Enter password to view Website Audit

Analysis

Website

Physical Intelligence (π)

Analysis

Website

Physical Intelligence (π)

Analysis

Website

Physical Intelligence (π)

Summary

About

Company

Physical Intelligence (π)

Overall Score of Website

19

Analysed on 2026-03-24

Description

Foundation model company for robotics. Founded 2024 by researchers from DeepMind, Google Brain, Stanford, and Carnegie Mellon. Mission: build general-purpose AI foundation models that give robots the ability to act in any environment. Products: π0 (3B parameter VLA model, February 2025, trained on 10,000+ hours/7 robot types/68 tasks, open-sourced); π0 FAST (November 2025, autoregressive VLA model with FAST tokenizer); business model: $300/robot/month subscription. Hardware-agnostic — works across Woven by Toyota, Zipline, Boston Dynamics. Named 'Android for robots.' Demonstrated tasks: kitchen cleaning, laundry folding, bed making, package packing. Funding: $1.1B total ($70M seed March 2024; $400M Series A November 2024 at $2B valuation; $600M Series B November 2025, led by CapitalG/Alphabet at $5.6B valuation; investors: Jeff Bezos, OpenAI, Thrive Capital, Lux Capital, Bond, Redpoint, Sequoia, T. Rowe Price, NVIDIA NVentures, Index Ventures). Customers: Woven by Toyota, Zipline. Market context: 400M intelligent robots expected online in 4 years; $30B physical AI market.

Market

Robotics AI / Physical AI / Robot Foundation Models / Vision-Language-Action Models / Industrial Automation

Audience

Robotics engineers and R&D teams at industrial automation companies evaluating general-purpose foundation models; hardware manufacturers needing AI software that works across their robot embodiments; enterprise automation buyers evaluating $300/robot/month AI capabilities; robotics researchers fine-tuning on domain-specific tasks

HQ

San Francisco, CA, USA

Visualisation

Spider Chart

ContentContentContentContentContentContentStrategySEOContentFreshness

Content

5

Content

8

Content

10

Content

13

Content

16

Content

20

Strategy

24

SEO

27

Content

30

Freshness

33

Content

$600M Series B (November 2025) — $5.6B Valuation — CapitalG Lead — Not in Hero

Score

5

Severity

High

Finding

Bloomberg and Sacra confirm: '$600 million Series B in November 2025 led by CapitalG (Alphabet's growth equity fund) at a $5.6 billion valuation. Other investors: Lux Capital, Bond, Redpoint Ventures, Sequoia Capital, T. Rowe Price, and NVIDIA through NVentures.' Total funding: $1.1 billion. CapitalG leading and NVIDIA investing signals that two of the most important technology companies in the world believe Physical Intelligence is building the most important robotics infrastructure.

Recommendation

Feature the Series B: '$600M Series B (November 2025) · $5.6B valuation · Led by Alphabet's CapitalG · With NVIDIA, Sequoia, T. Rowe Price. Total $1.1B raised. The leading foundation model company for robotics. [About π →]'

Content

$600M Series B (November 2025) — $5.6B Valuation — CapitalG Lead — Not in Hero

Score

5

Severity

High

Finding

Bloomberg and Sacra confirm: '$600 million Series B in November 2025 led by CapitalG (Alphabet's growth equity fund) at a $5.6 billion valuation. Other investors: Lux Capital, Bond, Redpoint Ventures, Sequoia Capital, T. Rowe Price, and NVIDIA through NVentures.' Total funding: $1.1 billion. CapitalG leading and NVIDIA investing signals that two of the most important technology companies in the world believe Physical Intelligence is building the most important robotics infrastructure.

Recommendation

Feature the Series B: '$600M Series B (November 2025) · $5.6B valuation · Led by Alphabet's CapitalG · With NVIDIA, Sequoia, T. Rowe Price. Total $1.1B raised. The leading foundation model company for robotics. [About π →]'

Content

$600M Series B (November 2025) — $5.6B Valuation — CapitalG Lead — Not in Hero

Score

5

Severity

High

Finding

Bloomberg and Sacra confirm: '$600 million Series B in November 2025 led by CapitalG (Alphabet's growth equity fund) at a $5.6 billion valuation. Other investors: Lux Capital, Bond, Redpoint Ventures, Sequoia Capital, T. Rowe Price, and NVIDIA through NVentures.' Total funding: $1.1 billion. CapitalG leading and NVIDIA investing signals that two of the most important technology companies in the world believe Physical Intelligence is building the most important robotics infrastructure.

Recommendation

Feature the Series B: '$600M Series B (November 2025) · $5.6B valuation · Led by Alphabet's CapitalG · With NVIDIA, Sequoia, T. Rowe Price. Total $1.1B raised. The leading foundation model company for robotics. [About π →]'

Content

π0 — 3 Billion Parameter VLA Model — 10,000+ Hours of Training Data — 68 Tasks — Open-Sourced — Not in Hero

Score

8

Severity

High

Finding

Multiple sources confirm: 'π0 is a 3 billion parameter transformer model trained on over 10,000 hours of real-world robot data spanning 7 robot embodiments and 68 tasks. Physical Intelligence open-sourced the π0 code and weights.' Open-sourcing π0 while charging for production use is the canonical AI infrastructure go-to-market — it's how PyTorch, Linux, and Android built ecosystems.

Recommendation

Feature π0: 'π0: 3 billion parameters. 7 robot types. 68 tasks. 10,000+ hours of training data. Open-sourced. The foundation model for robotics, available for anyone to fine-tune — in production for enterprises at $300/robot/month. [Explore π0 →]'

Content

π0 — 3 Billion Parameter VLA Model — 10,000+ Hours of Training Data — 68 Tasks — Open-Sourced — Not in Hero

Score

8

Severity

High

Finding

Multiple sources confirm: 'π0 is a 3 billion parameter transformer model trained on over 10,000 hours of real-world robot data spanning 7 robot embodiments and 68 tasks. Physical Intelligence open-sourced the π0 code and weights.' Open-sourcing π0 while charging for production use is the canonical AI infrastructure go-to-market — it's how PyTorch, Linux, and Android built ecosystems.

Recommendation

Feature π0: 'π0: 3 billion parameters. 7 robot types. 68 tasks. 10,000+ hours of training data. Open-sourced. The foundation model for robotics, available for anyone to fine-tune — in production for enterprises at $300/robot/month. [Explore π0 →]'

Content

π0 — 3 Billion Parameter VLA Model — 10,000+ Hours of Training Data — 68 Tasks — Open-Sourced — Not in Hero

Score

8

Severity

High

Finding

Multiple sources confirm: 'π0 is a 3 billion parameter transformer model trained on over 10,000 hours of real-world robot data spanning 7 robot embodiments and 68 tasks. Physical Intelligence open-sourced the π0 code and weights.' Open-sourcing π0 while charging for production use is the canonical AI infrastructure go-to-market — it's how PyTorch, Linux, and Android built ecosystems.

Recommendation

Feature π0: 'π0: 3 billion parameters. 7 robot types. 68 tasks. 10,000+ hours of training data. Open-sourced. The foundation model for robotics, available for anyone to fine-tune — in production for enterprises at $300/robot/month. [Explore π0 →]'

Content

$300/Robot/Month Subscription — Hardware-Agnostic Model — Primary Business Model Not Communicated

Score

10

Severity

High

Finding

Sacra confirms: '$300 monthly subscription per connected robot, yielding recurring revenue that scales with fleet deployments.' This pricing model is extraordinarily elegant — as robot fleets grow, revenue grows proportionally, with near-zero marginal cost.

Recommendation

Feature the business model: 'Physical Intelligence: $300/robot/month. Not hardware. Not services. Software that works on the robots you already have. As your fleet grows, your AI gets smarter. [Pricing →]'

Content

$300/Robot/Month Subscription — Hardware-Agnostic Model — Primary Business Model Not Communicated

Score

10

Severity

High

Finding

Sacra confirms: '$300 monthly subscription per connected robot, yielding recurring revenue that scales with fleet deployments.' This pricing model is extraordinarily elegant — as robot fleets grow, revenue grows proportionally, with near-zero marginal cost.

Recommendation

Feature the business model: 'Physical Intelligence: $300/robot/month. Not hardware. Not services. Software that works on the robots you already have. As your fleet grows, your AI gets smarter. [Pricing →]'

Content

$300/Robot/Month Subscription — Hardware-Agnostic Model — Primary Business Model Not Communicated

Score

10

Severity

High

Finding

Sacra confirms: '$300 monthly subscription per connected robot, yielding recurring revenue that scales with fleet deployments.' This pricing model is extraordinarily elegant — as robot fleets grow, revenue grows proportionally, with near-zero marginal cost.

Recommendation

Feature the business model: 'Physical Intelligence: $300/robot/month. Not hardware. Not services. Software that works on the robots you already have. As your fleet grows, your AI gets smarter. [Pricing →]'

Content

Woven by Toyota + Zipline + Boston Dynamics — Named Enterprise Partners Not in Hero

Score

13

Severity

Medium

Finding

Crunchbase and SiliconANGLE confirm: 'Toyota Motor Co.'s mobility subsidiary Woven' and 'Zipline International.' Crunchbase also shows Boston Dynamics as a related company. Woven by Toyota is one of the most demanding autonomy programs in the world.

Recommendation

Feature the enterprise partners: '[Woven by Toyota] [Zipline] — two of the most demanding autonomy programs on earth trust Physical Intelligence to power their robots. The Android for robots, now powering the world's most sophisticated robotic systems. [See partners →]'

Content

Woven by Toyota + Zipline + Boston Dynamics — Named Enterprise Partners Not in Hero

Score

13

Severity

Medium

Finding

Crunchbase and SiliconANGLE confirm: 'Toyota Motor Co.'s mobility subsidiary Woven' and 'Zipline International.' Crunchbase also shows Boston Dynamics as a related company. Woven by Toyota is one of the most demanding autonomy programs in the world.

Recommendation

Feature the enterprise partners: '[Woven by Toyota] [Zipline] — two of the most demanding autonomy programs on earth trust Physical Intelligence to power their robots. The Android for robots, now powering the world's most sophisticated robotic systems. [See partners →]'

Content

Woven by Toyota + Zipline + Boston Dynamics — Named Enterprise Partners Not in Hero

Score

13

Severity

Medium

Finding

Crunchbase and SiliconANGLE confirm: 'Toyota Motor Co.'s mobility subsidiary Woven' and 'Zipline International.' Crunchbase also shows Boston Dynamics as a related company. Woven by Toyota is one of the most demanding autonomy programs in the world.

Recommendation

Feature the enterprise partners: '[Woven by Toyota] [Zipline] — two of the most demanding autonomy programs on earth trust Physical Intelligence to power their robots. The Android for robots, now powering the world's most sophisticated robotic systems. [See partners →]'

Content

400 Million Intelligent Robots Expected Online in 4 Years — Market Context Not in Hero

Score

16

Severity

Medium

Finding

SiliconANGLE confirms: 'more than 400 million intelligent robots will come online in the next four years, pushing the physical AI industry to eclipse $30 billion annually.' 400 million robots in 4 years is the market context that justifies $1.1 billion in funding.

Recommendation

Feature the market: '400 million intelligent robots expected online in the next 4 years. $30 billion physical AI market. Physical Intelligence provides the foundation models that will power the majority of them. [The market opportunity →]'

Content

400 Million Intelligent Robots Expected Online in 4 Years — Market Context Not in Hero

Score

16

Severity

Medium

Finding

SiliconANGLE confirms: 'more than 400 million intelligent robots will come online in the next four years, pushing the physical AI industry to eclipse $30 billion annually.' 400 million robots in 4 years is the market context that justifies $1.1 billion in funding.

Recommendation

Feature the market: '400 million intelligent robots expected online in the next 4 years. $30 billion physical AI market. Physical Intelligence provides the foundation models that will power the majority of them. [The market opportunity →]'

Content

400 Million Intelligent Robots Expected Online in 4 Years — Market Context Not in Hero

Score

16

Severity

Medium

Finding

SiliconANGLE confirms: 'more than 400 million intelligent robots will come online in the next four years, pushing the physical AI industry to eclipse $30 billion annually.' 400 million robots in 4 years is the market context that justifies $1.1 billion in funding.

Recommendation

Feature the market: '400 million intelligent robots expected online in the next 4 years. $30 billion physical AI market. Physical Intelligence provides the foundation models that will power the majority of them. [The market opportunity →]'

Content

π0 FAST (November 2025) — Autoregressive VLA Model — Not in Hero

Score

20

Severity

Medium

Finding

Multiple sources confirm: 'In November 2025, the firm introduced π0 FAST, an autoregressive vision-language-action model using the Flow matching Action Space Tokenizer.' π0 FAST is Physical Intelligence's newest model — and the basis for the generation of models coming after π0.

Recommendation

Feature π0 FAST: 'π0 FAST (November 2025): Our newest vision-language-action model — autoregressive, faster, and more precise than π0. The Flow matching Action Space Tokenizer enables real-time robot control at higher accuracy than any previous physical AI model. [Technical details →]'

Content

π0 FAST (November 2025) — Autoregressive VLA Model — Not in Hero

Score

20

Severity

Medium

Finding

Multiple sources confirm: 'In November 2025, the firm introduced π0 FAST, an autoregressive vision-language-action model using the Flow matching Action Space Tokenizer.' π0 FAST is Physical Intelligence's newest model — and the basis for the generation of models coming after π0.

Recommendation

Feature π0 FAST: 'π0 FAST (November 2025): Our newest vision-language-action model — autoregressive, faster, and more precise than π0. The Flow matching Action Space Tokenizer enables real-time robot control at higher accuracy than any previous physical AI model. [Technical details →]'

Content

π0 FAST (November 2025) — Autoregressive VLA Model — Not in Hero

Score

20

Severity

Medium

Finding

Multiple sources confirm: 'In November 2025, the firm introduced π0 FAST, an autoregressive vision-language-action model using the Flow matching Action Space Tokenizer.' π0 FAST is Physical Intelligence's newest model — and the basis for the generation of models coming after π0.

Recommendation

Feature π0 FAST: 'π0 FAST (November 2025): Our newest vision-language-action model — autoregressive, faster, and more precise than π0. The Flow matching Action Space Tokenizer enables real-time robot control at higher accuracy than any previous physical AI model. [Technical details →]'

Strategy

'Android for Robots' — Category Framing Not in Hero

Score

24

Severity

Medium

Finding

Sacra describes Physical Intelligence's strategy as: 'hardware-agnostic AI models that work across different robot embodiments, described as an Android for robots.' This framing — Android, not iOS — is the most powerful competitive positioning in the robotics AI market because it signals openness, ecosystem, and scale.

Recommendation

Feature the Android framing: 'Physical Intelligence: The Android for robots. We don't build hardware. We build the open foundation model that runs on all hardware — Woven, Zipline, Boston Dynamics, and the 400 million robots coming online in the next 4 years. [Why hardware-agnostic →]'

Strategy

'Android for Robots' — Category Framing Not in Hero

Score

24

Severity

Medium

Finding

Sacra describes Physical Intelligence's strategy as: 'hardware-agnostic AI models that work across different robot embodiments, described as an Android for robots.' This framing — Android, not iOS — is the most powerful competitive positioning in the robotics AI market because it signals openness, ecosystem, and scale.

Recommendation

Feature the Android framing: 'Physical Intelligence: The Android for robots. We don't build hardware. We build the open foundation model that runs on all hardware — Woven, Zipline, Boston Dynamics, and the 400 million robots coming online in the next 4 years. [Why hardware-agnostic →]'

Strategy

'Android for Robots' — Category Framing Not in Hero

Score

24

Severity

Medium

Finding

Sacra describes Physical Intelligence's strategy as: 'hardware-agnostic AI models that work across different robot embodiments, described as an Android for robots.' This framing — Android, not iOS — is the most powerful competitive positioning in the robotics AI market because it signals openness, ecosystem, and scale.

Recommendation

Feature the Android framing: 'Physical Intelligence: The Android for robots. We don't build hardware. We build the open foundation model that runs on all hardware — Woven, Zipline, Boston Dynamics, and the 400 million robots coming online in the next 4 years. [Why hardware-agnostic →]'

SEO

'Robot Foundation Model' / 'Physical Intelligence pi' / 'Robotics AI Platform' — Category Terms

Score

27

Severity

Low

Finding

Physical Intelligence's primary search terms: 'generalist robot AI model,' 'vision language action model robotics,' 'robot foundation model fine-tuning,' 'Physical Intelligence vs. Covariant vs. Skild AI.' These come from robotics engineers, autonomous vehicle teams, and industrial automation operators.

Recommendation

Create comparison content: physicalintelligence.ai/vs-covariant. 'Physical Intelligence vs. Covariant: Both are building general-purpose robot foundation models. Physical Intelligence differentiates with open-sourced π0 weights (enabling developer ecosystem formation), 10,000+ hours of diverse training data across 7 robot types, and $300/robot/month pricing that makes enterprise deployment economically rational.'

SEO

'Robot Foundation Model' / 'Physical Intelligence pi' / 'Robotics AI Platform' — Category Terms

Score

27

Severity

Low

Finding

Physical Intelligence's primary search terms: 'generalist robot AI model,' 'vision language action model robotics,' 'robot foundation model fine-tuning,' 'Physical Intelligence vs. Covariant vs. Skild AI.' These come from robotics engineers, autonomous vehicle teams, and industrial automation operators.

Recommendation

Create comparison content: physicalintelligence.ai/vs-covariant. 'Physical Intelligence vs. Covariant: Both are building general-purpose robot foundation models. Physical Intelligence differentiates with open-sourced π0 weights (enabling developer ecosystem formation), 10,000+ hours of diverse training data across 7 robot types, and $300/robot/month pricing that makes enterprise deployment economically rational.'

SEO

'Robot Foundation Model' / 'Physical Intelligence pi' / 'Robotics AI Platform' — Category Terms

Score

27

Severity

Low

Finding

Physical Intelligence's primary search terms: 'generalist robot AI model,' 'vision language action model robotics,' 'robot foundation model fine-tuning,' 'Physical Intelligence vs. Covariant vs. Skild AI.' These come from robotics engineers, autonomous vehicle teams, and industrial automation operators.

Recommendation

Create comparison content: physicalintelligence.ai/vs-covariant. 'Physical Intelligence vs. Covariant: Both are building general-purpose robot foundation models. Physical Intelligence differentiates with open-sourced π0 weights (enabling developer ecosystem formation), 10,000+ hours of diverse training data across 7 robot types, and $300/robot/month pricing that makes enterprise deployment economically rational.'

Content

Founders From DeepMind + Google Brain + CMU — Technical Pedigree Not in Hero

Score

30

Severity

Low

Finding

Multiple sources confirm Physical Intelligence was founded by researchers from DeepMind, Google Brain, Stanford, and Carnegie Mellon — the four institutions that produced the majority of the world's best robotics AI researchers.

Recommendation

Feature the founding team: 'Physical Intelligence was founded by researchers from DeepMind, Google Brain, Stanford, and Carnegie Mellon — the institutions that have produced most of the world's breakthrough robotics AI research. We didn't just hire AI researchers. We assembled the team that wrote the papers. [Meet the team →]'

Content

Founders From DeepMind + Google Brain + CMU — Technical Pedigree Not in Hero

Score

30

Severity

Low

Finding

Multiple sources confirm Physical Intelligence was founded by researchers from DeepMind, Google Brain, Stanford, and Carnegie Mellon — the four institutions that produced the majority of the world's best robotics AI researchers.

Recommendation

Feature the founding team: 'Physical Intelligence was founded by researchers from DeepMind, Google Brain, Stanford, and Carnegie Mellon — the institutions that have produced most of the world's breakthrough robotics AI research. We didn't just hire AI researchers. We assembled the team that wrote the papers. [Meet the team →]'

Content

Founders From DeepMind + Google Brain + CMU — Technical Pedigree Not in Hero

Score

30

Severity

Low

Finding

Multiple sources confirm Physical Intelligence was founded by researchers from DeepMind, Google Brain, Stanford, and Carnegie Mellon — the four institutions that produced the majority of the world's best robotics AI researchers.

Recommendation

Feature the founding team: 'Physical Intelligence was founded by researchers from DeepMind, Google Brain, Stanford, and Carnegie Mellon — the institutions that have produced most of the world's breakthrough robotics AI research. We didn't just hire AI researchers. We assembled the team that wrote the papers. [Meet the team →]'

Freshness

$600M Series B November 2025 + π0 FAST November 2025 — Both 4 Months Old — Not in Hero

Score

33

Severity

Low

Finding

The Series B ($600M, $5.6B valuation, November 2025) and π0 FAST launch (November 2025) are both 4 months old and should be the dominant homepage narrative.

Recommendation

Rebuild the hero: '$600M Series B (November 2025) · $5.6B valuation · π0 FAST launched · 400M robots expected online in 4 years. Physical Intelligence is building the foundation model layer for the robotic economy. [Start building →]'

Freshness

$600M Series B November 2025 + π0 FAST November 2025 — Both 4 Months Old — Not in Hero

Score

33

Severity

Low

Finding

The Series B ($600M, $5.6B valuation, November 2025) and π0 FAST launch (November 2025) are both 4 months old and should be the dominant homepage narrative.

Recommendation

Rebuild the hero: '$600M Series B (November 2025) · $5.6B valuation · π0 FAST launched · 400M robots expected online in 4 years. Physical Intelligence is building the foundation model layer for the robotic economy. [Start building →]'

Freshness

$600M Series B November 2025 + π0 FAST November 2025 — Both 4 Months Old — Not in Hero

Score

33

Severity

Low

Finding

The Series B ($600M, $5.6B valuation, November 2025) and π0 FAST launch (November 2025) are both 4 months old and should be the dominant homepage narrative.

Recommendation

Rebuild the hero: '$600M Series B (November 2025) · $5.6B valuation · π0 FAST launched · 400M robots expected online in 4 years. Physical Intelligence is building the foundation model layer for the robotic economy. [Start building →]'

Let's discuss how we can get Physical Intelligence (π)'s website to the next level

Let's discuss how we can get Physical Intelligence (π)'s website to the next level

Let's discuss how we can get Physical Intelligence (π)'s website to the next level