Naveen Rao

Naveen Rao is an American computer architect and neuroscientist, founder and chief executive officer of Unconventional AI, and the prior founder of MosaicML (acquired by Databricks for $1.3 billion) and Nervana Systems (acquired by Intel).
Naveen Rao

Bio

Naveen Rao is an American computer architect and neuroscientist, the founder and chief executive officer of Unconventional AI, the brain-inspired AI hardware company he established in 2025 with a $475 million seed round at a $4.5 billion post-money valuation. He previously founded Nervana Systems (acquired by Intel for approximately $350 million in 2016) and MosaicML (acquired by Databricks for approximately $1.3 billion in 2023). As of May 2026, Rao leads Unconventional AI in San Francisco, with the December 2025 seed round one of the largest seed financings in AI hardware history and a reported planned cumulative round of up to $1 billion.

At a glance

  • Education: Bachelor of Science in electrical engineering and computer science, Duke University (1997); graduate study in electrical engineering at Stanford University (1997 to 1998); PhD in neuroscience, Brown University (2011).
  • Current role: Founder and Chief Executive Officer of Unconventional AI, 2025 to present.
  • Key contributions: founder of Nervana Systems (2014, acquired by Intel for approximately $350 million in 2016); founder of MosaicML (2021, acquired by Databricks for approximately $1.3 billion in 2023); founder of Unconventional AI (2025); developed the MPT open-weights model line and the MosaicML training-and-inference infrastructure during the MosaicML period; developed the Nervana Neural Network Processor product line at Intel.
  • Recognition: Square Inc. board director; scientific advisor at the Allen Institute for AI; long-running speaker on the AI-hardware-and-training infrastructure circuit.
  • X / Twitter: @NaveenGRao
  • LinkedIn: naveen-rao-bba5b01

Origins

Rao earned a Bachelor of Science in electrical engineering and computer science from Duke University in 1997. He continued at Stanford University from 1997 to 1998 for graduate study in electrical engineering before transitioning to industrial chip-design roles at companies including Sun Microsystems, CALY Networks, and Kealia Inc. He returned to academia for doctoral study, completing the PhD in neuroscience at Brown University in 2011 with research on neural-circuit dynamics and computational neuroscience.

The hybrid background, combining systems-level chip-design experience with computational-neuroscience research, has been the structural feature distinguishing Rao's subsequent AI ventures. The thesis across both Nervana and Unconventional AI has been that biology-derived computational principles, applied at the silicon-architecture level rather than the software-abstraction level, can produce energy-efficiency and capability gains beyond what conventional digital scaling delivers.

Career

Rao's professional career organizes around three sequential AI ventures, with researcher and corporate-executive roles at Qualcomm and Intel between them.

He joined Qualcomm as a researcher from 2012 to 2014, focused on neuromorphic computing and neural computation. The Qualcomm period preceded the founding of Nervana Systems in 2014, his first AI-hardware startup. Nervana developed deep-learning compute architectures for AI training, with the Nervana Neural Network Processor as the central product. Intel acquired Nervana in 2016 for approximately $350 million, with Rao continuing at Intel as Corporate Vice President and General Manager of the Artificial Intelligence Products Group from May 2017 through March 2020. The acquisition seeded Intel's Nervana Neural Network Processor product line, though the product was eventually discontinued in 2020.

Rao founded MosaicML in 2021, his second AI-infrastructure venture. MosaicML built training-and-inference infrastructure that allowed enterprises to train custom large-language models efficiently using their own data, with the Composer training library and the MPT open-weights model line as the principal technical outputs. MosaicML raised $37 million in venture capital and reached a $222 million valuation in October 2021. Databricks acquired MosaicML in July 2023 for approximately $1.3 billion. Rao continued at Databricks as Vice President of Generative AI and head of AI through 2025, with the MosaicML team and infrastructure becoming the foundation of Databricks's Mosaic AI platform.

Rao founded Unconventional AI in 2025 in San Francisco, his third and most ambitious venture in AI hardware. The thesis returns to the neuromorphic-and-analog-compute direction Rao first explored at Qualcomm, with the strategic premise that running neural networks directly on the nonlinear physics of silicon, rather than simulating them through digital abstractions, can recover a meaningful share of the biological-brain energy advantage. The December 2025 seed round of $475 million at a $4.5 billion post-money valuation, co-led by Andreessen Horowitz and Lightspeed Venture Partners with Lux Capital, DCVC, and Jeff Bezos participating, was one of the largest seed financings in AI-hardware history. Rao contributed $10 million of his own capital. Industry coverage has reported the round as the first installment toward a planned cumulative round of up to $1 billion.

Affiliations

  • Sun Microsystems, CALY Networks, Kealia Inc.: Engineering roles, late 1990s to 2000s.
  • Qualcomm: Researcher on neuromorphic machines and neural computation, 2012 to 2014.
  • Nervana Systems: Founder and Chief Executive Officer, 2014 to 2016 (acquired by Intel).
  • Intel: Corporate Vice President and General Manager, Artificial Intelligence Products Group, May 2017 to March 2020.
  • MosaicML: Founder and Chief Executive Officer, 2021 to 2023 (acquired by Databricks).
  • Databricks: Vice President of Generative AI and head of AI, 2023 to 2025.
  • Unconventional AI: Founder and Chief Executive Officer, 2025 to present.

Notable contributions

Rao's published record runs through three sequential AI ventures, with technical output spanning AI-hardware architecture, large-model training infrastructure, and open-weights model releases.

  • Nervana Systems (2014). Founded the AI-hardware startup; Nervana Neural Network Processor product line; acquired by Intel for approximately $350 million in 2016.
  • Nervana Neural Network Processor (Intel, 2017 to 2020). Led product development of Intel's first dedicated AI training chip, eventually discontinued.
  • MosaicML (2021). Founded the training-infrastructure-and-models company; Composer training library; MPT open-weights model line; acquired by Databricks for approximately $1.3 billion in 2023.
  • Unconventional AI (2025). Founded the brain-inspired AI hardware company; $475 million December 2025 seed round at $4.5 billion post-money valuation; planned cumulative round of up to $1 billion.
  • Public commentary. Long-running speaker and writer on AI-hardware-and-training infrastructure, datacenter-power economics, and the architectural-bet diversity in the AI compute supply chain.

Investments and boards

  • Unconventional AI (AI hardware): Founder and Chief Executive Officer, 2025 to present. Privately held; $4.5 billion post-money valuation following the December 2025 seed.
  • Square Inc. (now Block): Board director.
  • Allen Institute for AI: Scientific advisor.

The cap-table participation pattern across Rao's three ventures has been characterized in industry coverage as one of his structural advantages in capital formation. The combination of two prior successful AI-hardware exits and the operating record at Intel and Databricks supports founder-led capital formation at frontier-lab-scale round size at the pre-product stage.

Network

Rao's senior research-and-engineering relationships span the prior MosaicML, Nervana, and broader AI-hardware ecosystems. The Unconventional AI founding team draws on the prior MosaicML training-infrastructure team and the prior Nervana AI-chip design team, with additional academic neuromorphic-computing recruitment. The Andreessen Horowitz, Lightspeed, Lux Capital, DCVC, and Jeff Bezos investor cohort overlaps with the broader AI-and-deeptech investor base that has tracked Rao across the three ventures. Industry coverage of Unconventional AI's capital formation has consistently cited Rao's prior track record as the principal credibility signal at the pre-product stage.

Position in the field

As of May 2026, Rao is one of the principal serial founders in AI hardware, with two prior successful exits cumulating to approximately $1.65 billion in transaction value and a third venture at frontier-lab-scale capitalization. The closest peer comparators within the AI-hardware founder cohort are Andrew Feldman at Cerebras and the broader founder-operator cohort at AI-chip startups; Rao differs in the architectural commitment to neuromorphic and analog compute rather than digital architectural innovation, and in the prior software-and-training-infrastructure experience from MosaicML.

The Unconventional AI thesis represents one of the principal architectural bets against incumbent GPU-based AI compute, alongside Cerebras's wafer-scale architecture, Groq's deterministic-pipeline architecture, Tenstorrent's RISC-V-based architecture, and Etched's transformer-specialized ASIC. The diversification of architectural directions, with multiple distinct approaches receiving meaningful funding through 2025 to 2026, reflects the structural shift in datacenter-power economics that has made energy efficiency a first-order constraint on AI scaling.

Outlook

Open questions over the next 6 to 18 months:

  • First architectural disclosures. Pace of architectural-disclosure milestones across 2026 to 2027 as Unconventional AI moves from pre-product seed-stage development toward first prototype demonstrations.
  • Subsequent fundraising. Progression of any subsequent funding rounds toward the reported planned cumulative $1 billion at a $5 billion-class valuation.
  • Programmability of the analog-and-neuromorphic substrate. Translation of the architectural thesis into substrates that support diverse AI workloads.
  • Competitive dynamics. Direct competition with established GPU incumbents (NVIDIA, AMD) and peer AI-chip startups across power-efficient AI compute.
  • Senior-talent recruitment. Continued hiring leveraging the prior MosaicML and Nervana teams plus academic neuromorphic-computing research.

Sources

About the author
Nextomoro

Nextomoro

nextomoro tracks progress for AI research labs, models, and what's next.

AI Research Lab Intelligence

nextomoro tracks progress for AI research labs, models, and what's next.

AI Research Lab Intelligence

Great! You’ve successfully signed up.

Welcome back! You've successfully signed in.

You've successfully subscribed to AI Research Lab Intelligence.

Success! Check your email for magic link to sign-in.

Success! Your billing info has been updated.

Your billing was not updated.