← back to rankingQS · QuantumScape Corporation
Auto - Parts · mkt cap $5.7B · calls: Q1 FY2026 vs Q4 FY2025
41.0 conviction · conf-adj 41
conf 2/10 partial
enthusiasm:21.0 · trend:8 · quantifies:0 · impact:0 · under_radar:5 · credibility:0 · business_impact:4 · disruption:0 · commitment:0 · confirmation:3
Enthusiasm latest 7 / prev 5 (rising)
Across the two calls, QS’s AI story splits into (1) internal ML/AI on the Eagle Line for metrology-driven quality/reliability and faster learning loops—most concrete in Q1 FY2026—and (2) AI data centers as a demand vertical for solid-state in-rack power, with no dollars, yields, or productivity metrics tied to AI itself. Enthusiasm rises quarter-over-quarter as management moves from a single CFO line on deploying ML/AI tools to CEO-level claims of “substantive progress” from integrated AI models, but credibility on business impact remains qualitative: no quantified revenue, cost, margin, or throughput attribution to AI.
GROUNDED NEXT-FY IMPACT vs CONSENSUS
Grounded on actual base — revenue $0.0B · net income $-0.4B · net margin n/a · diluted EPS -0.76
These are next-fiscal-year annual uplift estimates, not next-quarter numbers.
Aggregate next-FY est. rev uplift: % · next-FY EPS uplift: % · vs analysts: unclear · priced in: medium · confidence: 2/10
| Claim | Figure | Arithmetic | Next-FY Rev % | Next-FY EPS % |
|---|
$1M GPUs per rack (customer asset at risk)
other · soft | $1,000,000 GPUs per rack | Customer-side asset value, explicitly 'not QS AI ROI'. No QS revenue, capture rate, rack count, or ASP mapped to it. QS current revenue = $0 → rev_uplift_pct undefined; NI = -$435.05M (loss) → eps_uplift_pct N/A per loss-making guardrail. | | |
Fire 'millions in damage and downtime' (customer loss)
other · soft | 'millions' (unanchored) | Unbounded ($2M–$99M+) customer-loss figure with no QS wallet share, pricing, or units. Sizes the customer's pain point, not QS revenue. Unanchored — not mappable to QS FY26 P&L without inventing. | | |
800-volt DC data center power shift
other · soft | 800-volt DC designs | Qualitative market-architecture context; no $, kWh, units, or timing-to-revenue. No arithmetic path to QS rev or EPS. | | |
PowerCo 5 GWh non-auto manufacturing capacity (incl. data centers)
revenue · soft | 5 GWh capacity | Disclosed physical base: 5 GWh = 5,000,000 kWh (partner PowerCo capacity, not QS recognized revenue). No disclosed $/kWh, QS share %, or FY26 phasing anywhere in inputs → cannot form claim_$ for rev_uplift_pct=100×claim_$/$0 without inventing an ASP/allocation. Capacity ≠ revenue. Soft because a revenue base is not obtainable from inputs. | | |
Eagle Line data volume / learning / quality (manufacturing AI)
productivity · soft | none quantified | Internal operational AI/learning-cycle narrative; no $ opex save, yield %, or volume lift. rev_uplift_pct ≈ 0; no cost $ and loss-making NI → eps_uplift_pct N/A. Unanchored. | | |
QSE-5 ramp + AI data center samples (early-days add-on)
revenue · soft | none quantified | Management frames AI data centers as early-days sampling add-on; only QSE-5/Eagle-Line samples shipping in Q2 FY2026, no bookings/revenue $ or FY26 conversion rate. No ratable next-FY revenue to recognize. Unanchored. | | |
Assumptions: Earnings basis: consensus EPS (adjusted-style, ~-$0.67 FY26). Current FY25 base revenue = $0 and net income = -$435.05M (diluted EPS -$0.76 on 575.95M shares) → rev % denominators undefined/meaningless and EPS uplift % suppressed per loss-making guardrail. Tax 21% (default, irrelevant — loss-making). No incremental margin/phasing applied: no quantified QS $ revenue or cost saves disclosed. Bookings: none. Side = supplier for the data-center battery opportunity (QS sells INTO the AI buildout); Eagle Line manufacturing learning is the only adopter-side element and is unquantified. DISCLOSED-BASE: 5 GWh and $1M/rack appear in claims but yield no QS P&L $ without undisclosed ASP, QS share, and unit volume.
Top line: No quantifiable topline impact. The only dollar figures disclosed are customer-side ($1M GPU value/rack, 'millions' in fire damage) — they size the customer's pain point, not QS revenue. The lone QS-related capacity figure (PowerCo 5 GWh non-auto) carries no disclosed price, and QS FY25 revenue is $0, so no rev_uplift_pct is computable. Supplier narrative (5 GWh, 800V DC, GPU-rack safety) is TAM/market context, not FY26 recognized revenue; the AI data-center opportunity is explicitly 'early days' with only Q2 FY2026 sample shipments. Consensus models only ~$56k (FY26) and ~$28.5M (FY27) total revenue.
Bottom line: No quantifiable bottom-line impact. QS is deeply loss-making (net income -$435.05M, EPS -$0.76), so any EPS-uplift % is meaningless per guardrail and set to null. None of the AI claims are cost/productivity savings — they are forward revenue optionality with no anchored figure. Consensus FY26–FY27 net income improves -$402.7M → -$365.6M (EPS -$0.67/-$0.64) without any disclosed AI-driven bridge.
[impact n/m (all claims soft/unanchored); EPS uplift n/m (loss-making base)] Management supplied ZERO QS-revenue dollar figures for the AI opportunity — only a customer's $1M/rack GPU exposure, unbounded 'millions' in fire damage, and an unpriced 5 GWh PowerCo capacity number. Consensus already models QS as essentially pre-revenue (FY26 revenue $55,962, FY27 $28.5M) with deep losses and no separate AI line. The 'early-days add-on' does not quantify upside above that trajectory, and no analyst is underwriting a data-center battery ramp at scale. The gap vs claims is therefore unquantifiable rather than clearly mispriced; supplier-side AI optionality is real but unsized, so it is not clearly priced in.
MODEL CONSENSUS (impact)
partial
Both find no quantified QS AI revenue/EPS off a $0-revenue, loss-making base; reconciled on side, verdict, priced_in, and the PowerCo soft flag toward the better-justified/conservative reading.
Conflicts reconciled
- ai_revenue_side: X=both vs Y=supplier -> used both because an adopter-side element (Eagle Line manufacturing AI) is present, though unquantified
- vs_analyst_expectations: X=inline vs Y=unclear -> used unclear because zero quantified figures make the gap unquantifiable, not demonstrably inline
- priced_in: X=high vs Y=medium -> used medium because no analyst underwrites a data-center ramp, so the unsized optionality is not clearly priced in
- math[PowerCo 5 GWh].soft: X=false vs Y=true -> used true; capacity has no disclosed $/kWh so no revenue base is obtainable (matches the soft definition)
- math item count: X=6 vs Y=4 -> kept X's fuller set (added Eagle Line adopter + QSE-5 sampling rows) as more complete
| Field | Opus 4.8 | GPT-5.5 |
|---|
| Rev uplift % | – | – |
| EPS uplift % | – | – |
| Priced in | medium | – |
| vs analysts | unclear | – |
| Confidence | 2 | – |
| Top line | No quantifiable topline impact. The only dollar figures disclosed are customer-side ($1M GPU value/rack, 'millions' in fire damage) — these size the CUSTOMER'S pain point, not QS revenue. The lone QS capacity figure (PowerCo 5 GWh non-auto) carries no disclosed price, and QS's current revenue is $0, so no rev_uplift_pct is computable. The AI data-center opportunity is explicitly 'early days' with only sample shipments planned for Q2 FY2026. | – |
| Bottom line | No quantifiable bottom-line impact. QS is deeply loss-making (net income –$435.05M, diluted EPS –$0.76 on 575.95M shares), so any EPS-uplift % is meaningless per the guardrail and is set to null. None of the AI claims are cost/productivity savings; they are forward revenue optionality with no anchored figure. | – |
| Reasoning | Consensus already models QS as essentially pre-revenue: 2026 revenueAvg $55,962, 2027 $28.5M, with net income staying –$402.7M (2026) to –$365.6M (2027) and EPS –$0.67/–$0.64. No analyst is underwriting a data-center battery ramp at scale. Management supplied ZERO QS-revenue dollar figures for the AI opportunity — only a customer's $1M/rack GPU exposure and an unpriced 5 GWh capacity number. There is therefore no adopter-side figure that points above the consensus near-zero-revenue, deep-loss trajectory; the gap is unquantifiable, not clearly mispriced. Supplier-side AI optionality is real but unsized. | – |
Rows highlighted where the two models disagreed.
QUANTIFICATIONS
GPU rack asset value at risk (customer pain point, not QS AI ROI): $1 million GPUs per rack (not specified, topline)
“you absolutely cannot have a fire in racks with $1 million GPUs”
Data center fire/downtime damage (customer pain point, not QS AI ROI): millions in damage and downtime (not specified, topline)
“a fire in a GPU rack could easily cost millions in damage and downtime”
Data center power architecture shift (market context for AI-data-center batteries): 800-volt DC designs (transitioning now, topline)
“data centers are transitioning to 800-volt DC designs”
PowerCo non-automotive manufacturing capacity (commercialization path for new markets incl. data centers): 5 gigawatt hours of capacity (not specified, topline)
“Additionally, PowerCo has 5 gigawatt hours of capacity tenor markets outside the automotive.”
PAST (realized)
- Q1 FY2026 — Siva: "we have seen substantive progress on cell quality and reliability" from AI models on the Eagle Line.
- Q1 FY2026 — Siva: reliability improvements already "enabled by some of our new AI models" using metrology data.
- Q4 FY2025 — Kevin: "We're well along in deploying machine learning and AI tools" for development cycles and engineering productivity.
CURRENT (now)
- Q1 FY2026 — Siva: "We've been integrating advanced AI models into the Eagle Line" and improving uptime, throughput, control systems, and process stability.
- Q1 FY2026 — Siva: AI models on metrology are making quality determinations "faster, more accurately and more consistently than a human could possibly do," accelerating feedback/feed-forward loops.
- Q4 FY2025 — Kevin: ongoing deployment of ML/AI tools to accelerate development and improve engineering productivity.
- Q1 FY2026 — Siva: "increasing inbound customer interest" in AI data centers (and defense/aerospace), with samples from Eagle Line.
FORWARD (guidance)
- Q1 FY2026 — Siva: Eagle Line capacity to drive "a virtuous cycle of higher data volume, more rapid learning cycles and enhanced quality."
- Q1 FY2026 — Siva: Q2 ramp of QSE-5 production including demand from "new markets"; shipping Eagle Line samples for AI data center and other non-auto interest.
- Q1 FY2026 — Siva: AI data center opportunity framed as early-days add-on; PowerCo 5 GWh capacity for non-auto markets; ecosystem (Corning/Murata) to support ramps.
- Q4 FY2025 — Siva/Kevin: 2026 goal #3 to "expand into high-value markets" including data centers; Eagle Line as sampling currency for non-auto trials.
- Q4 FY2025 — Siva: ecosystem partners needed for "software and AI systems" support.
TRACK RECORD — PROMISE vs DELIVERY
—/100 (no quantified promises) no-quantified-promises 6 calls reviewed
Across six earnings calls (Q4 FY2024–Q1 FY2026), QuantumScape management never issued a quantified AI/ML commitment with a number and deadline—only late-2025/Q1-2026 status updates on deploying AI models on the Eagle Line for metrology and engineering productivity, with no prior measurable targets to judge against.
PRICED-IN (REFINED)
MEDIUMEst. revisions flat · Fwd P/E -11.6 · EV/Sales 0.0x
Analyst ratings show no buys and a stable Hold-heavy mix (holds ~7, sells easing from 4 to 2), with no fresh price-target updates (lastMonth/quarter counts zero; lastYear avg ~$10.22 vs ~$9.20 price). Forward consensus still embeds large losses (FY25–27 EPS roughly -$0.79 to -$0.64) and erratic tiny revenue, not a sharp upward revision path. Valuation is stretched on fundamentals (~$5.6B market cap, ~0 EV/Sales and negative EV/EBITDA; negative fwd P/E is not meaningful), but revisions are not rising—so per the framework this is mixed: optionality is partly paid for, yet estimates are not being ratcheted up to fully price an AI/battery upside case; segment data is absent so no revenue line maps the claim.
COVERAGE — ENTHUSIASM TRAJECTORY + CATALYSTS
2Q4 FY20242Q1 FY20252Q2 FY20252Q3 FY20253Q4 FY20256Q1 FY2026
AI enthusiasm across 6 calls — trend ↗ rising
Five calls cited factory automation only; Q1 FY2026 first tied advanced AI models on Eagle Line to metrology-driven cell quality gains.
RECENT AI CATALYSTS & NEWS
BUSINESS IMPACT - QUALITATIVE MATERIALITY
5/10 qualitative impact moderate medium-term · soft evidence
Where AI matters: Eagle Line manufacturing quality/ramp
Deployed metrology AI on Eagle Line plausibly accelerates yield, reliability, and learning loops on the core ramp, but management cites no yield, cost, throughput, or margin attribution and QS remains pre-revenue with no AI-linked P&L bridge.
Caveats: AI data-center TAM is supplier-side narrative with no bookings, ASP, or FY26 revenue tied to AI; Manufacturing AI benefits are qualitative with no disclosed yield or unit-economics proof; Rising AI enthusiasm may outrun analyst estimates still modeling ~$0–$28M revenue and deep losses; Competitors could use AI/ML to accelerate conventional or alternative battery development, compressing differentiation over time
AI DISRUPTION / CANNIBALIZATION RISK tailwind · 2/10
QS monetizes physical solid-state cells and manufacturing scale-up, not billable knowledge work or content; GenAI does not automate away electrochemistry or in-rack energy storage, and the main AI-data-center angle is demand optionality rather than cannibalization of what they sell.
OPTIONS / MARKET STRUCTURE
option liquidity: fair
proxy inputs — dollar-ADV $177M · beta 2.584 · px $9.20
source: proxy (no options chain on FMP)
FMP /stable/ exposes no options-chain endpoint on this key, so ATM IV, bid-ask spread and open interest are unavailable. Liquidity below is a PROXY from dollar-ADV, beta and price level (a stand-in for option depth), not measured option-market data.
CONFIRMATION — INSIDERS · 13F · LANGUAGE
Mixed — insiders selling, institutions adding, management language 6/10 measured.
INSIDERS selling 27 open-market sell(s) vs 0 buy(s) — net distribution
INSTITUTIONS (13F) adding as of 2026-03-31: 59 new / 90 closed positions; 200 increased / 126 reduced; institutional ownership +4.08pp; -37 net 13F holders
MGMT LANGUAGE 6/10 measured Internal AI on Eagle Line is real and results-oriented; most other AI talk is data-center demand, hedged or exploratory.
commit “We've been integrating advanced AI models into the Eagle Line, and we have seen substantive progress on cell quality and reliability.”
commit “The improvements we have made in reliability are enabled by some of our new AI models that take data from our metrology”
hedge “We believe that the increased production capacity at the Eagle Line will help drive a virtuous cycle of higher data volume, more rapid learning cycles”
VERBATIM AI QUOTES
“We've been integrating advanced AI models into the Eagle Line, and we have seen substantive progress on cell quality and reliability.”
— Siva Sivaram, Q1 FY2026
“We believe that the increased production capacity at the Eagle Line will help drive a virtuous cycle of higher data volume, more rapid learning cycles and enhanced quality.”
— Siva Sivaram, Q1 FY2026
“The improvements we have made in reliability are enabled by some of our new AI models that take data from our metrology make determinations of the quality faster, more accurately and more consistently than a human could possibly do.”
— Siva Sivaram, Q1 FY2026
“This enables an accelerated feedback and feed-forward loop which drives the continuous improvement cycle faster.”
— Siva Sivaram, Q1 FY2026
“We believe our high-performance solid-state design has compelling attributes to address the evolving energy storage needs of AI data centers, where conventional lithium-ion technology faces safety and performance limitations.”
— Siva Sivaram, Q1 FY2026
“Driven by massive compute demand, data centers are transitioning to 800-volt DC designs, and adopting power systems architecture and technology from the electric vehicle industry.”
— Siva Sivaram, Q1 FY2026
“In-rack energy storage and power delivery is a large and fast-growing market, and the higher energy density of our battery technology can enable increased compute density for AI factories.”
— Siva Sivaram, Q1 FY2026
“The speed of change and growth in the AI data center market is breathtaking.”
— Siva Sivaram, Q1 FY2026
“Sam, we think the opportunity for our technology in AI data centers is obvious and compelling.”
— Siva Sivaram, Q1 FY2026
“And safety really matters for a data center, where operating temperatures are higher and a fire in a GPU rack could easily cost millions in damage and downtime.”
— Siva Sivaram, Q1 FY2026
“provide power smoothing for these AI workloads, which from the battery's point of view is almost like being on a racetrack.”
— Siva Sivaram, Q1 FY2026
“And in all of these AI data centers and the AI factories, what you need is to be able to maximize the compute density.”
— Siva Sivaram, Q1 FY2026
“we'll be shipping samples from Eagle Line to meet the increasing inbound customer interest.”
— Siva Sivaram, Q1 FY2026
“We're well along in deploying machine learning and AI tools to accelerate development cycles and improve engineering productivity.”
— Kevin Hettrich, Q4 FY2025
“For example, in a data center, you have high ambient temperatures, but you absolutely cannot have a fire in racks with $1 million GPUs.”
— Siva Sivaram, Q4 FY2025
“AI demonstrators -- data center, safety.”
— Kevin Hettrich, Q4 FY2025
“whether it be in things like software and AI systems, there are places where we need help.”
— Siva Sivaram, Q4 FY2025
ANALYST QUESTIONS ON AI
Q (Q1 FY2026, Sam Kamara (scripted investor Q&A)): On the new markets you've described, what makes the opportunity in AI data centers and defense interesting?
A: Siva: AI data centers are "obvious and compelling," early but a strong add-on to automotive; requirements (energy density, power/charge-discharge for workload smoothing "almost like being on a racetrack," safety near GPUs) fit QS; differentiated tech enables "last-meter" power; samples from Eagle Line for rising inbound interest; board/adviser hires (Niebergall, Maybury) support defense/aerospace/space alongside data centers.
Q (Q1 FY2026, Yan Dong (Deutsche Bank)): On expansion to new markets—can you talk about potential investments, time frame to launch a product for data center use, and whether substantial technology/product changes are needed vs. an easy transfer?
A: Siva: Most automotive learning transfers; data centers moving to 800-volt with similar density/power/cycle needs plus safety/no-compromise positioning for "last-meter" power and compute density—"natural transition" from automotive to data center. Kevin: Platform strategy serves multiple markets; incremental tailoring; 2026 plan and reiterated EBITDA/CapEx/billings guidance already contemplate goal #3 (high-value markets).
Q (Q1 FY2026, Laisha Zaack Carrillo (HSBC)): How does going into new markets (e.g., data centers) change automotive goal timing, human capital, and resourcing—either/or or workforce expansion?
A: Siva: Automotive remains core focus; new markets added to portfolio, not a diversion; data center and related markets are good product fits; resourcing was sized in the annual operating plan—"not an either or."
Q (Q4 FY2025, Sam Kamara (scripted investor Q&A)): You've highlighted growing interest beyond automotive. How are you thinking about those opportunities while maintaining focus on automotive commercialization?
A: Siva: Automotive remains core and largest; electrification continues; autonomous-vehicle fleets strengthen EV economics; QS cell traits (safety, temperature range, power+energy) valuable elsewhere—e.g., data centers with high ambient temps and fire risk near expensive GPUs, drones needing density and discharge power; versatile cathode chemistries; logical to pursue fast-growing emerging applications.
Q (Q4 FY2025, Yan Dong (Deutsche Bank)): Among verticals (data centers, robotics, aviation, consumer electronics, etc.), where is the technology most suitable, and why is lithium-metal competitive vs. LFP in stationary storage?
A: Siva: Ceramic-separator "no-compromise" cell (energy, power, safety, cycle life, cost profile) maps to each market's needs; automotive still largest/longest-cycle; Eagle Line sampling enables parallel market trials. Kevin: Consumer electronics values volumetric density; "AI demonstrators -- data center, safety"; drones/aviation value gravimetric savings; grid load-shifting values cost per round-trip—management sequences parallel pursuits under 2026 goal #3.