Synthszr Charts — die großen AI-Marken im Wettkampf ums Podium

RoboTTT

Nvidia · od 15. Juli 2026 (Paper/Projektseite veröffentlicht) · 2× · naposledy 15. 9. 2026

Momentum

RoboTTT is a research framework from NVIDIA (with Stanford and UT Austin) that integrates Test-Time Training (TTT) into vision-language-action robot policies, scaling visuomotor context to 8,000 timesteps — three orders of magnitude beyond prior state-of-the-art policies — without increasing inference latency. It is instantiated on the pretrained GR00T N1.7 model (3-billion parameters) by adding TTT layers to its Diffusion Transformer action head. On challenging real-robot manipulation tasks (bimanual assembly on a YAM setup), RoboTTT improves overall performance by 87% over a single-step baseline and is the only model to fully complete a five-minute, ten-stage assembly task. It is a research project with a published paper and project page, not a commercial product with pricing.

Vývoj momenta

02.08.31.10.

Vlastnosti

Deployment Model	Research prototype with published paper/project page and demo videos; no commercial product deployment documented
Use Case Scope	Bimanual robot manipulation/assembly (e.g. Circuit Assembly, Pup Go Car, Gear Bot) on YAM robot platform with 4 RGB cameras
Integrations	Instantiated on GR00T N1.7; training uses sequence action forcing + truncated backpropagation through time on 16 NVIDIA GB200 GPUs
Platform	Built on NVIDIA GR00T N1.7 (Vision-Language-Action model, 3B parameters), TTT layers added to its 16 DiT layers in the Diffusion Transformer action head
Release Date	July 15, 2026 (arXiv paper and project page)

Další produkty v této kategorii: Robotika a Embodied AI

Zdroje (2)

Company Analysis: Nvidia

As of 24/10/2026

HOLDSynthszr Vote

Data over the last ~90 days supports exceptional operating momentum (Q2 FY2027 revenue $96.2B, +106% YoY; ~75% GAAP gross margin). (investor.nvidia.com) However, the investment case is now more balanced by (1) explicit exclusion of China data center compute revenue from forward outlook (visibility/TAM constraint) and (2) valuation that remains elevated on trailing earnings (~26–27x) even if forward multiples are lower. (investor.nvidia.com) For EUR-based investors without a current price provided, a data-driven stance is to HOLD: maintain exposure to strong fundamentals while managing concentration and monitoring margin trajectory, working-capital terms (DSO), and policy developments that can alter guidance and demand patterns. (investor.nvidia.com)

Summary

NVIDIA is a fabless semiconductor and platform company whose core business is designing accelerated computing hardware (GPUs, networking, and increasingly CPU/SoC components) and monetizing it through systems, software, and developer ecosystems. The company’s core competency is end-to-end AI infrastructure: tightly coupled silicon (e.g., Blackwell generation), high-speed interconnect (networking), and a mature software stack that lowers total cost of ownership and time-to-deployment for hyperscalers and enterprises. Market position remains dominant in data-center AI acceleration, supported by scale, a broad partner ecosystem, and high switching costs created by software tooling and optimized libraries. Recent results indicate demand remains strong and supply-constrained in the near term. In fiscal Q2 2027 (quarter ended July 26, 2026), NVIDIA reported revenue of $96.2B (+18% QoQ, +106% YoY) with GAAP gross margin of 75.0% (non-GAAP 74.9%). (investor.nvidia.com) Management commentary also highlighted working-capital dynamics (DSO rising to ~60 days) tied to extended payment terms for large, investment-grade customers shipping over multiple quarters. (investor.nvidia.com) A key recent development is the company’s forward outlook explicitly excluding China data center compute revenue due to geopolitical uncertainty, which is material for modeling regional mix and risk. (investor.nvidia.com) The Q2 10-Q also notes that shipments of certain data center products to China were less than 1% of Data Center revenue in the quarter, indicating limited current-quarter exposure but continued policy sensitivity. (investor.nvidia.com) On valuation, market data sources in EUR indicate a trailing P/E around ~26–27x and forward P/E around ~17–18x (ratios are currency-invariant; the EUR listing reflects the same underlying economics). (stockanalysis.com) Profitability remains exceptional for a large-cap semiconductor company, with gross margin at ~75% in the latest quarter. (investor.nvidia.com) Short- to medium-term outlook is primarily driven by (1) sustained AI infrastructure capex by hyperscalers and sovereign/enterprise deployments, (2) product mix and memory/supply-chain costs affecting gross margin trajectory, and (3) export-control/geopolitical constraints affecting addressable market and visibility. The company also increased its quarterly dividend earlier in 2026, signaling confidence in cash generation and shareholder returns. (sec.gov)

Key Takeaways

Q2 FY2027 results show continued hyper-growth: $96.2B revenue (+106% YoY) with ~75% GAAP gross margin, indicating strong pricing power and scale benefits. (investor.nvidia.com)
Management’s forward outlook excludes China data center compute revenue due to geopolitical uncertainty, a meaningful constraint on incremental growth and visibility. (investor.nvidia.com)
Working-capital terms are loosening: DSO increased to ~60 days, reflecting extended payment terms for large customers and potentially higher receivables risk/volatility. (investor.nvidia.com)
Valuation in EUR terms remains elevated on trailing earnings (~26–27x) but materially lower on forward earnings (~17–18x), implying the market is underwriting continued earnings expansion. (stockanalysis.com)
Capital return profile improved: the quarterly dividend was increased to $0.25/share (from $0.01) earlier in 2026, supporting total-return appeal beyond pure multiple expansion. (sec.gov)

Action Ideas

BUY

12–24 month accumulation thesis: NVIDIA’s latest quarter demonstrates sustained demand for AI compute infrastructure with very high gross margins (~75%) and rapid YoY growth. The forward P/E (~17–18x) suggests the market expects continued earnings growth; if execution remains consistent and supply ramps without major margin compression, risk-adjusted returns can remain attractive even without further multiple expansion. Focus on position sizing and staged entries given volatility and policy risk.

Horizon: 18 mo.

HOLD

6–12 month risk-managed hold: For investors already positioned, the data supports holding through the next earnings cycle given strong reported growth and profitability. However, with a still-high trailing P/E (~26–27x), near-term upside may depend on continued upside to guidance and stable margins. Maintain exposure but consider rebalancing if position concentration is high relative to portfolio risk limits.

Horizon: 9 mo.

SELL

3–6 month tactical trim thesis: If an investor’s objective is to reduce drawdown risk, trimming can be justified because (a) the stock embeds strong expectations (high trailing multiple), and (b) management’s explicit exclusion of China data center compute revenue from forward outlook increases uncertainty around incremental growth sources. This is most appropriate for portfolios with high NVDA concentration or strict volatility constraints.

Horizon: 6 mo.

Contrarian Insights

• Despite headline growth, the rise in DSO (~60 days) and extended payment terms can be interpreted as a less-discussed form of demand financing and could increase cash-flow volatility versus earnings, especially if macro conditions tighten. (investor.nvidia.com)
• Consensus often frames China exposure as either large or binary; NVIDIA’s filings indicate current-quarter China-related data center shipments were <1% of Data Center revenue, suggesting near-term revenue dependence may be smaller than many narratives imply—while the forward-looking exclusion still matters for incremental growth. (investor.nvidia.com)