendum, and subprocessor list.

Kill Criteria

Kill Criterion	Evidence That Would Trigger It
Only a token sub-0.2 percent no-pro-rata stake is available at the full valuation.	Cap table and allocation terms show no defendable ownership path.
Revenue is majority low-margin forward-deployed services.	Segmented financials show services dominate revenue and margin does not improve.
No customer-controlled paid production reference is available.	Customer list resolves to pilots, partnerships, investor-overlap logos, or unreferenceable accounts.
NVIDIA or Siemens terms subordinate PhysicsX's roadmap, intellectual property, data, or distribution.	Contract summaries show partner control, exclusivity, hard lock-in, or unfavorable revenue share.
Model-catalog components reset per customer and do not reuse across programs.	Technical diligence shows no reusable model assets or improving deployment efficiency.
Series B or financial metric discrepancies reflect a material misstatement rather than a harmless presentation difference.	Closing documents and board materials contradict public claims without a credible explanation.

Double-Down Criteria

Double-Down Criterion	Evidence That Would Justify More Diligence
A meaningful lower-basis allocation with pro-rata is available.	Allocation terms and cap-table evidence fit the fund's return math.
Majority high-margin recurring software revenue is verified.	Audited or management accounts show recurring software annual recurring revenue and software-like gross margin.
At least one paid production customer validates retained value.	Customer-controlled reference shows deployment metrics, renewal, and expansion.
Partner terms are favorable and portability is real.	Counterparty-validated summaries and a non-NVIDIA deployment artifact show independent value capture.
Model-catalog compounding is evidenced.	Cross-program reuse and improving accuracy or cost validate the platform thesis.
The team bench is deep and regulated-market ready.	Haag evidence, senior engineering leadership, export-control program, and trust materials are complete.

Founder Action Plan

Timeframe	Action	Output
Before the next investor call	Prepare the ownership and rights package.	Cap table, allocation, effective basis, pro-rata rights, information rights, founder vesting, and board summary.
Before the next investor call	Disclose revenue quality.	Annual recurring revenue, annual contract value, recognized revenue, bookings bridge, services split, gross margin, retention, and concentration.
Before the next investor call	Bring one production customer proof packet.	Customer-controlled reference, deployment metrics, paid status, renewal, expansion, and reference permission.
Early diligence	Summarize NVIDIA and Siemens economics.	Counterparty-validated terms covering intellectual property, data, exclusivity, distribution, revenue share, roadmap control, and portability.
Early diligence	Evidence the compounding asset.	Cross-program Deep Physics Model or model-catalog reuse artifact with improving accuracy or cost.
Early diligence	Close the team-depth gap.	Haag prior-work evidence, senior engineering bench biographies, references, and org chart.
Early diligence	Complete regulated-buyer readiness.	Export-control memo, ISO certificate scope, SOC 2 Type II report, data processing addendum, subprocessor list, and security owner.
Ongoing tracking	Define revisit triggers.	Revenue-per-head trend, pilot-to-production conversion, gross-margin trend, partner-term updates, and allocation availability.

Decision

Screen: pass for this fund, with tracking warranted.
Confidence: medium.
Rationale: PhysicsX is a high-quality, well-validated deep-tech company, but a seed-sized fund cannot underwrite a fund-returning stake at roughly $2.4 billion without a special lower-basis allocation, pro-rata rights, verified high-margin recurring revenue, production-customer proof, and partner economics that preserve independent value capture.
What would move this to pursue: A meaningful lower-basis allocation with pro-rata, audited or management evidence of majority high-margin recurring software, one customer-controlled paid production reference, favorable NVIDIA and Siemens terms, and off-NVIDIA portability would move the verdict to pursue.
What would move this to pass: A token no-pro-rata stake, services-heavy low-margin revenue, no paid production reference, partner-subordinated value capture, no cross-program model reuse, or material financial inconsistency would strengthen the pass.
Recommended next step: Track the company and take at most one tightly scoped founder call if the fund wants optionality on the company rather than the current deal.
Founder preparation standard: Lead with ownership, revenue quality, production proof, partner economics, and model reuse rather than another partnership announcement or broad platform narrative.

How We Would Miss This One

The miss scenario: PhysicsX could be the rare company where the founder-market-fit thesis is exactly right, the forward-deployed model creates the proprietary physics-AI library incumbents cannot copy, and a structured lower-basis allocation becomes available despite the headline valuation.
Flip conditions: The pass should be revisited if the company verifies majority high-margin recurring software, shows a reusable model-catalog asset across programs, produces a customer-controlled paid production reference, proves favorable NVIDIA and Siemens terms with off-NVIDIA portability, and offers meaningful pro-rata-bearing ownership.
Revisit trigger / date: Revisit by 2027-06-15, or sooner if PhysicsX provides the ownership package plus the revenue-mix and production-reference proof packet.

Source Log

PhysicsX homepage
physicsx.ai
Company positioning, sectors, and public profile
Retrieved 2026-06-15company-controlled pagehigh that PhysicsX states the claims
PhysicsX about page
physicsx.ai
Founder roster, forward-deployed model, and team narrative
Retrieved 2026-06-15company-controlled pagehigh that PhysicsX states the claims
PhysicsX platform
physicsx.ai
Platform modules, data isolation, Deep Physics Models, Model Catalog, and deployment claims
Retrieved 2026-06-15company-controlled pagehigh that PhysicsX states the claims
PhysicsX industries
physicsx.ai
Target verticals and use-case claims
Retrieved 2026-06-15company-controlled pagehigh that PhysicsX states the claims
PhysicsX careers
physicsx.ai
Careers route and team narrative
Retrieved 2026-06-15company-controlled pagehigh
PhysicsX Greenhouse job board
job-boards.eu.greenhouse.io
Open roles, role mix, and company-stated Series B amount
Retrieved 2026-06-15company-linked applicant tracking pagehigh for role visibility
PhysicsX Series C announcement
physicsx.ai
Company-stated Series C, growth multiples, employee count, and investor list
Retrieved 2026-06-15company-controlled announcementmedium
PhysicsX Fourier neural operator blog
physicsx.ai
Technical narrative, FNO limitations, and Large Physics Model context
Retrieved 2026-06-15company-controlled technical blogmedium
PhysicsX pilots blog
physicsx.ai
Pilot-led go-to-market claim
Retrieved 2026-06-15company-controlled blogmedium
PhysicsX NVIDIA announcement
physicsx.ai
NVIDIA, PhysicsNeMo, Opora, and native-stack claims
Retrieved 2026-06-15company-controlled announcementmedium
PhysicsX Deutsche Telekom and NVIDIA announcement
physicsx.ai
Native NVIDIA, sovereign artificial-intelligence infrastructure, and Teamcenter claims
Retrieved 2026-06-15company-controlled announcementmedium
PhysicsX ISO announcement
physicsx.ai
ISO 27001 claim
Retrieved 2026-06-15company-controlled announcementmedium
PhysicsX Microsoft Discovery article
physicsx.ai
Microsoft Discovery, Azure, SOC 2, and agentic-engineering claims
Retrieved 2026-06-15company-controlled articlemedium
PhysicsX privacy policy
cdn.prod.website-files.com
Privacy entity and public legal surface
Retrieved 2026-06-15company-controlled legal documenthigh
LinkedIn company profile
linkedin.com
Company profile and headcount proxy
Retrieved 2026-06-15public profilemedium
Jacomo Corbo LinkedIn profile
linkedin.com
Founder public profile URL and current association
Retrieved 2026-06-15public profilemedium
Robin Tuluie LinkedIn profile
linkedin.com
Founder public profile URL and current association
Retrieved 2026-06-15public profilemedium
Nicolas Haag LinkedIn profile
linkedin.com
Founder public profile URL and role claim
Retrieved 2026-06-15public profilemedium
Crunchbase PhysicsX
crunchbase.com
Investor roster, employee band, and company funding profile
Retrieved 2026-06-15startup database (Crunchbase)medium
PhysicsX Instagram
instagram.com
Public social profile
Retrieved 2026-06-15public social profilemedium
Companies House — PHYSICSX LIMITED
find-and-update.company-information.service.gov.uk
UK entity, incorporation, company number, and prior name
Retrieved 2026-06-15public registryhigh
Companies House — PHYSICSX LIMITED officers
find-and-update.company-information.service.gov.uk
Directors and officer status
Retrieved 2026-06-15public registryhigh
Companies House — PHYSICSX LIMITED persons with significant control
find-and-update.company-information.service.gov.uk
Persons-with-significant-control status and control opacity
Retrieved 2026-06-15public registryhigh
Companies House accounts — PhysicsX Ltd group accounts to 31 Dec 2024
find-and-update.company-information.service.gov.uk
FY2024 loss, headcount, Series B timing, revenue policy, and turnover limitation
Retrieved 2026-06-15public registry filinghigh
Harvard SEAS alumni profile Jacomo Corbo
seas.harvard.edu
Corbo PhD, Formula 1, QuantumBlack, and McKinsey background
Retrieved 2026-06-15reputable institution profilehigh
Motor Sport Magazine Robin Tuluie profile
motorsportmagazine.com
Tuluie Formula 1, Mercedes, Ducati, Bentley, and simulation background
Retrieved 2026-06-15reputable press profilehigh
Bloomberg — PhysicsX Series C
bloomberg.com
Series C amount, valuation, prior valuation, and investor context
Retrieved 2026-06-15reputable press (Bloomberg)high
Yahoo Finance — PhysicsX raises $300 million
finance.yahoo.com
Series C, valuation, revenue quote, customer names, and investor context
Retrieved 2026-06-15reputable press syndication (Yahoo Finance)medium
Sifted — PhysicsX funding round
sifted.eu
Series C valuation, prior valuation, and headcount growth
Retrieved 2026-06-15reputable press (Sifted)medium
TechCrunch — PhysicsX Series A
techcrunch.com
Series A amount and company launch context
Retrieved 2026-06-15reputable press (TechCrunch)high
Tech.eu — PhysicsX Series B
tech.eu
Series B amount, timing, and total-after-B context
Retrieved 2026-06-15reputable press (Tech.eu)high
General Catalyst — investment in PhysicsX
generalcatalyst.com
Series A narrative and Velo3D reference
Retrieved 2026-06-15investor-authored postmedium
Siemens — PhysicsX collaboration
news.siemens.com
Siemens collaboration existence and LGM-Aero context
Retrieved 2026-06-15vendor and partner page (Siemens)high for collaboration existence
Microsoft UK Stories — PhysicsX
ukstories.microsoft.com
Microsoft-authored PhysicsX feature and partnership narrative
Retrieved 2026-06-15vendor and partner story (Microsoft)medium
NVIDIA PhysicsNeMo
developer.nvidia.com
NVIDIA physics-AI framework and partner-as-competitor context
Retrieved 2026-06-15vendor documentation (NVIDIA)high
GitHub — NVIDIA PhysicsNeMo
github.com
Open-source PhysicsNeMo repository evidence
Retrieved 2026-06-15open-source repositoryhigh
Siemens Simcenter PhysicsAI
siemens.com
Incumbent AI-assisted simulation product page
Retrieved 2026-06-15competitor product page (Siemens)medium
Ansys AI products
ansys.com
Incumbent AI product positioning
Retrieved 2026-06-15competitor product page (Ansys)medium
Synopsys multiphysics simulation
synopsys.com
Incumbent multiphysics simulation positioning
Retrieved 2026-06-15competitor product page (Synopsys)medium
Dassault SIMULIA
3ds.com
Incumbent simulation product positioning
Retrieved 2026-06-15competitor product page (Dassault Systemes)medium
COMSOL homepage
comsol.com
Incumbent simulation alternative
Retrieved 2026-06-15competitor homepage (COMSOL)medium
Altair PhysicsAI
altair.com
Incumbent AI simulation positioning and Siemens context
Retrieved 2026-06-15competitor product page (Altair)medium
Luminary Cloud homepage
luminarycloud.com
AI-native or cloud-native simulation challenger positioning
Retrieved 2026-06-15competitor homepagemedium
Monolith AI homepage
monolithai.com
AI for engineering test and validation positioning
Retrieved 2026-06-15competitor homepagemedium
SimScale homepage
simscale.com
Cloud simulation alternative positioning
Retrieved 2026-06-15competitor homepagemedium
SimScale pricing
simscale.com
Pricing-model proxy
Retrieved 2026-06-15competitor pricing pagemedium
Rescale AI Physics
rescale.com
AI physics and high-performance-computing alternative
Retrieved 2026-06-15competitor product pagemedium
Neural Concept homepage
neuralconcept.com
AI engineering competitor positioning
Retrieved 2026-06-15competitor homepagemedium
Navier AI homepage
navier.ai
Agentic computational-fluid-dynamics challenger positioning
Retrieved 2026-06-15competitor homepagelow
Pasteur Labs homepage
pasteurlabs.ai
Early AI science or physics-model challenger positioning
Retrieved 2026-06-15competitor homepagelow
Crunchbase Neural Concept
crunchbase.com
Peer funding and scale proxy
Retrieved 2026-06-15startup database (Crunchbase)medium
Crunchbase Luminary Cloud
crunchbase.com
Peer funding and scale proxy
Retrieved 2026-06-15startup database (Crunchbase)low
Crunchbase SimScale
crunchbase.com
Peer funding and scale proxy
Retrieved 2026-06-15startup database (Crunchbase)medium
Grand View Research — CAE market
grandviewresearch.com
Narrow CAE market-size public summary
Retrieved 2026-06-15market-sizing aggregator (Grand View Research)low
MarketsandMarkets — CAE market
marketsandmarkets.com
Narrow CAE market-size cross-check
Retrieved 2026-06-15market-sizing aggregator (MarketsandMarkets)low
Fortune Business Insights — Simulation Software Market
fortunebusinessinsights.com
Broad simulation software public-summary size
Retrieved 2026-06-15market-sizing aggregator (Fortune Business Insights)low
MarketsandMarkets — Simulation Software Market
marketsandmarkets.com
Broad simulation software public-summary size
Retrieved 2026-06-15market-sizing aggregator (MarketsandMarkets)low
Dassault Systemes 2025 Registration Document
investor.3ds.com
Broad software-domain TAM, software revenue, and incumbent benchmark
Retrieved 2026-06-15public-company filinghigh
Ansys 2024 Form 10-K
sec.gov
Simulation revenue, gross margin, and incumbent benchmark
Retrieved 2026-06-15public-company filinghigh
Altair Engineering 2024 Form 10-K
sec.gov
Software revenue, gross margin, and simulation comp
Retrieved 2026-06-15public-company filinghigh
Synopsys 2025 Form 10-K
sec.gov
Incumbent absorption and Ansys acquisition context
Retrieved 2026-06-15public-company filinghigh
Siemens Report 2025
siemens.com
Digital Industries, industrial software, and Altair acquisition context
Retrieved 2026-06-15public-company filinghigh
BLS Occupational Outlook — Architecture and Engineering
bls.gov
Labor and engineering occupation context
Retrieved 2026-06-15government statistics (BLS)high
Capterra SIMUL8 Professional Pricing
capterra.com
Per-user simulation software pricing proxy
Retrieved 2026-06-15marketplace listing (Capterra)medium
Vendr ANSYS Buyer Guide
vendr.com
Enterprise contract value proxy
Retrieved 2026-06-15pricing aggregator (Vendr)low
LinkedIn Jobs — PhysicsX senior data scientist
linkedin.com
Public technical hiring signal
Retrieved 2026-06-15public job postmedium

Access limitations: This research did not have access to PhysicsX's private data room, financial model, audited revenue detail, customer contracts, partner agreements, cap table, shareholder agreement, board materials, founder vesting, security reports, export-control materials, or live founder interviews. Public checks did not surface customer-controlled paid production case studies, public pricing, a full trust center, a subprocessor list, a data processing addendum, third-party ISO certificate scope, SOC 2 Type II report, or PhysicsX-specific export-control disclosure. Search-result and visibility checks are lower-confidence signals and are not treated as market share, customer proof, or proof of absence beyond the exact databases and queries checked.

Public disclaimer: This is an independent public teardown based solely on publicly available information retrieved on 2026-06-15. It is research support, not investment, legal, tax, or financial advice. Company-provided statements are labeled as claims, not verified facts. PhysicsX did not participate in or review this report. Errors and omissions are possible; final decisions remain with the reader.

What The Company Does#

Key Takeaways#

What Makes This Potentially Fundable#

The Four Holes To Close Before Fundraising#

Decision Snapshot#

IC Disagreement Map#

Scenario Range#

What Investors Will Test#

Claim Reconciliation: Inconsistencies Investors Will Catch#

Diligence Findings: Issues To Fix Before You Raise#

Data Room Readiness Checklist#

First-Call Agenda For The Startup#

Business Model And Revenue Signals#

Revenue Quality Checklist#

Funding And Ownership Context#

Founder, Team, And Related Entities#

Founder And Team#

Founder Competency Coverage#

Public Professional Footprint#

Founder-Market Fit Read#

Related Entities And Founder-Associated Companies#

Traction#

Customer Status Table#

Traction Signals#

Hiring And Org Momentum#

Traction Quality Read#

Competitive Landscape#

Category Visibility Snapshot#

Pricing And Competitive Benchmark#

Competitive Wedge#

Risks And Open Questions#

Pre-Mortem: The Most Likely Obituary#

Decision-Critical Unknowns#

Diligence Questions#

First Call#

Follow-Up#

Kill Criteria#

Double-Down Criteria#

Founder Action Plan#

Decision#

How We Would Miss This One#

Source Log#