Loop AI: Enterprise SLM Platform Founded 2012 (2026)
Loop AI is a San Francisco enterprise AI company founded in 2012 that trains client-owned language models, valued at $4.2 billion after its Dec 2025 round.
Loop AI helps large enterprises like banks, automakers, and telecoms build their own small language models instead of depending on API calls to Anthropic or OpenAI. Its Agents Orchestra platform lets non-technical teams manage the full agent lifecycle, from training through deployment, while keeping every model and dataset on the client's own infrastructure.
Founded in 2012 in San Francisco, Loop AI is an enterprise AI company that trains client-owned small language models on proprietary data instead of reselling third-party LLM APIs. Its Loop Q Cognitive Platform and Agents Orchestra tool let banks, telecoms, and automakers deploy on-premises models with autonomous learning, aiming for full data sovereignty and reduced vendor lock-in.
Founded: 2012 · HQ: San Francisco, CA, USA · Team: 20-50 · CEO: Gianmauro Calafiore · Funding: $100M raised in a Dec 2025 round led by AI Fund; historical investors reported include United Ventures, WI Harper Group, and Club Italia Investimenti (participation in the 2025 round unconfirmed). Prior disclosed round: Series A, 2017, undisclosed amount. · Valuation: $4.2B (reported Dec 2025, post-money after $100M round)
About Loop AI
Loop AI, legally Loop AI Labs and also known as Loop AI Group, was founded in 2012 in San Francisco by Gianmauro (GM) Calafiore, who serves as Founder and President. The founding team drew on backgrounds from SRI International's AI Center and Stanford's Computational Semantics Lab, with roots tracing back to work on DARPA's CALO project, the government-funded cognitive assistant research effort that also fed into Siri. The founding thesis was that enterprises should not have to rent access to a third party's general-purpose model to get value from AI on their own data. Instead, Loop AI built tooling for companies to train and own small language models (SLMs) trained exclusively on their proprietary data. The company's product line centers on two platforms. Loop Q Cognitive Platform lets enterprises train custom SLMs on their own structured and unstructured data using unsupervised learning, with autonomous, continuous learning cycles and support for on-premises or edge inference rather than cloud-only API calls. Loop AI Agents Orchestra is a no-code, drag-and-drop interface for managing the full lifecycle of AI agents, training, deployment, and runtime performance optimization, and has reportedly been used in Fortune 100 production environments since 2019. Both products target regulated, high-stakes industries: automotive, telecom, banking, insurance, media, and healthcare, where firms want AI outputs without sending proprietary data to an external model provider. The most significant recent move is a $100 million funding round closed in December 2025, reported by outlets including OpenPR and Dealroom at a $4.2 billion valuation, led by AI Fund, the venture studio founded by Andrew Ng. This marks a large jump from Loop AI's last confirmed round, a 2017 Series A of undisclosed size, and signals renewed investor appetite for enterprise-owned model infrastructure as a category distinct from foundation model API resale. Prior aggregate investor lists (Crunchbase and Craft) also name United Ventures, WI Harper Group, and Club Italia Investimenti among historical backers, though it is not confirmed which of these participated in the December 2025 round specifically. Loop AI's business model is enterprise licensing rather than self-serve subscriptions or per-token API billing. There is no published pricing tier; engagement happens through direct sales and "book a demo" requests aimed at large account procurement cycles measured in months. No public revenue or ARR figures have been disclosed by the company or a credible third party as of 2026. GM Calafiore leads the company as Founder and President; some third-party sources also list him as CEO. Other named team members include Marco Torresi in communications, Emanuel Mazzilli in engineering with prior experience at Facebook and Twitter, and analyst Tom Davenport in an advisory capacity. PitchBook estimates the company at roughly 45 employees as of 2026, a lean headcount relative to its reported valuation, consistent with an enterprise-sales-heavy, engineering-light operating model rather than a large research organization. The company's stated mission is to give enterprises what it calls Artificial Human Capacity: real-time, continuous, autonomous learning from a business's own data, explicitly positioned against symbolic, rules-based automation. Loop AI does not publish detailed model cards, safety research, or a responsible scaling policy in the way frontier labs like Anthropic or OpenAI do; its public materials focus on data ownership and deployment sovereignty rather than model safety research output. Competitively, Loop AI sits closer to enterprise automation and applied-AI platform vendors, such as C3.ai on enterprise AI platform positioning, UiPath on agent and workflow orchestration, and Automation Anywhere on no-code agent tooling, than to frontier model labs. Its differentiator is that clients own the trained model outright and can run it on-premises, versus renting inference from a shared model behind an API. The tradeoff is that Loop AI has far less public benchmark data, developer mindshare, or documentation than either camp, making enterprise technical due diligence slower. No compliance certifications (SOC 2, ISO 27001, HIPAA, GDPR-specific attestations) or a public trust center were found as of 2026, which is notable given the company's enterprise, regulated-industry customer base; this is likely handled through direct contractual security review rather than a published compliance program. No IPO filing, acquisition, or merger activity has been reported, and the company remains privately held with no disclosed stock ticker.
Mission
To give enterprises Artificial Human Capacity: real time, continuous, autonomous learning from their own data instead of manual symbolic logic.
Products
- Loop Q Cognitive Platform (Enterprise SLM training platform): https://www.loop.ai/loop-q-cognitive-platform
- Loop AI Agents Orchestra (No-code agent orchestration platform): https://www.loop.ai/loop-ai-agents-orchestra
Links
Frequently Asked Questions
What is Loop AI and what do they build?
Loop AI is a San Francisco enterprise AI company founded in 2012 by Gianmauro Calafiore, a founding team that grew out of research work at SRI International's AI Center and Stanford's Computational Semantics Lab. Rather than reselling access to a general-purpose LLM through an API, Loop AI builds tooling that lets enterprises train small language models (SLMs) on their own proprietary data and keep full ownership of the resulting model. Its two named products are the Loop Q Cognitive Platform, which handles unsupervised training on structured and unstructured client data, and Loop AI Agents Orchestra, a no-code interface for building, deploying, and tuning AI agents. Loop Q supports on-premises and edge inference so the trained model never has to leave a client's own infrastructure. Agents Orchestra has reportedly been used in Fortune 100 production environments since 2019, targeting regulated sectors such as banking, insurance, telecom, automotive, media, and healthcare. Access is through direct enterprise sales rather than a self-serve signup or public API key. Loop AI's market position is as a data-sovereignty-first alternative to buying inference from shared foundation model APIs.
Who founded Loop AI and who is the CEO?
Loop AI was founded in 2012 in San Francisco by Gianmauro (GM) Calafiore, who holds the title of Founder and President; some third-party listings also describe him as CEO. Calafiore's background traces to AI research tied to SRI International and the DARPA CALO project, the government-funded cognitive assistant program that also contributed early technology to Siri. Named team members include Marco Torresi handling communications, Emanuel Mazzilli in engineering with prior experience at Facebook and Twitter, and analyst Tom Davenport, well known for his enterprise AI research, listed in an advisory capacity. No distinct, separately named CTO or CSO has been publicly confirmed as of 2026. There is no public record of a major leadership change or founder departure since 2012, which is unusual longevity for a private AI company of this age. The company has not disclosed a formal governance structure such as a benefit trust or long-term stewardship board. Loop AI Labs is the legal entity name behind the Loop AI and Loop AI Group branding used across its marketing site and press coverage.
How much funding has Loop AI raised?
Loop AI closed a $100 million funding round in December 2025 at a reported $4.2 billion valuation, according to OpenPR and Dealroom coverage from that month. The round was led by AI Fund, the venture studio founded by Andrew Ng, marking a large step up from Loop AI's last confirmed disclosed round, a 2017 Series A of undisclosed size. Historical investor lists aggregated by Crunchbase and Craft also name United Ventures, WI Harper Group, and Club Italia Investimenti as backers, though it is not confirmed whether all three participated in the December 2025 round specifically. No public revenue or annual recurring revenue figure has been disclosed by Loop AI or by an independent credible source as of 2026. The company has not announced an IPO timeline or acquisition talks. Total disclosed funding across all rounds, based on public reporting, is the $100 million December 2025 round plus an undisclosed 2017 Series A amount. No government grants or sovereign wealth fund participation have been reported.
What products does Loop AI make?
Loop AI's two named products are Loop Q Cognitive Platform and Loop AI Agents Orchestra, both aimed at enterprise buyers rather than individual developers or consumers. Loop Q Cognitive Platform trains custom small language models on a client's own structured and unstructured data using unsupervised learning, with support for on-premises or edge deployment so the trained model runs inside the client's own infrastructure rather than a shared cloud API. Loop AI Agents Orchestra is a no-code, drag-and-drop interface for managing the full lifecycle of AI agents, covering training, deployment, and ongoing runtime optimization, and has reportedly been in Fortune 100 production use since 2019. Neither product has published self-serve pricing; both are sold through direct enterprise contracts, with pricing negotiated per account rather than listed in fixed tiers. There is no confirmed free tier, trial signup, or open-source release of either product. Loop AI does not currently offer a consumer-facing chat product or a public developer API comparable to OpenAI or Anthropic's offerings.
Where is Loop AI headquartered and how big is the team?
Loop AI is headquartered in San Francisco, California, with public company records citing an address on Mission Street in the city's South of Market district. PitchBook estimates the company's headcount at roughly 45 employees as of 2026, a comparatively lean team relative to its reported $4.2 billion valuation. This is consistent with an enterprise-sales-driven operating model rather than a large in-house research organization; no separate office locations or per-hub headcount breakdowns have been publicly disclosed. There is no public reporting of a major hiring surge or layoff round tied to the company in 2025 or 2026. Named team members span founding leadership, communications, and engineering, with an outside analyst serving in an advisory role. No remote-work or distributed-team policy has been publicly described. No new country offices or expansions have been reported as of 2026.
What is Loop AI's mission or research focus?
Loop AI's stated mission is to give enterprises what it calls Artificial Human Capacity, meaning real-time, continuous, and autonomous learning from a business's own data rather than relying on manually programmed, rules-based logic. In practice this means the company positions itself against symbolic, hand-coded automation and in favor of neural models that keep learning from a client's live data streams. Loop AI does not publish detailed technical research papers, model cards, or system cards in the way frontier labs like Anthropic, OpenAI, or Google DeepMind do. It also has not published a formal responsible scaling policy, usage policy, or named external red-teaming partners as of 2026. Its safety and governance posture is oriented around data sovereignty, on-premises deployment, and client ownership of trained models, rather than published frontier-model safety research. This differs sharply from research-first labs whose public output centers on interpretability, alignment, or scaling-law papers. The company's public materials focus on business outcomes for enterprise clients rather than open research contributions to the broader AI field.
Is Loop AI compliant with SOC 2, GDPR, HIPAA?
No SOC 2, ISO 27001, HIPAA, or GDPR-specific compliance certification, and no public trust center, was found for Loop AI as of 2026, despite the company selling into regulated industries such as banking, insurance, and healthcare. This is likely handled through direct contractual security review with individual enterprise customers rather than a published, self-serve compliance program, which is common for smaller enterprise vendors with a lean headcount. No public data retention policy, data processing agreement template, or standard contractual clauses document was located. The company's on-premises and edge deployment model is marketed as a data-sovereignty advantage in itself, letting clients keep trained models and data inside their own infrastructure rather than relying on a vendor's cloud retention policy. No EU AI Act classification has been publicly disclosed; the company's customer base as described in its marketing appears to be primarily US enterprise accounts. Prospective enterprise buyers evaluating Loop AI for regulated workloads should expect to request compliance documentation directly rather than finding it published online. No HIPAA business associate agreement availability has been confirmed.
Who are Loop AI's main competitors?
Loop AI competes most closely with enterprise AI and automation platform vendors rather than frontier foundation model labs, given its focus on client-owned small language models and agent orchestration. C3.ai is a comparable rival on broad enterprise AI platform positioning, though C3.ai leans more on prebuilt vertical applications while Loop AI leans on training bespoke client-owned models. UiPath is a competitor on agent and workflow orchestration, with UiPath having far greater public market presence and a listed stock (NYSE: PATH) versus Loop AI's private, low-profile status. Automation Anywhere competes on no-code agent tooling similar to Loop AI's Agents Orchestra product. Loop AI's primary differentiator is that clients train and own the resulting small language model outright rather than renting inference from a shared API, which appeals to data-sovereignty-sensitive industries like banking and healthcare. Its main weakness against these rivals is a near-total lack of public benchmark data, documented compliance certifications, and case study depth, which slows enterprise buyer due diligence. An emerging competitive pressure not on the radar a few years ago is the rise of general-purpose foundation model vendors offering their own fine-tuning and on-premises deployment options, which narrows Loop AI's structural advantage over time.