TensorWave

June 16, 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. TensorWave did not participate in or review this report. Errors and omissions are possible; final decisions remain with the reader.

AMD, not NVIDIA. That is the entire bet. TensorWave is the cloud wagering that the chip everyone ranks second can run the next era of AI, and the public record gives the wager real footing: the Series A sits in SEC filings, and AMD itself hosts a TensorWave evaluation page on amd.com. Then the scale story arrives entirely in the company's own voice — a $350 million round at a $1.55 billion valuation, thousands of GPUs online, gigawatts of capacity — and none of it has surfaced in the registry yet. Five partners spent one day on those 59 public sources, citing every claim and trusting none of the pitch. The case turns on a figure no announcement discloses, the one that separates a durable infrastructure compounder from capital that quietly evaporates. The teardown follows that figure to the edge of the answer. Read it before the next round sets the price for you.

What The Company Does

TensorWave describes itself as an all-AMD Instinct AI cloud for training, fine-tuning, and inference, with bare-metal GPU infrastructure, managed AI infrastructure, ROCm support, fleet-management claims, and a sales-led request flow (TensorWave homepage; TensorWave docs; TensorWave connect) company claim. The pitch is not "another generic GPU cloud"; it is a specialized bet that AMD Instinct hardware, ROCm-era software, and AMD ecosystem alignment can create a durable alternative to NVIDIA-centered compute clouds.

The public record supports that TensorWave is real, late-stage, and capital-intensive. Its May 2025 Form D shows $100M sold of a $125M offering with 11 investors (SEC Form D) source-backed fact, while the company later announced a $350M Series B at a $1.55B valuation (TensorWave Series B blog) company claim. The core underwriting problem is that the customer, utilization, backlog, margin, debt, and depreciation evidence needed to value a GPU-capacity business is not public.

Key Takeaways

Investor reaction
TensorWave is impressive enough to keep on a watch list, but the current deal is not fundable for a seed mandate without a different entry instrument and corroborated economics.
Current verdict
pass with medium-high confidence. This is a reasoned pass, not a claim that the company is weak or fraudulent.
Why investors might lean in
The AMD compute shift is real enough to study, the category is enormous, AMD is visibly connected to TensorWave, the Series A is registry-confirmed, and the team has already shown unusual capital-formation ability.
Why investors might pull back
The $1.55B Series B is company-reported and not yet registry-confirmed, the 8,192-GPU fleet and more than 2 GW capacity are company claim, and the value-capture variables are unknown.
Highest-leverage fix
Share audited cohort unit economics, contracted backlog, utilization, GPU depreciation policy, capex financing terms, and AMD allocation terms under NDA.
Best next move
Keep TensorWave on a watch list and reopen only if a fund-fit entry vehicle appears alongside premium contracted backlog and durable margins.

What Makes This Potentially Fundable

The fundable version of TensorWave is the independent AMD cloud that becomes the default place for serious AI teams to run workloads when they want an alternative to NVIDIA-dependent capacity. In that version, AMD gains durable share, TensorWave receives better allocation or roadmap alignment than larger generic clouds, customers migrate enough ROCm workloads to create switching costs, and contracted demand lets the company finance capacity without being crushed by depreciation.

That is a real upside shape, but the public evidence does not yet prove the mechanism. IDC forecasts global AI infrastructure spending of $758B by 2029 (IDC) source-backed fact, and AMD's FY2024 filing shows Data Center revenue growth (AMD Form 10-K) source-backed fact. TensorWave-specific backlog, utilization, margins, debt tenor, and price realization remain private. For this fund, the public case fails on constructability: a company-reported $1.55B post-money valuation is not a seed-stage entry where a small check can reach or defend meaningful ownership.

The Four Holes To Close Before Fundraising

Hole	Investor Fear	What To Bring
Unit economics are invisible.	GPU clouds can create revenue while destroying equity value if utilization, depreciation, debt tenor, and gross margins do not clear the hurdle.	Cohort P&L, utilization logs, realized price by GPU model, depreciation policy, and lender terms.
Capacity and customer claims are not independently corroborated.	The company may be building ahead of demand, and logo strips may reflect pilots or partnerships rather than paid production.	Contracted backlog, take-or-pay schedule, customer references, and independent capacity substantiation.
AMD is supplier, investor, and competitor.	The same party that helps TensorWave can also squeeze allocation, pricing, or roadmap access while selling to larger rivals.	TensorWave-specific allocation agreement, pricing terms, roadmap commitments, and non-compete or conflict-management terms.
The entry does not fit a seed fund.	A $0.5M-$2.5M check at a company-reported $1.55B post cannot reach or defend an 8%-12% ownership target.	A materially earlier entry, secondary, SPV, strategic co-invest, or other structure that makes ownership constructable.

Decision Snapshot

One-sentence company description: TensorWave is a later-stage AMD Instinct GPU cloud that wants to become the independent compute platform for AI workloads moving beyond NVIDIA-only infrastructure.
Screen: pass
Confidence: medium-high
Suggested next action: Watch-list the company with explicit flip conditions rather than opening a normal seed-stage diligence process.
Why this matters now: AI infrastructure demand is enormous, but capital intensity and commodity pricing can erase venture value even when revenue grows.
Investor-readiness diagnosis: The public story is strong on category, fundraising, and team signal, but weak on the private economics that matter most.
Best founder use of this report: Prepare the economics, backlog, AMD-terms, and customer-proof packet that converts scale narrative into underwritable value capture.

IC Disagreement Map

The partners reviewed the same Evidence dossier independently: the power-law partner focused on fund-returning upside, the prepared-mind partner focused on category structure, the founder-jockey partner focused on team fit, the risk-reduction partner focused on fatal flaws, and the long-horizon partner focused on ownership and compounding. The final split was four pass votes and one hold vote, with the founder-jockey partner preserving the live dissent.

Power-law partner
passmedium confidence
The AMD compute shift could be a real omission-risk opportunity, but the public upside is unfalsifiable without backlog and premium pricing, and the company-reported entry does not support seed-fund ownership.
Would flip on
Contracted take-or-pay backlog at premium pricing plus a reachable sub-$1.55B entry would move this partner toward hold or pursue; uncontracted capacity or full-price Series B access confirms pass.
Prepared-mind partner
passmedium confidence
The why-now is real, but AMD specialization has not yet been proven as durable structure rather than a supply-arbitrage window.
Would flip on
AMD allocation priority, ROCm migration lock-in, and premium pricing would reopen the case; commodity pricing with no allocation edge confirms pass.
Founder-jockey partner
holdmedium confidence
The team is the best part of the case: founders are corroborated, the Series A is registry-confirmed, and capital formation is unusually strong for a neocloud.
Would flip on
Verified operating-at-scale depth and clean prior outcomes would move this partner toward pursue; contradicted outcomes or unverifiable technical leadership would move them toward pass.
Risk-reduction partner
passhigh confidence
Single-silicon dependence, AMD's triple role, commodity pricing, capital intensity, and unknown margins stack into a value-capture trap.
Would flip on
Durable margins, long-tenor financing, contracted backlog, and aligned AMD terms would move this partner to hold; thin margins or discretionary AMD supply confirms pass.
Long-horizon partner
passhigh confidence
This may be a good company, but it is the wrong vehicle: ownership is unreachable, follow-on needs are too large, and the compounding moat is unproven.
Would flip on
A below-market entry, proven compounding moat, and aligned AMD terms would move this partner to hold; no reachable ownership confirms pass.

Round-two review changed one vote: the prepared-mind partner moved from hold to pass because public evidence did not produce AMD allocation priority, ROCm switching-cost evidence, or realized-pricing proof. The founder-jockey partner confirmed hold, and the remaining partners confirmed pass, with one confidence increase.

Scenario Range

Scenario	What The Company Looks Like In 3-5 Years	Falsifiable Trigger To Watch	Earliest Evidence
Strikeout	TensorWave builds capacity ahead of contracted demand, pricing compresses toward commodity AMD rates, depreciation and financing costs outrun margins, and the company is absorbed or recapitalized at a loss.	Capacity announcements keep growing while contracted backlog, utilization, and cohort margins remain undisclosed or weak.	New capacity press without RPO disclosure, short-tenor debt, weak utilization, or a flat/down financing.
Base	TensorWave becomes a real AMD specialist cloud with revenue and recognizable customers, but margins remain cyclical and capital needs continue to dilute shareholders.	Realized pricing tracks alternatives like Vultr, and follow-on rounds remain necessary to keep fleet scale current.	Public pricing comparison, repeated nine-figure financings, and no evidence of premium take-or-pay demand.
Home run	TensorWave becomes the default independent AMD compute cloud, with priority supply, ROCm migration switching costs, contracted backlog, and premium pricing.	Take-or-pay backlog grows at premium pricing while AMD confirms TensorWave-specific allocation or roadmap advantage.	RPO schedule, customer migration case studies, AMD terms, and expanding cohort margins over 12-24 months.

What Investors Will Test

What We Looked For	Current Read	Investor Implication	Founder Prep Priority
Falsifiable fund-returning path	The market is large, but the ownership and economics path is not visible at the claimed stage.	A seed investor cannot underwrite this as a normal early-stage check.	Bring a fund-fit entry structure or accept that this is a later-stage-only opportunity.
Evidence beyond company narrative	Series A and founders are supported; Series B terms, capacity, customers, and certifications remain claims or private.	Load-bearing claims must be downgraded until corroborated.	Provide registry filings, customer proof, capacity evidence, and scoped audit reports.
Durable AMD specialization	AMD timing is plausible; exclusivity, allocation priority, and switching costs are not public.	The all-AMD wedge may be a moat or a trap.	Show allocation terms, ROCm migration evidence, and realized premium pricing.
Team coverage for the hardest build	Capital formation is evidenced; operating-at-scale and founder-level ROCm depth are thin.	The team can earn a call from some investors, but not override the economics gap.	Bring role ownership, operating metrics, references, and senior technical-leadership proof.
Incentive alignment and value capture	AMD helps TensorWave but also supplies rivals and participates in the same ecosystem.	The company may create value that AMD or larger clouds capture.	Show AMD terms, customer lock-in, and margin evidence.

Claim Reconciliation: Inconsistencies Investors Will Catch

Claim Or Metric	Where It Appears	Conflicting / Unreconciled Versions	Why Investors Flag It	How To Reconcile
Total capital raised	TensorWave about page TensorWave Series B blog	The about page says more than $140M raised, while the Series B announcement says $350M.	Stale or inconsistent funding totals create avoidable trust friction.	Publish a current financing timeline and reconcile with securities filings and cap table.
Series B legal status	TensorWave Series B blog Business Wire SEC EDGAR listing	The company and release-distribution sources state the round; the reviewed SEC listing did not show a matching Series B Form D on the retrieval date.	The absence may be a filing lag, but investors will ask for issuer, instrument, closing status, and terms.	Provide financing documents, filing accession if applicable, and counsel's explanation for filing status.
Co-founder count and title narrative	SEC Form D Crunchbase TensorWave Series B blog	SEC and Crunchbase support three co-founders; the Series B blog quotes Horton and Tomasik as co-founders but does not equally surface Tatarchuk.	Title drift can become a cap-table and founder-equity diligence question.	Share charter, cap table, 409A summary, and founder equity/title history.
Piotr Tomasik prior outcome	UNLV TechCrunch Publicis WSJ	Public sources support the Influential acquisition, but the role and economics attributed to Tomasik need founder validation.	Prior-outcome inflation is a team-quality diligence issue.	Provide role scope, equity/proceeds evidence, and references.
HIPAA readiness	TensorWave Trust Center TensorWave DPA	The trust portal markets HIPAA-related compliance, while the DPA prohibits customers from uploading PHI and other sensitive data.	A regulated buyer will test whether the badge maps to actual contractual scope.	Provide scoped reports, BAA terms, data-handling boundaries, and current customer eligibility rules.
Customer status	TensorWave homepage TensorWave Moreh case study Modular case study	The public record shows logos, case studies, and partner claims, but not paid production status for the full logo strip.	Investors distinguish paid active customers from pilots, evals, and partner marketing.	Provide customer-controlled references, contracts, committed volume, and renewal status.

Diligence Findings: Issues To Fix Before You Raise

2 deal-breaker3 high1 medium

No public utilization, margin, depreciation, or financing structure exists for a business where those variables are the investment case.
deal-breaker
Financial and KPI diligence
Fix
Prepare a cohort economics packet with lender terms and GPU useful-life policy.
Capacity claims are not independently substantiated and are not tied to contracted backlog.
deal-breaker
Commercial and infrastructure diligence
Fix
Provide third-party capacity evidence plus RPO and take-or-pay schedules.
AMD's supplier, investor, and competing-cloud role creates a value-capture conflict.
high
Commercial, legal, and strategy diligence
Fix
Share AMD allocation, pricing, roadmap, and conflict protections.
The company-reported Series B makes the opportunity structurally mismatched for a seed fund.
high
Portfolio construction and financing diligence
Fix
Identify a fund-fit entry or accept that this is not an early-stage mandate fit.
Customer logos and partner pages do not prove paid production use.
high
Customer diligence
Fix
Produce referenceable customers, contracts, committed volume, and renewal evidence.
Compliance marketing and DPA scope need reconciliation.
medium
Legal and procurement diligence
Fix
Provide scoped audit reports, BAA position, DPA updates, and procurement packet.

Data Room Readiness Checklist

Procurement-readiness score: partial
Rationale: TensorWave has public docs, a trust portal, and legal pages, but audit reports are gated, HIPAA-related messaging needs reconciliation with the DPA, and enterprise buyers will still require security, data-processing, export-control, and infrastructure-resilience evidence.

Data Room Area	What Investors Expect	Current Read	Status	Priority
Company overview & corporate	One-pager, incorporation, good standing, entity structure	TensorWave Inc. and the founder roster are supported by SEC and profile evidence, but corporate documents are private.	`partial`	High
Financials	Monthly P&L, three-year model, burn, runway, cohort economics	No public revenue, utilization, margin, burn, or debt structure is available.	`missing`	Critical
Cap table & funding history	Clean cap table, prior rounds, SAFEs or notes, 409A	Series A is registry-confirmed; seed and Series B terms need reconciliation.	`partial`	High
Legal & IP	Bylaws, board consents, IP assignments, material contracts	Terms, DPA, privacy, and trust surfaces are public; IP and material contracts are private.	`partial`	Medium
Product & technology	Roadmap, architecture, security/compliance docs, operational reliability	Docs and GitHub signals exist; fleet, reliability, capacity, and ROCm lock-in evidence remain thin.	`partial`	High
Team	Org chart, key employment/advisor agreements, vesting	Founders and headcount proxies are public; role scope and senior technical-bench proof are private.	`partial`	Medium
Customers & traction	Retention cohorts, ARR bridge, pipeline, references	Logos and case studies are public; paid status and backlog are not public.	`missing`	Critical
Market & competition	TAM/SAM/SOM, competitive landscape, pricing	Market sizing and competitor pages exist; TensorWave-specific SOM and realized pricing are unknown.	`partial`	High

First-Call Agenda For The Startup

Time	Topic	Founder Goal	Evidence To Bring
0-10 minutes	Frame the AMD thesis	Explain why AMD specialization is durable structure, not a temporary supply window.	AMD roadmap, customer demand drivers, and allocation terms.
10-25 minutes	Unit economics and financing	Prove GPU capacity earns durable value after depreciation and financing costs.	Cohort P&L, utilization, gross margin, debt tenor, and GPU useful-life policy.
25-35 minutes	Backlog and customers	Convert logo and capacity narrative into contracted demand.	RPO, take-or-pay contracts, customer references, and committed volume.
35-45 minutes	AMD relationship	Defuse supplier/investor/competitor concentration risk.	Allocation, pricing, roadmap, and non-compete or conflict terms.
45-55 minutes	Team operating depth	Show the team can operate at scale, not only raise capital.	Data-center commissioning records, uptime, senior-leader references, and ROCm ownership.
55-60 minutes	Fund-fit entry	Decide whether this fund can even construct a position.	Entry options, secondary or SPV terms, and next financing plan.

Business Model And Revenue Signals

Sales-led GPU cloud model with request-for-quote flow.
inferenceCompany sales pagemedium
SourceTensorWave connect
Supports enterprise/direct-sales posture, not revenue quality.
Series A capital raised is registry-confirmed.
source-backed factSEC filing and listinghigh
SourceSEC Form D SEC EDGAR listing
Confirms capital formation, not revenue.
Series B amount and valuation are company-reported.
Company announcement and release distributionmedium
SourceTensorWave Series B blog Business Wire
Treat as stage signal until registry or documents confirm.
Reported GPU-hour pricing and ARR figures appear in press citing company context.
Reputable trade press citing company or market sourceslow-medium
SourceNext Platform TechCrunch
Useful for scenarios only; do not treat as audited revenue.
Public revenue, ARR, gross margin, backlog, utilization, burn, and debt terms are not available.
Public record reviewedhigh
SourceSEC submissions API SEC EDGAR listing
This is the central diligence gap.

TensorWave's homepage, docs, and connect page were checked and showed no public self-serve rate card. This carries little weight because enterprise GPU clouds often price through sales contact rather than published list rates.

No public source reviewed here reports audited TensorWave revenue, profit, backlog, utilization, gross margin, debt structure, or burn. Any revenue or margin scenario is an estimate until the company provides account-level and cohort-level evidence.

Revenue Quality Checklist

Before TensorWave can move from pass to hold or pursue for this fund, the founder should be ready to show annual recurring revenue, contracted backlog, remaining performance obligations, take-or-pay versus on-demand mix, utilization by GPU cohort, realized blended price per GPU-hour, gross margin by workload, depreciation policy, debt tenor and covenants, customer concentration, renewal rates, churn, expansion, and gross revenue retention or net revenue retention where applicable. These metrics must reconcile across contracts, billing, accounting exports, lender materials, and the board model.

Funding And Ownership Context

Item	Public Read	Evidence Label	Diligence Request
Total raised / rounds	Series A is registry-confirmed at $100M sold; the company reports a later $350M Series B and earlier press described seed financing.	`source-backed fact` for Series A and `company claim` for Series B	Share full financing history, closing documents, and cap table.
Instruments (equity / SAFE / notes)	Series A Form D is visible; seed and Series B instruments are not public.	`unknown` for full instrument stack	Share SAFEs, notes, preferred stock terms, side letters, and conversion history.
Investors	AMD and Magnetar are named in company and press materials; Series A filing confirms investors existed but not full terms.	`company claim` for investor roles and `source-backed fact` for offering filing	Reconcile investor list and ownership with cap table.
Post-money valuation	Company reports $1.55B for the Series B.	`company claim`	Share financing documents and 409A context.
Cap table / ownership	Not public.	`unknown`	Show founder ownership, investor ownership, option pool, and pro forma next-round ownership.
Burn & runway	Not public.	`unknown`	Share burn, cash, committed capex, debt service, and runway by utilization case.
Next-round plan	Not public.	`unknown`	Tie the next financing to utilization, backlog, margin, and capacity milestones.

SEC EDGAR was checked and showed no Series B Form D filing for TensorWave as of the retrieval date. The absence may reflect filing lag rather than a failed close, but the company-reported $1.55B valuation should stay labeled company claim until registry or closing documents confirm it.

Founder And Team

Darrick Horton
Co-founder and CEO; central to the company and capital-formation story.
Darrick Horton LinkedIn Darrick Horton personal site SEC Form D Forbes
Evidencesource-backed fact for identity and public profile signals; detailed operating scope needs validation.
Confidencemedium-high
Piotr Tomasik
Co-founder and President/COO; relevant to operations and prior-outcome narrative.
Piotr Tomasik LinkedIn UNLV SEC Form D TechCrunch
Evidencesource-backed fact for identity and public profile signals; prior-outcome economics need founder validation.
Confidencemedium
Jeff Tatarchuk
Co-founder and Chief Growth Officer; relevant to go-to-market and prior infrastructure history.
SEC Form D Crunchbase TechCrunch
Evidencesource-backed fact for roster presence; detailed profile depth is incomplete.
Confidencemedium
TensorWave company profile
LinkedIn dataset indicates a 101-250 employee band and public company profile signal.
TensorWave LinkedIn Crunchbase
Evidencesource-backed fact
Confidencemedium
Open roles and operating footprint
Ashby careers page showed many open roles, including data-center technician roles across multiple locations.
TensorWave careers
Evidencesource-backed fact
Confidencemedium

Founder Competency Coverage

Domain depth
Darrick Horton
claimed
Piotr Tomasik
claimed
Jeff Tatarchuk
claimed
What The Evidence Is
Public sources support prior startup and infra-adjacent narratives, but detailed GPU fleet operating depth is not independently proven.
Technical build capability
Darrick Horton
claimed
Piotr Tomasik
claimed
Jeff Tatarchuk
unknown
What The Evidence Is
TensorWave's GitHub and ScalarLM are org-level artifacts; founder-level authorship is not public.
Product
Darrick Horton
claimed
Piotr Tomasik
claimed
Jeff Tatarchuk
unknown
What The Evidence Is
Company pages and docs show product breadth, but founder-specific product ownership is private.
GTM / sales
Darrick Horton
unknown
Piotr Tomasik
claimed
Jeff Tatarchuk
claimed
What The Evidence Is
Tatarchuk's role implies go-to-market relevance, but public proof of repeatable enterprise sales execution is thin.
Leadership / hiring
Darrick Horton
claimed
Piotr Tomasik
claimed
Jeff Tatarchuk
claimed
What The Evidence Is
Employee-band and careers signals support growth, while org chart and executive scope are private.
Fundraising history
Darrick Horton
evidenced
Piotr Tomasik
evidenced
Jeff Tatarchuk
evidenced
What The Evidence Is
The Series A filing and press support major capital formation under this team.
Prior founding outcomes
Darrick Horton
claimed
Piotr Tomasik
needs founder validation
Jeff Tatarchuk
claimed
What The Evidence Is
VMAccel and Influential signals exist, but outcomes, scale, and individual economics require validation.

Operating GPU and data-center capacity at high utilization
high
Team coverage today
claimed and not independently evidenced
What would close it
Uptime, utilization, commissioning history, and references from customers or operators.
AMD/ROCm software depth
high
Team coverage today
Org-level source-backed fact; founder-level unknown
What would close it
Named technical owners, contribution history, and ROCm migration case studies.
Capital formation for infrastructure buildout
low
Team coverage today
evidenced
What would close it
Cap table and financing terms would strengthen, not create, the signal.
Enterprise and regulated-buyer readiness
medium
Team coverage today
partial
What would close it
Scoped audit reports, DPA/BAA clarity, export-control program, and buyer references.

Public Professional Footprint

These checks of public technical artifacts, professional social presence, and education or credential traces carry low decision weight by design; they sharpen founder-call questions but do not alone change the verdict.

Technical artifacts (repo level)
Darrick Horton
unknown
Piotr Tomasik
unknown
What The Public Record Shows
TensorWave's GitHub organization and ScalarLM repository exist, but public evidence does not attribute core product authorship to individual founders.
Professional social content
Darrick Horton
evidenced
Piotr Tomasik
evidenced
What The Public Record Shows
LinkedIn profiles exist for Horton and Tomasik; the analysis uses them for role and profile context, not as proof of underlying claims.
Education / credentials
Darrick Horton
evidenced
Piotr Tomasik
evidenced
What The Public Record Shows
Forbes and UNLV provide public profile or credential context; the weight is low relative to operating evidence.
Publications / patents / certifications
Darrick Horton
unknown
Piotr Tomasik
unknown
What The Public Record Shows
No decision-critical founder-level patent or certification proof changes the verdict.

Founder-Market Fit Read

What is promising: The team has already assembled a large capital stack and public employee footprint in a category where capital formation is a real operating competency.
What is missing: Verified GPU fleet operating depth, ROCm/software ownership, data-center uptime evidence, and clean prior-outcome economics.
What to prepare: Founder role scope, operating references, data-center and utilization track record, senior technical-leader references, and cap-table/title reconciliation.

TensorWave Inc.
Delaware corporation and issuer in the Series A Form D.
SEC Form D SEC submissions API
Evidencesource-backed fact
Confidencehigh
Follow-upConfirm full corporate structure and financing history in the data room.
VMAccel
Prior infrastructure venture associated with Horton and Tatarchuk in press sources.
TechCrunch
Evidencesource-backed fact for public mention and needs founder validation for scale/outcome
Confidencemedium
Follow-upVerify operating scale, ownership, outcome, and lessons transferred to TensorWave.
Influential
Prior company associated with Tomasik in some public sources; Publicis acquired Influential.
UNLV Publicis WSJ
Evidencesource-backed fact for acquisition and needs founder validation for Tomasik role/economics
Confidencemedium
Follow-upVerify role, equity, proceeds, and references.
TensorWave GitHub / ScalarLM
Public technical surface for TensorWave.
TensorWave GitHub ScalarLM
Evidencesource-backed fact for repo existence and activity
Confidencemedium
Follow-upDo not overweight until customer lock-in and author ownership are shown.

Traction

Customer Status Table

AMD
AMD hosts a TensorWave evaluation page, and TensorWave names AMD as a strategic investor and partner (AMD TensorWave eval page; TensorWave Series B blog).
Yes for the AMD page; no for investor terms.MarketplaceUnknownUnknownUnknownmedium
Moreh
TensorWave publishes a Moreh case study about MI355X inference and model-weight download time (TensorWave Moreh case study).
No; this is TensorWave-controlled.Claim onlyUnknownUnknownUnknownlow
Modular
Modular publishes a TensorWave case study (Modular case study).
Yes, but it is partner-controlled rather than proof of paid production.MarketplaceUnknownUnknownUnknownmedium
Meta
TensorWave lists Meta in a trusted-by logo strip (TensorWave homepage).
No customer-controlled TensorWave confirmation found in reviewed public searches.Claim onlyUnknownUnknownUnknownlow
Luma
TensorWave lists Luma in a trusted-by logo strip (TensorWave homepage).
No customer-controlled TensorWave confirmation found in reviewed public searches.Claim onlyUnknownUnknownUnknownlow

Meta and Luma's public sites were checked and showed no customer-controlled TensorWave confirmation. These absences carry weight because homepage logos imply tier-one customers without independent proof.

Traction Signals

Registry-confirmed Series A offering
source-backed facthigh
SourceSEC Form D SEC EDGAR listing
Full cap table and instrument terms.
Company-reported Series B and expansion narrative
medium
SourceTensorWave Series B blog Business Wire
Registry filing, financing documents, and cap table.
Public docs and customer portal posture
medium
SourceTensorWave docs TensorWave homepage
Active customer usage and support history.
GitHub organization and ScalarLM repository
source-backed factmedium
SourceTensorWave GitHub ScalarLM
Customer lock-in and founder/employee authorship.
Public hiring footprint
source-backed factmedium
SourceTensorWave careers TensorWave LinkedIn
Actual headcount, role fill rate, and org chart.
Glassdoor small-sample employee signal
source-backed factmedium
SourceGlassdoor
More reviews and internal retention data.

Hiring And Org Momentum

LinkedIn company page
source-backed factPublic company dataset supports a 101-250 employee band and company profile signal.
SourceTensorWave LinkedIn
Consistent with a later-stage infrastructure company.
Company careers page
source-backed factAshby careers page showed roughly 30 open roles, including data-center technician roles in multiple states.
SourceTensorWave careers
Supports operating footprint buildout, though not role fill or quality.
Glassdoor
source-backed factGlassdoor company profile showed a small review sample and positive rating.
SourceGlassdoor
Useful as a weak org signal, not a culture conclusion.
Company technical footprint
source-backed factTensorWave GitHub and ScalarLM exist.
SourceTensorWave GitHub ScalarLM
Shows technical surface, not production reliability or individual founder depth.

Indeed was checked and showed no matching employer listing. This carries little weight because Ashby remains the primary public careers hub and LinkedIn surfaced separate job URLs.

Role-mix read: The public hiring footprint is consistent with infrastructure buildout and data-center operations, but it does not prove operating quality or utilization.
Momentum read (growth / steady / contraction / unknown): growth intent is plausible from public hiring and financing posture, while realized operating momentum remains unknown.

Traction Quality Read

Traction Dimension	Current Status	Good Enough For First Call?	Needed For Deep Diligence
Paid production demand	`unknown`	No; it is a gating question.	Contracts, references, committed spend, and renewal evidence.
Utilization	`unknown`	No; it is central to the investment case.	Fleet-level and cohort-level utilization exports.
Customer quality	`company claim` and partner-controlled evidence	Yes, only as a question list.	Customer-controlled references and status mapping.
Capital formation	`source-backed fact` for Series A and `company claim` for Series B	Yes.	Cap table, instruments, and closing documents.
Technical maturity	`company claim` and org-level `source-backed fact`	Yes, as diligence input.	Reliability metrics, customer workloads, and architecture review.

Competitive Landscape

Segment	Examples	Customer Alternative	Pressure On Company
Direct GPU and AI infrastructure clouds	CoreWeave, Lambda, Crusoe, Nebius, Together AI, Nscale	Buy capacity from larger or better-known AI infrastructure providers.	Scale, financing, availability, and customer trust can compound against TensorWave.
AMD-capacity alternatives	Vultr, Crusoe, Azure	Buy AMD GPU capacity without betting on a single specialist.	The AMD wedge is not unique unless allocation or software lock-in is proven.
Hyperscalers and incumbents	AWS, Google Cloud, Azure, Oracle Cloud, NVIDIA	Use a broader cloud or incumbent AI stack.	Distribution, procurement trust, and existing spend relationships are strong.
Marketplaces and lower-cost substitutes	Runpod, Paperspace, Voltage Park, Vast.ai	Source capacity from more flexible or lower-level platforms.	Price transparency and commodity capacity can compress premium pricing.

Category Visibility Snapshot

Google / USAMD GPU cloud

Captured page-one results (2026-06-15): TensorWave appeared on page one alongside AMD-specific and cloud provider results.
Was the company present?: Yes

Google / GBAMD GPU cloud

Captured page-one results (2026-06-15): TensorWave appeared near the top of AMD-specific results.
Was the company present?: Yes

Google Search in the United States and United Kingdom was checked for broader GPU-cloud and NVIDIA-alternative queries and did not surface TensorWave on page one. These absences carry limited weight because TensorWave still appears on AMD-specific category queries.

Visibility read (inference, low confidence): TensorWave has AMD-specific visibility, but it does not yet own generic GPU-cloud discovery. That supports a focused AMD niche while showing broader category mindshare risk.

Pricing And Competitive Benchmark

Alternative	What It Offers	Public Price Signal	Price Vs. This Company	Evidence Label
TensorWave	AMD Instinct GPU cloud and managed AI infrastructure.	No public self-serve TensorWave rate card found; Next Platform cites reported TensorWave pricing.	Baseline remains private and partially press-cited.	`unknown` for official price book and `company claim` for press-cited figures.
Vultr	Public cloud with AMD GPU availability.	Public page supports AMD GPU cloud positioning and price comparison context (Vultr).	Direct AMD-capacity pressure.	`source-backed fact` for public page
CoreWeave	Large public AI infrastructure cloud.	Filings show scale, backlog, and economics rather than simple list pricing (CoreWeave IPO prospectus).	Benchmark for contracted backlog and depreciation risk.	`source-backed fact`
Lambda and Runpod	GPU cloud alternatives with public product/pricing surfaces.	Public pages and pricing surfaces reviewed (Lambda instances; Runpod pricing).	Lower-friction alternatives can pressure price expectations.	`source-backed fact` for public pages
Hyperscalers	Cloud GPU services inside broader platforms.	Public GPU pages reviewed (AWS; Google Cloud; Azure; Oracle Cloud).	Incumbents can bundle compute with procurement trust.	`source-backed fact` for public pages

Price positioning read: TensorWave may be able to compete on AMD performance-per-dollar, but public evidence does not prove premium pricing, gross margin, or long-tenor contracts.
Price claims to correct or substantiate: Publish or share the official price book, discount schedule, realized ARPU, and price comparison against AMD-capacity alternatives.

Competitive Wedge

The credible wedge is a deep AMD specialization: Instinct hardware focus, ROCm-era software, possible memory-capacity advantages, and an AMD relationship that could help allocation or roadmap access. That wedge becomes defensible only if customers migrate real workloads, sign long contracts, and face switching costs or better service levels that generic clouds cannot match. Today, public evidence shows a focused position but not a moat. Capacity is replicable, AMD supplies a broader ecosystem, and the business can become commodity unless TensorWave proves protected allocation and customer lock-in.

Risks And Open Questions

Risk	Severity	Evidence	What To Ask
Value capture may be structurally weak.	high	TensorWave's margins and financing are `unknown`; CoreWeave filings show depreciation-heavy economics.	What are cohort gross margins after depreciation and debt service?
Capacity may be built ahead of demand.	high	8,192 GPUs and more than 2 GW are `company claim`; contracted backlog is not public.	What fraction is take-or-pay, and what is the RPO schedule?
AMD may be a conflicted gatekeeper.	high	AMD is publicly connected while same-category capacity appears elsewhere.	What allocation, pricing, roadmap, and conflict protections does TensorWave have?
Fund ownership may be unreachable.	high	Company-reported Series B valuation is $1.55B.	Is any fund-fit sub-$1.55B entry available?
Customer logos may not be paid production.	high	Logos and case studies are visible; customer-controlled paid status is absent.	Which logos are paid, production, committed, and referenceable?
Export controls may narrow global expansion.	medium	Federal Register and AMD filing context support advanced-computing controls.	What export-compliance program governs customer eligibility and deployment?

Pre-Mortem: The Most Likely Obituary

Cause of death (one sentence): TensorWave built ahead of contracted demand on fast-depreciating AMD GPU capacity, and when utilization and pricing failed to clear the financing burden, it could not raise the next aligned round.
The causal chain (3-5 steps from today to the shutdown): First, the company scaled capacity faster than take-or-pay demand. Second, AMD capacity became available from larger clouds and price pressure increased. Third, utilization and gross margin lagged while depreciation and debt service kept compounding. Fourth, a new round required flat or down terms that neither strategic nor financial investors wanted to lead. Fifth, the assets were sold, the company was recapitalized, or a larger cloud absorbed the team and capacity.
The earliest observable warning sign: Capacity headlines continue while contracted backlog, utilization, and margins remain undisclosed.
The question that defuses this chain today: What fraction of the fleet is contracted take-or-pay, at what realized price, financed with what debt tenor and depreciation policy?

Decision-Critical Unknowns

Unknown	Why It Is Decision-Critical	Best Evidence	Decision Effect
Utilization and gross margin by cohort	It determines whether revenue converts to value.	Cohort P&L, utilization logs, and margin bridge.	Durable margins move toward hold; weak margins confirm pass.
Contracted backlog and take-or-pay mix	It separates committed demand from speculative capacity.	RPO schedule, contracts, customer references, and revenue recognition policy.	Strong backlog reopens diligence; mostly on-demand demand confirms pass.
Debt tenor and depreciation policy	It bounds downside for fast-depreciating GPUs.	Lender terms, covenants, balance sheet, and useful-life schedule.	Long-tenor matched financing improves the case; short-tenor debt confirms risk.
AMD allocation and conflict terms	It decides whether AMD is a moat or a trap.	Supply agreement, roadmap terms, pricing, and non-compete or conflict protections.	Priority terms reopen the AMD thesis; discretionary terms confirm commodity risk.
Customer status by logo	It controls traction quality.	Paid contracts, usage logs, renewal status, and on-record references.	Paying production customers improve confidence; pilot-only logos weaken it.
Fund-fit entry	It determines whether this fund can participate constructably.	Entry terms, secondary or SPV terms, and ownership math.	A reachable entry reopens the case; full-price access confirms pass.

Diligence Questions

First Call

Question	Why It Matters	Good Evidence
What are cohort utilization, gross margin, depreciation policy, and capex financing terms?	It is the core value-capture test.	Cohort P&L, utilization logs, lender terms, and useful-life schedule.
What share of current capacity is contracted take-or-pay versus on-demand or idle?	It separates financed demand from speculative capacity.	RPO schedule, committed contracts, and revenue-recognition policy.
What are TensorWave-specific AMD allocation, pricing, roadmap, and conflict terms?	It tests whether AMD is strategic moat or conflicted supplier.	AMD agreement, roadmap commitments, and non-compete or conflict protections.
Which named logos are paying production customers?	It converts logo traction into underwritable demand.	Customer references, contracts, committed volume, usage, and renewal data.
Is there any fund-accessible entry below the company-reported post-money valuation?	It determines whether this fund can construct a position.	Secondary, SPV, earlier allocation, or co-invest terms.

Follow-Up

Question	Why It Matters	Good Evidence
Has the Series B legally closed, and when will the Form D or equivalent filing appear?	It reconciles the funding narrative.	Filing accession, financing documents, or counsel explanation.
What is the export-control compliance program for global deployment?	It bounds customer eligibility and geography.	Export-control policy, customer screening, license history, and legal review.
What customer switching costs exist after ROCm migration?	It tests the compounding moat.	Migration case studies, reserved contracts, engineering effort logs, and renewal behavior.
What was each founder's role in operating prior infrastructure ventures?	It validates founder-market fit.	References, cap table or role records, operational metrics, and outcome documentation.
How do trust badges map to contractual scope and sensitive-data restrictions?	It determines procurement readiness.	Scoped reports, DPA, BAA position, and customer security questionnaire history.

Kill Criteria

Kill Criterion	Evidence That Would Trigger It
Capacity is materially uncontracted and gross margins are thin or negative after depreciation and debt service.	Cohort P&L, utilization, RPO, and lender terms show low utilization or short-tenor financing against fast-depreciating GPUs.
AMD supply is discretionary while AMD competes and supplies larger rivals on equal or better terms.	AMD terms show no priority allocation, pricing protection, roadmap advantage, or conflict protection.
No fund-fit entry exists.	Only full-price Series B or later-stage access is available, leaving no path to meaningful ownership.
The named customer logos are pilots, partner marketing, or evals rather than paid production.	Customer references and contracts fail to support paid committed use.
Series B terms cannot be reconciled with legal closing records or financing documents.	The company cannot provide documents or a credible filing explanation.

Double-Down Criteria

Double-Down Criterion	Evidence That Would Justify More Diligence
Contracted take-or-pay backlog exists at premium pricing.	RPO and contracts show durable demand above commodity AMD alternatives.
Cohort economics are durable after depreciation and financing costs.	Gross margins, utilization, and lender terms support equity value creation.
AMD gives TensorWave priority or protected economics.	Written AMD terms show allocation, roadmap, pricing, and conflict advantages.
Customer migration creates ROCm or workflow switching costs.	Customers confirm migration effort, retained usage, and expansion.
A reachable entry vehicle appears.	A secondary, SPV, strategic co-invest, or earlier allocation makes ownership constructable for the fund.

Founder Action Plan

Timeframe	Action	Output
Before the next investor call	Assemble cohort unit economics, utilization, depreciation policy, and capex financing evidence.	An exportable economics packet that lets investors model value capture.
Before the next investor call	Produce contracted backlog and take-or-pay schedule.	Demand-certainty packet with RPO, committed volumes, and customer concentration.
Before the next investor call	Substantiate fleet and power capacity independently.	Leases, power agreements, interconnection evidence, and facility confirmations.
Within 2-4 weeks	Document AMD allocation and roadmap terms.	AMD relationship memo with pricing, allocation, and conflict protections.
Within 2-4 weeks	Convert logo claims into customer-controlled proof.	Reference list, contracts, usage logs, and renewal status.
Within 2-4 weeks	Reconcile funding history and title/cap-table questions.	Financing timeline, Form D status, cap table, founder title history, and prior-outcome evidence.

Decision

Screen: pass
Confidence: medium-high
Rationale: TensorWave is a serious company in a real market, but the fund cannot underwrite the current opportunity without a constructable entry and the private economics that prove durable value capture.
What would move this to pursue: A fund-fit entry plus contracted take-or-pay backlog at premium pricing, durable cohort margins, matched financing, and protected AMD allocation.
What would move this to pass: No fund-fit entry, uncontracted capacity, thin or negative margins after depreciation, discretionary AMD supply, or pilot-only logo traction.
Recommended next step: Keep on watch list and revisit on the next financing event, a disclosed economics packet, or a specific sub-$1.55B entry opportunity.
Founder preparation standard: Bring audited/exportable economics and contracts, not more capacity or funding narrative.

How We Would Miss This One

The miss scenario: TensorWave could become the one that got away if AMD demand keeps compounding, TensorWave has already locked in priority supply and contracted backlog privately, and the company is building a durable ROCm migration surface that public evidence cannot yet see.
Flip conditions: The screen should flip if a reachable entry exists, contracted backlog is premium-priced and growing, unit economics are durable after depreciation and financing, and AMD terms protect rather than squeeze TensorWave.
Revisit trigger / date: Revisit within 12 months, at the next financing event, or immediately if TensorWave discloses contracted backlog, margin evidence, and AMD allocation terms or offers a fund-fit entry.

Source Log

TensorWave homepage
tensorwave.com
Product positioning, logo strip, AMD-cloud narrative, official profiles.
Retrieved 2026-06-15Company-controlled homepagehigh
TensorWave Series B announcement blog
tensorwave.com
Company-reported Series B, valuation, capacity, investor narrative, and expansion claims.
Retrieved 2026-06-15Company-controlled funding announcementmedium
TensorWave connect page
tensorwave.com
Sales-led go-to-market posture.
Retrieved 2026-06-15Company-controlled sales pagehigh
TensorWave about page
tensorwave.com
Funding narrative, team narrative, and total-raised inconsistency.
Retrieved 2026-06-15Company-controlled about pagemedium
TensorWave DPA
tensorwave.com
Sensitive-data restrictions and DPA scope.
Retrieved 2026-06-15Company-controlled legal policyhigh
TensorWave Trust Center
security.tensorwave.com
SOC 2, ISO 27001, HIPAA-related trust claims and gated report posture.
Retrieved 2026-06-15Company-controlled trust portalmedium
TensorWave docs homepage
docs.tensorwave.com
Product and technical documentation posture.
Retrieved 2026-06-15Company-controlled documentationhigh
TensorWave Moreh case study
tensorwave.com
Moreh relationship and MI355X case-study claim.
Retrieved 2026-06-15Company-controlled case studymedium
TensorWave Series A announcement
tensorwave.com
Company-reported Series A narrative and capacity claims.
Retrieved 2026-06-15Company-controlled funding announcementmedium
TensorWave careers
jobs.ashbyhq.com
Public hiring footprint and role mix.
Retrieved 2026-06-15Company-linked ATS pagemedium
TensorWave LinkedIn company page
linkedin.com
Employee band and company profile signal.
Retrieved 2026-06-15Public professional profile datasetmedium
TensorWave Crunchbase profile
crunchbase.com
Founder roster, profile, and company database context.
Retrieved 2026-06-15Crunchbase company datasetmedium
Darrick Horton LinkedIn profile
linkedin.com
Founder identity and public profile context.
Retrieved 2026-06-15Public professional profile datasetmedium
Darrick Horton personal site
darrickhorton.com
Founder public profile context.
Retrieved 2026-06-15Founder-controlled personal sitemedium
Piotr Tomasik LinkedIn profile
linkedin.com
Founder identity and public profile context.
Retrieved 2026-06-15Public professional profile datasetmedium
Forbes Darrick Horton profile
forbes.com
Forbes 30 Under 30 public profile signal.
Retrieved 2026-06-15Reputable press profilemedium
UNLV Piotr Tomasik article
unlv.edu
Piotr Tomasik public biography and prior-company narrative.
Retrieved 2026-06-15University profile articlemedium
Publicis Influential acquisition press release
publicisgroupe.com
Publicis acquisition of Influential.
Retrieved 2026-06-15Publicis corporate press releasehigh
WSJ Influential acquisition article
wsj.com
Influential founder/CEO attribution and acquisition context.
Retrieved 2026-06-15Independent business pressmedium
TechCrunch TensorWave founding coverage
techcrunch.com
Founders, seed context, ARR claims, and early company narrative.
Retrieved 2026-06-15Independent technology pressmedium
Reuters TensorWave Series A coverage
reuters.com
Series A reporting cross-check.
Retrieved 2026-06-15Independent business presshigh
Business Wire TensorWave Series B
businesswire.com
Series B release-distribution copy.
Retrieved 2026-06-15Release-distribution servicemedium
DCD TensorWave Series B coverage
datacenterdynamics.com
Series B and data-center context.
Retrieved 2026-06-15Data-center trade pressmedium
Next Platform TensorWave AMD cluster article
nextplatform.com
Pricing, ARR, and financing-context claims.
Retrieved 2026-06-15Technical trade pressmedium
SEC Form D for TensorWave
sec.gov
Series A offering amount, founders/officers, entity context.
Retrieved 2026-06-15SEC Form D filinghigh
SEC EDGAR TensorWave Form D listing
sec.gov
Public filing listing and checked absence context for later Form D filings.
Retrieved 2026-06-15SEC EDGAR issuer listinghigh
SEC TensorWave submissions API
data.sec.gov
Public filing index context.
Retrieved 2026-06-15SEC submissions JSONhigh
TensorWave GitHub organization
github.com
Public technical footprint.
Retrieved 2026-06-15Public repository organizationmedium
TensorWave ScalarLM repository
github.com
ROCm/training-stack software artifact.
Retrieved 2026-06-15Public repositorymedium
Glassdoor TensorWave company profile
glassdoor.com
Employee-review proxy and small-sample org signal.
Retrieved 2026-06-15Glassdoor company datasetmedium
IDC AI infrastructure spending forecast
my.idc.com
AI infrastructure market size and timing.
Retrieved 2026-06-15Analyst press summaryhigh
McKinsey cost of compute article
mckinsey.com
AI data-center capex and power-demand context.
Retrieved 2026-06-15Consulting research articlemedium
AMD FY2024 Form 10-K
sec.gov
AMD Data Center revenue and export-control risk context.
Retrieved 2026-06-15Public-company SEC filinghigh
NVIDIA FY2025 Form 10-K
sec.gov
NVIDIA Data Center revenue benchmark.
Retrieved 2026-06-15Public-company SEC filinghigh
CoreWeave IPO prospectus
sec.gov
Neocloud revenue, margin, backlog, and depreciation benchmark.
Retrieved 2026-06-15Public-company SEC filinghigh
CoreWeave Q1 2026 filing
sec.gov
CoreWeave quarterly revenue, backlog, and depreciation context.
Retrieved 2026-06-15Public-company SEC filinghigh
AMD TensorWave eval page
amd.com
AMD-hosted TensorWave evaluation relationship.
Retrieved 2026-06-15AMD vendor-controlled pagemedium
AMD Advancing AI 2025 roadmap press release
ir.amd.com
AMD roadmap and ecosystem positioning.
Retrieved 2026-06-15AMD corporate press releasemedium
Federal Register advanced-computing export controls
federalregister.gov
Advanced-computing export-control baseline.
Retrieved 2026-06-15Federal Register rulehigh
Federal Register advanced-computing license review revision
federalregister.gov
Updated license-review context.
Retrieved 2026-06-15Federal Register rule revisionhigh
Modular TensorWave case study
modular.com
Modular partnership/case-study signal.
Retrieved 2026-06-15Partner-controlled case studymedium
Vultr homepage
vultr.com
AMD GPU cloud competitive pressure and pricing context.
Retrieved 2026-06-15Competitor-controlled homepagehigh
Crusoe homepage
crusoe.ai
AI cloud competitor positioning.
Retrieved 2026-06-15Competitor-controlled homepagehigh
Azure Virtual Machines
azure.microsoft.com
Hyperscaler GPU and virtual-machine alternative.
Retrieved 2026-06-15Incumbent-controlled product pagehigh
AWS EC2 P5 GPU page
aws.amazon.com
Hyperscaler GPU alternative.
Retrieved 2026-06-15Incumbent-controlled product pagehigh
Google Cloud GPU docs
cloud.google.com
Hyperscaler GPU alternative.
Retrieved 2026-06-15Incumbent-controlled documentationhigh
Oracle Cloud GPU page
oracle.com
Hyperscaler GPU alternative.
Retrieved 2026-06-15Incumbent-controlled product pagehigh
NVIDIA Data Center page
nvidia.com
Incumbent silicon ecosystem context.
Retrieved 2026-06-15Incumbent-controlled product pagehigh
CoreWeave homepage
coreweave.com
Direct AI infrastructure competitor.
Retrieved 2026-06-15Competitor-controlled homepagehigh
Lambda homepage
lambda.ai
GPU cloud competitor.
Retrieved 2026-06-15Competitor-controlled homepagehigh
Lambda instances page
lambda.ai
GPU instance product context.
Retrieved 2026-06-15Competitor-controlled product pagehigh
Runpod homepage
runpod.io
GPU cloud competitor.
Retrieved 2026-06-15Competitor-controlled homepagehigh
Runpod pricing
runpod.io
GPU pricing benchmark.
Retrieved 2026-06-15Competitor-controlled pricing pagehigh
Nebius homepage
nebius.com
AI infrastructure competitor.
Retrieved 2026-06-15Competitor-controlled homepagehigh
Together AI homepage
together.ai
AI infrastructure and model-platform competitor.
Retrieved 2026-06-15Competitor-controlled homepagehigh
Nscale homepage
nscale.com
AI infrastructure competitor.
Retrieved 2026-06-15Competitor-controlled homepagehigh
Paperspace homepage
paperspace.com
GPU cloud substitute.
Retrieved 2026-06-15Competitor-controlled homepagehigh
Voltage Park homepage
voltagepark.com
GPU infrastructure substitute.
Retrieved 2026-06-15Competitor-controlled homepagehigh
Vast.ai marketplace homepage
vast.ai
GPU marketplace substitute.
Retrieved 2026-06-15Competitor-controlled homepagehigh

Access limitations: This research did not include TensorWave's private data room, customer contracts, utilization exports, billing records, lender documents, AMD supply agreements, security reports, cap table, board materials, or founder interview. Customer status for Meta, Luma, Moreh, Modular, and AMD could not be resolved into paid production use from public sources, and exact category search results were treated as visibility signals rather than market share. Some registry, social, and job-board checks were empty, blocked, or snippet-grade, so they constrain the report's ability to verify Series B filing status, founder profile depth, and hiring momentum. The absence of public backlog, unit economics, security-report bodies, and Series B Form D evidence is a public-record limitation, not proof those materials do not exist privately.

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. TensorWave 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#