Positron
Memory-first AI inference silicon for LLMs
Technical Product Manager - AI Inference & Software
+ Equity
About the job
- At least 10 years of experience with a strong mix of technical skills involving machine learning systems, inference infrastructure, or serving-stack engineering, including direct ownership of performance or architecture trade-offs.
- Hands-on knowledge of transformer internals at the operator level, including attention variants, mixture-of-experts routing, key-value cache mechanics, and quantization formats with their hardware implications.
- Experience with production inference serving at scale, including multi-tenancy, latency service-level agreements such as time to first token and time per output token, batching, scheduling, disaggregated serving, key-value cache management, and observability.
- Knowledge of the open-source inference ecosystem, including runtimes such as vLLM and SGLang, kernels, model ingestion, and how models are released, quantized, and adopted.
- Ability to perform model and systems performance analysis, including utilization reasoning, tokens-per-dollar and tokens-per-watt calculations, and benchmark design, and to build and defend the underlying math personally.
- Experience with competitive landscaping and analysis of inference providers and serving stacks at a model laboratory, inference application programming interface provider, or artificial intelligence hardware company.
- At least 10 years of experience with a strong mix of personal and analytical skills, including the ability to span model-architecture details through fleet-scale serving systems and stay current with the model and inference landscape.
- Ability to communicate effectively, navigate uncertainty, and drive the socialization of ideas and decisions.
- Ability to operate as the technically grounded representative in go-to-market discussions and the market-aware representative in engineering discussions.
- Ability to own decision history by documenting and explaining why technical choices were made with executive leadership.
- Demonstrated daily, hands-on use of agentic artificial intelligence in technical work, including building and running agent workflows for research, analysis, and drafting requirements, with the judgment to verify and own agent-produced work.
- Candidates must currently be authorized to work in the United States.
- Lead all aspects of artificial intelligence inference and software technical product planning.
- Write requirements for the inference software stack covering model coverage, numerics, inference-engine features and modes, serving-stack features and modes, and managed-service capabilities.
- Track the model frontier and convert changes in models and inference systems into requirements before they become customer escalations.
- Work with engineering and go-to-market teams to communicate a roadmap.
- Work with go-to-market teams to understand customer and market needs.
- Define and manage the software product lifecycle, including versions and release trains, feature modes, the model catalog, and deprecation policy.
- Create competitive briefings.
- Work with ecosystem partners, including model laboratories, open-source runtimes, and serving and orchestration partners, to understand their technology roadmaps.
- Create scope and feasibility frameworks that convert model and inference-system innovations into tangible engineering requests.
- Ensure software products are well documented.
- Streamline and automate product planning processes using agentic artificial intelligence.
About the company
Positron develops memory-first AI inference silicon and systems to scale large-language-model serving. Its Atlas systems rely on commodity LPDDR5X memory rather than expensive HBM/CoWoS to boost bandwidth and capacity, enabling more efficient LLM deployment at hyperscale. The company’s next-generation Asimov silicon (tape-out planned for late 2026; production in 2027) pairs compute with very large memory per chip (roughly 288 GB to 2,304 GB) and is designed to be combined into Titan configurations (four to eight Asimov chips) for extremely large models and long context windows. Positron has already deployed Atlas systems in the cloud, including 50+ racks at Oracle Cloud Infrastructure and partnerships with Parasail, Jump Trading, and i3d.net. Its goal is to provide scalable, memory-rich AI inference hardware that lowers cost and increases throughput for serving large language models.
Company Size
51-200
Company Stage
Series C
Total Funding
$1.2B
Headquarters
Reno, Nevada
Founded
2023
Get referred to Positron
See people who can refer or advise you
Simplify's Take
What believers are saying
- September 10, 2026 financing delivered $875 million at a $5 billion valuation.
- The round funds Asimov tapeout on TSMC N3P and a 2-megawatt engineering data center.
- Oracle, Parasail, Jump Trading, and i3d.net validate demand for Atlas capacity.
What critics are saying
- A failed N3P tapeout or LPDDR5X shortage kills Positron's custom-silicon thesis.
- Oracle Cloud Infrastructure anchors current usage; a switching decision would slash credibility.
- Nvidia, Groq, and Cerebras can outspend Positron on performance and customer acquisition.
What makes Positron unique
- Atlas runs at Oracle Cloud Infrastructure, with 50+ racks deployed by September 2026.
- Asimov uses commodity LPDDR5X, avoiding HBM and CoWoS bottlenecks.
- Titan targets 16-trillion-parameter, 10-million-token inference in one node.
Help us improve and share your feedback! Did you find this helpful?
Benefits
Health Insurance
Dental Insurance
Vision Insurance
Life Insurance
Disability Insurance
Unlimited Paid Time Off
Paid Holidays
Remote Work Options
401(k) Company Match
Company Equity
Growth & Insights and Company News
6 month growth
↑ 4%1 year growth
↑ 3%2 year growth
↓ -4%Tech ceos' doomsaying is a distraction from real, existing harms of AI. This essay argues that the fashionable talk of AI apocalypse is itself a kind of misdirection. By framing the debate as a binary between extinction and salvation, the piece contends, tech executives draw attention away from the concrete harms their systems already cause, from sprawling data centers to surveillance tools to the enormous wealth concentrating in a few hands. The argument has bite because the loudest doom warnings often come from the very people profiting most from the buildout. If the future is cast as an inevitable collision with superintelligence, then present-day choices about power use, labor, and military applications start to look like footnotes rather than decisions someone is actively making. Whether you find the existential worries credible or overblown, the essay makes a sharp point about attention as a resource. Framing a technology as destiny tends to foreclose the boring, winnable fights over regulation and accountability. It is a bracing counterweight to a week dominated by apocalyptic headlines, insisting that the real harms are here, specific, and addressable now. Free · one email each morning The daily AI briefing, tuned for humans. Unsubscribe anytime. Wired 1h ago Simon Willison 1h ago AI Snake Oil 16h ago The Decoder 16h ago The Next Web 1d ago Bloomberg 2d ago From the wortins blog. * BRKZ Funding: How Much It Raised, Its Valuation & Investors BRKZ raised $31M in a Series B round announced September 2026, led by Wa'ed Ventures and 500 Global. Here's the full breakdown of the round, the backers, and its AI procurement plan for Saudi Arabia's building boom. * Positron AI Funding: How Much It Raised, Its Valuation & Investors Positron AI raised $875M in a Series C at a $5B valuation, announced September 2026 and led by NEA, Atreides Management and Valor Equity Partners. Here's the full breakdown of the round and its inference chip plan. * TAR Funding: How Much It Raised, Its Valuation & Investors TAR raised $120M in a Series A at a $1B valuation, announced September 2026 and led by Spark Capital. Here's the full breakdown of the round, the backers, and its off-grid AI power plan.
Positron AI secures $875M in series C and C-1 funding rounds. NEA, Andra Capital, Atreides Management, Valor Equity Partners, and Dylan Patel's SemiAnalysis Capital lead the investment in Positron AI's $875 million funding rounds, focusing on energy-efficient hardware and software for AI inference. AI-assisted, human-published
Liberty Global Tech Ventures invests in AI inference hardware and software company, Positron AI. * Positron AI develops hardware and software for memory-first AI inference systems * Positron joins growing portfolio of AI firms that Liberty Global has backed including Higgsfield, XBOW, Legora and ElevenLabs PALO ALTO, Calif., Sept. 11, 2026 (GLOBE NEWSWIRE) - Liberty Global Tech Ventures, the technology investment arm of Liberty Global, today announced an investment in Positron AI, a leader in hardware and software for memory-first AI inference systems. Liberty Global joins a select group of investors in the oversubscribed $875 million funding round, which values the company at $5 billion. The round was co-led by NEA, Valor Equity Partners, Atreides Management, Andra Capital, Dylan Patel's SemiAnalysis Capital and Jim Clark. Liberty Global's selection reflects the strategic value it brings through its global technology footprint and experience scaling digital platforms. Founded in 2023, Positron develops hardware and software designed specifically for AI inference, the stage at which trained AI models process and generate outputs. Its memory-first architecture is designed to lower total cost of ownership and process requests faster by addressing the memory bottlenecks that constrain AI inference. Positron's memory-first inference systems use LPDDR5X memory, avoiding constrained HBM and CoWoS supply chains, while achieving more than 90% memory bandwidth utilization. Its Atlas systems can be installed in existing data center architecture without changes to cooling systems. Bobbie Maltiel, Partner at Liberty Global Tech Ventures, said: "While AI is being adopted by consumers and enterprises at a phenomenal rate, the ability to deliver token usage at the lowest possible cost is key to the full benefits being delivered for society. The Positron team have a relentless focus on this and we are delighted to be joining them on their mission to deliver high-speed AI at the lowest possible cost to businesses and consumers." The market for inference compute, the computing power needed to run AI models rather than train them, is expected to be worth $1.3tn by 2032 driven by hyperscaler demand*. Positron already counts Oracle, Jump Trading and Parasail among its customers. Its next-generation custom silicon, Asimov, tapes out in late 2026 with production in the second half of 2027. Asimov powers Titan, a multi-terabyte-memory inference system for long-context and next-generation AI workloads. Liberty Global Tech Ventures invests in category-defining AI and technology businesses and Positron joins firms like ElevenLabs, XBOW, Legora and Higgsfield in selecting Liberty Global as a strategic investor. ABOUT LIBERTY GLOBAL Liberty Global Ltd. (Nasdaq: LBTYA, LBTYB, LBTYK) delivers long-term shareholder value through the strategic management of two complementary platforms: Liberty Telecom and Liberty Growth. Liberty Telecom is a world leader in converged broadband, video and mobile communications, providing approximately 80 million fixed and mobile connections across Europe through advanced fiber and 5G networks that empower customers and strengthen national economies. The business generates aggregate revenue of $22 billion, including approximately $18 billion from nonconsolidated joint ventures and $4 billion from consolidated operations. Liberty Growth invests in scalable businesses across the technology, media, sports and infrastructure sectors, with a portfolio of roughly 70 companies and funds valued at $3.4 billion.* Together, these platforms reflect Liberty Global's focus on operating, enabling and investing in businesses with strong strategic fit and the potential to deliver sustainable long-term returns. *As independently valued as of December 31, 2025. This release contains forward-looking statements within the meaning of the Private Securities Litigation Reform Act of 1995, including statements with respect to business strategies, future growth prospects; valuation expectations and other information and statements that are not historical fact. These forward-looking statements involve certain risks and uncertainties that could cause actual results to differ materially from those expressed or implied by these statements. These risks and uncertainties include events that are outside of Liberty Global's control. These forward-looking statements speak only as of the date of this release. Liberty Global expressly disclaims any obligation or undertaking to disseminate any updates or revisions to any forward-looking statement contained herein to reflect any change in its expectations with regard thereto or any change in events, conditions or circumstances on which any such statement is based. ABOUT POSITRON AI Positron AI builds hardware and software to make AI inference dramatically cheaper and more energy efficient. The company's shipping product, Atlas, is deployed at hyperscaler scale today, and its next-generation custom silicon, Asimov, tapes out in late 2026 with production in the second half of 2027. Asimov powers Titan, a multi-terabyte-memory inference system for long-context and next-generation AI workloads. Positron's energy-efficient architecture allows its systems to be deployed in air-cooled or liquid-cooled data centers at varying rack densities, without dependence on constrained HBM or CoWoS supply chains. Learn more at positron.ai. | Investor Relations | Corporate Communications | | Michael Bishop +44 20 8483 6246 | Pádraig McGarrigle +44 7474 736967 | | / | [email protected] |
AI chip startup Positron's valuation skyrockets in latest funding round. 10 Sep 2026 11:47PM Add CNA as a trusted source to help Google better understand and surface our content in search results. Sept 10: Positron AI, a startup developing chips to run AI models, said on Thursday it has raised $875 million in its latest funding round, more than quadrupling its valuation in seven months to $5 billion. The startup had raised $230 million in February at a valuation of $1.06 billion, according to data provider PitchBook. Here are some details: - The rapidly growing AI inference market, focused on running trained AI models, has become a key battleground as companies develop specialized processors to challenge Nvidia's dominance. - The financing was split into two tranches: a $375 million Series C at a $3.5 billion pre-money valuation and a Series C-1 of up to $500 million led by NEA and Jim Clark, Positron said. - Proceeds will fund the final design stage of Asimov, the company's next-generation chip. The company targets production in the second half of 2027. - The company said its AI inference system, Titan, will combine four to eight Asimov chips to serve models exceeding 16 trillion parameters and context windows beyond 10 million tokens in a single node. - The round was co-led by NEA, Atreides Management, Valor Equity Partners, Andra Capital, SemiAnalysis Capital and Jim Clark, the founder of Silicon Graphics. Qatar Investment Authority, Cisco Investments and Naver Ventures were among additional investors. - Positron also said it is deploying more than 50 racks of Atlas, its first-generation inference system, at Oracle Cloud Infrastructure. - As part of the financing, Forest Baskett of NEA, Gavin Baker of Atreides Management, Dylan Patel, the founder of SemiAnalysis, and Thomas Jermoluk from Jim Clark Office will join Positron's board.
Positron AI raises $875 million Series C at $5 billion valuation. Cris Tolomia Positron AI raised $875 million at a $5 billion post-money valuation on Thursday, as demand for AI inference hardware continues to climb. The raise was structured in two tranches. A $375 million Series C, priced at a $3.5 billion pre-money valuation, drew co-lead investments from NEA, Atreides Management, Valor Equity Partners, Andra Capital, and SemiAnalysis Capital, the fund run by analyst Dylan Patel. A follow-on Series C-1 of up to $500 million was anchored by NEA and Netscape co-founder Jim Clark. Additional investors include DFJ Growth, Qatar Investment Authority, Hudson River Trading, Cisco Investments, and Naver Ventures, among others. Positron's most recent prior funding round was a $230 million raise in February that valued the company at $1 billion, according to the Wall Street Journal. Thursday's round values the company five times higher than that February figure. Reno, Nevada-based Positron designs inference chips built around memory capacity and bandwidth rather than raw compute. Its next-generation chip, called Asimov, pairs the company's compute architecture with between 288 GB and 2,304 GB of memory per chip using commodity LPDDR5X memory, sidestepping supply constraints around high-bandwidth memory and advanced packaging. Asimov is scheduled to tape out on TSMC's N3P process at the end of 2026, with production targeted for the second half of 2027, the company said. The proceeds will fund the Asimov tapeout, the build-out of a 2-megawatt engineering data center and emulation platform, and the production ramp of Titan, a next-generation inference system that combines four to eight Asimov chips into a single node. Titan is designed to serve models beyond 16 trillion parameters and context windows beyond 10 million tokens, the company said. Positron's earlier product, the Atlas server rack, has been deployed at Oracle Cloud Infrastructure, with over 50 racks now in the field. Customers using that capacity include Parasail, Jump Trading, and i3d.net. As part of the financing, Forest Baskett of NEA, Gavin Baker of Atreides Management, Thomas Jermoluk from Jim Clark Office, and Dylan Patel will join Positron's board of directors. Positron chief executive Mitesh Agrawal said in a statement that deploying Atlas at Oracle informed the design of Asimov and Titan. "Our focus now is to tape out Asimov, bring Titan to production, and scale manufacturing to meet the demand in front of us," Agrawal said in a statement.