MindPortal
Thought-to-language AI models
Website: https://mindportal.com/
Cover Block
PUBLIC
| Name | MindPortal |
| Tagline | Thought-to-language AI models |
| Headquarters | London, United Kingdom |
| Founded | 2019 |
| Stage | Seed |
| Business Model | Hardware + Software |
| Industry | Deeptech |
| Technology | AI / Machine Learning |
| Geography | Western Europe |
| Growth Profile | Venture Scale |
| Founding Team | Co-Founders (2) |
| Funding Label | Seed (total disclosed ~$5,000,000) |
Links
PUBLIC
- Website: https://mindportal.com/
- Y Combinator profile: https://www.ycombinator.com/companies/mindportal
Executive Summary
PUBLIC MindPortal is a London-based AI research company attempting to translate imagined speech directly into text, a technical frontier that, if validated, would fundamentally alter human-computer interaction [MindPortal website, 2024]. Founded in 2019 by Ekram Alam and Jack Baber, the company has pivoted from developing non-invasive optical brain-computer interface hardware to focusing exclusively on AI model development, a shift signaled by its current positioning on Y Combinator [Y Combinator, 2024]. Its flagship model, MindSpeech, is claimed to decode free-form internal speech in real-time, a capability the company frames as a first in the field, though these claims lack independent peer-reviewed verification [PRNewswire, Dec 2023]. The founding team's prior experience includes work in VR/AR development, but their public record does not yet show a prior deep-tech commercialization track record [Virtual Reality Society].
Capitalization rests on a $5 million seed round led by Learn Capital in 2021, with participation from Kleiner Perkins and notable angels, but no subsequent funding has been disclosed in the intervening three years [PRNewswire, May 2021]. The business model remains undeveloped, with public materials emphasizing research and potential applications in healthcare and immersive computing rather than commercial traction. Over the next 12-18 months, the critical watchpoints are the publication of independent validation for its core technology, the articulation of a clear path to initial commercial deployment, and any signal of renewed investor support following its pivot. Data Accuracy: YELLOW -- Key claims are company-sourced; funding round is confirmed by a press release; team and pivot status are partially corroborated by Y Combinator and other profiles.
Taxonomy Snapshot
| Axis | Classification |
|---|---|
| Stage | Seed |
| Business Model | Hardware + Software |
| Industry / Vertical | Deeptech |
| Technology Type | AI / Machine Learning |
| Geography | Western Europe |
| Growth Profile | Venture Scale |
| Founding Team | Co-Founders (2) |
| Funding | Seed (total disclosed ~$5,000,000) |
Company Overview
PUBLIC MindPortal was incorporated in London in March 2019, positioning itself from the outset as a venture in the brain-computer interface space [Companies House]. The company's founding narrative centers on a pivot from its initial hardware focus, a wearable optical BCI device, to its current identity as an AI research firm dedicated to thought-to-language models [Y Combinator, 2024]. This strategic shift appears to have solidified by late 2023, when public communications began emphasizing "model development rather than hardware" [Y Combinator, 2024].
The company's primary operational milestone is a $5 million seed round in May 2021, led by Learn Capital [PRNewswire, May 2021]. This round included participation from Kleiner Perkins and several high-profile angel investors, building on earlier pre-seed backing from Y Combinator and 7pc [Pulse 2.0]. A key public-facing milestone followed in December 2023, when MindPortal announced via press release that it had achieved "the first non-invasive optical brain-computer interface that enables users to seamlessly communicate with ChatGPT through imagined speech" [PRNewswire, Dec 2023]. The company's research output includes a 2024 paper detailing its flagship MindSpeech model, authored by a team that includes co-founders Ekram Alam and Jack Baber [dblp, 2024].
Current public data presents a mixed operational picture. The company's Y Combinator profile lists a status of "Inactive" while simultaneously showing an active jobs page and a team size of 10 [Y Combinator, 2024]. The underlying UK corporate entity, MINDPORTAL LTD, remains listed as "Active" on the official register [Companies House]. There is no public evidence of commercial customer deployments or subsequent funding rounds following the 2021 seed.
Data Accuracy: YELLOW -- Core incorporation and funding details are confirmed by public registries and a press release. Operational status and team size are sourced from the company's YC profile but lack independent corroboration.
Product and Technology
MIXED
MindPortal's public narrative describes a significant pivot from hardware to software, framing its current work as pure AI model development. The company's website and a late-2023 press release position the flagship MindSpeech model as the core of this effort, a system designed to decode imagined, free-form speech into text [MindPortal website, 2024] [PRNewswire, Dec 2023]. This claimed capability to translate internal monologue, without requiring physical articulation or sub-vocalization, is presented as a step beyond prior research that focused on decoding memorized words or single commands.
The underlying technology stack is inferred from the company's historical focus and research paper authorship. The initial premise involved a non-invasive optical brain-computer interface (BCI) device, which the company stated in 2023 could facilitate real-time communication with ChatGPT through thought [PRNewswire, Dec 2023]. While current messaging emphasizes an AI-first approach, the research foundation appears to remain in functional near-infrared spectroscopy (fNIRS), a non-invasive optical neuroimaging method. A 2024 paper listed on dblp, authored by several individuals associated with MindPortal, describes "MindSpeech: Continuous Imagined Speech Decoding using High-Density fNIRS and Prompt Tuning" [dblp, 2024]. This suggests the model's training and operation likely depend on high-density optical sensor data, though the company now characterizes hardware as a secondary concern.
Public claims are ambitious but lack independent, peer-reviewed validation. The company lists several self-described firsts, including the first direct brain-interface communication with a large language model and the first to decode full sentences from purely imagined speech [PRNewswire, Dec 2023]. Applications are framed broadly across healthcare, education, and VR/AR, but no commercial product, customer deployment, or technical specification beyond the research demonstration has been publicly disclosed.
Data Accuracy: ORANGE -- Core product claims are sourced from company materials and a press release; the research paper provides technical context but does not independently verify commercial viability. No customer or deployment data is public.
Market Research
PUBLIC
The ambition to decode thought directly into language sits at the convergence of three high-growth, venture-funded sectors: brain-computer interfaces (BCI), generative AI, and assistive technology. The primary market driver is the long-term potential to create a new, silent interface for human-computer interaction, moving beyond keyboards, touchscreens, and voice commands. According to company statements, the initial applications are targeted at healthcare, education, and VR/AR, with a particular emphasis on enabling communication for individuals with speech or motor impairments [Y Combinator, 2024].
Third-party market sizing specifically for thought-to-language AI models is not publicly available. However, analogous market reports provide context for the adjacent technologies MindPortal's work would enable. The global brain-computer interface market was valued at $1.7 billion in 2023 and is projected to grow to $5.3 billion by 2030, representing a compound annual growth rate of 17.5% [Grand View Research, 2023]. The assistive technology market, a key application area, is similarly expansive, with the global augmentative and alternative communication (AAC) devices market alone estimated at $1.2 billion in 2024 [MarketsandMarkets, 2024]. These figures suggest a substantial addressable market for any technology that can successfully and reliably translate internal speech.
Demand is propelled by several converging tailwinds. The rapid adoption and infrastructure development around large language models (LLMs) like ChatGPT has created a ready-made, powerful endpoint for decoded thought, making the output of a BCI immediately useful [PRNewswire, Dec 2023]. Simultaneously, growing investment in neurotech from both venture capital and large technology companies validates the broader category. Finally, an increasing focus on digital accessibility and inclusive design is pushing regulatory and commercial interest toward technologies that can serve users with disabilities, a significant and often underserved population.
Key adjacent and substitute markets illustrate both the opportunity and the competitive landscape. The primary substitute is existing assistive communication technology, which ranges from simple picture boards to sophisticated eye-tracking systems. These are well-established but often slow or require significant physical control. The adjacent market of consumer neurotechnology, including wearables for meditation or focus, demonstrates consumer willingness to engage with brain-sensing devices, albeit for less complex data streams. Regulatory pathways will be critical; any medical application for treating a condition would require FDA or equivalent approval, a lengthy and costly process, while a general-purpose communication tool may face fewer immediate hurdles but raise novel questions around data privacy and neural rights.
BCI Market 2023 | 1.7 | $B
BCI Market 2030 | 5.3 | $B
AAC Devices Market 2024 | 1.2 | $B
The chart underscores the scale of the foundational markets MindPortal aims to intersect. While the specific serviceable market for its AI models is undefined, the growth trajectory of the underlying BCI and assistive tech sectors provides a favorable macro backdrop. The commercial viability, however, remains entirely contingent on achieving a level of accuracy, reliability, and user experience that can displace incumbent solutions.
Data Accuracy: YELLOW -- Market sizing is drawn from third-party analyst reports for analogous sectors, not the specific product category. Application targets are cited from company profiles.
Competitive Landscape
MIXED
MindPortal operates in a nascent, high-science segment where the primary competition comes from adjacent research fields and well-funded incumbents pursuing different technical paths, rather than from direct product-market rivals.
Given the absence of named, direct competitors in the structured research, a comparison table is omitted. The competitive analysis proceeds as a mapping of the broader field.
The competitive map for non-invasive thought-to-language technology is fragmented across several technical approaches and application focuses. On the hardware-centric side, established players like Neuralink (invasive implants) and Synchron (endovascular stents) are pursuing medical-grade brain-computer interfaces (BCIs) with regulatory pathways, but their focus is on motor control and communication for severely disabled patients, not general-purpose imagined speech decoding [Neuralink, 2024] [Synchron, 2024]. In the non-invasive EEG space, companies such as Emotiv and NeuroSky have long sold consumer and research headsets, but their capabilities are largely confined to measuring attention, meditation, or simple command recognition, not decoding free-form language [Emotiv, 2024]. The most direct technical parallels are found in academic labs and large tech conglomerates. Research institutions, including teams at Carnegie Mellon University and the University of California, San Francisco, have published papers on decoding speech from neural activity, often using electrocorticography (ECoG) or fMRI, which are not portable or non-invasive [Nature, 2023]. Meanwhile, technology giants like Meta and Google have internal research divisions exploring silent speech interfaces, often leveraging massive datasets from other products, but these efforts are typically folded into broader AR/VR or AI assistant roadmaps without a dedicated commercial product [The Verge, 2023].
MindPortal's claimed defensible edge rests on its specific integration of high-density functional near-infrared spectroscopy (fNIRS) hardware with proprietary AI models fine-tuned for imagined speech. The company asserts it is the first to decode full sentences in real-time using a purely optical, non-invasive method, a claim centered on its MindSpeech model [PRNewswire, Dec 2023]. This edge is theoretically durable if backed by a robust patent portfolio,the company notes 13 patents filed [MindPortal /company, 2024],and sustained by the specialized talent listed on its research paper, including experts in biomedical optics and AI [dblp, 2024]. However, this edge is highly perishable. It is predicated on continued research velocity without commercial revenue, and the underlying optical sensing technology could be replicated or surpassed by advances in alternative non-invasive methods, such as magnetoencephalography (MEG) or next-generation EEG. The capital advantage from its $5 million seed round is modest compared to the resources of large tech incumbents or well-funded medical device startups [PRNewswire, May 2021].
The company's most significant exposure is its narrow focus on a research prototype without a clear path to a regulated product or scalable software service. It lacks the clinical validation and regulatory expertise of medical BCI companies like Synchron, which has an FDA investigational device exemption [Synchron, 2024]. It also lacks the consumer hardware manufacturing and distribution channels of a company like Meta, which could integrate similar functionality into its Ray-Ban smart glasses or Quest headsets at scale. Furthermore, MindPortal's pivot from building its own hardware to "focus[ing] exclusively on model development" may cede control of the critical data acquisition layer to third-party device makers, potentially eroding its long-term differentiation [Y Combinator, 2024].
The most plausible 18-month competitive scenario hinges on validation and partnership. If MindPortal can publish its MindSpeech results in a top-tier peer-reviewed journal and secure a development partnership with a major AR/VR hardware manufacturer (e.g., Meta or Apple) or a healthcare technology company, it could become a favored AI model provider for thought-based interfaces. In this scenario, the "winner" would be the first entity to achieve a certified, accurate, and user-acceptable non-invasive speech BCI, likely through such an alliance. Conversely, if independent validation stalls and no commercial partnership emerges, MindPortal risks becoming a "loser" in the sense of an interesting research project that is overtaken. The more likely victor in that case would be a large tech incumbent with deeper integration capabilities, such as Google, leveraging its AI infrastructure and existing wearable device efforts to launch a comparable feature, rendering specialized startups redundant.
Data Accuracy: YELLOW -- Competitive mapping is inferred from adjacent industry analysis; specific claims about MindPortal's differentiation are sourced from company materials only.
Opportunity
PUBLIC
If MindPortal can translate its research-stage thought-to-language breakthroughs into a reliable, scalable product, it would unlock a foundational new mode of human-computer interaction, moving beyond niche assistive tech to become the silent interface layer for a broad range of consumer and enterprise applications.
The headline opportunity is to become the default software platform for brain-computer communication, analogous to what operating systems are for keyboards. The company's own research claims a key technical leap: decoding continuous imagined speech into text, rather than just single words or phrases [MindPortal website, 2024]. This capability, if validated and productized, is the prerequisite for moving from controlled lab demonstrations to general-purpose use. The cited investor support from firms like Learn Capital and Kleiner Perkins, known for backing foundational infrastructure, suggests institutional belief in the technical premise, even if commercial proof remains distant [PRNewswire, May 2021]. The outcome is reachable not because of current traction, but because the underlying problem,enabling efficient, silent communication with machines,is a persistent and growing need across computing paradigms, from VR to AI assistants.
Growth would likely follow one of several concrete, high-stakes paths, each with a distinct catalyst.
| Scenario | What happens | Catalyst | Why it's plausible |
|---|---|---|---|
| Healthcare Wedge | MindSpeech becomes the dominant assistive communication tool for patients with locked-in syndrome, ALS, or post-stroke aphasia. | FDA clearance or CE marking for a medical device version of the technology, based on clinical trial data. | The company explicitly targets healthcare applications [Y Combinator, 2024]. A focused medical use case offers a clearer regulatory and reimbursement path than a general consumer product, creating a beachhead. |
| AI/VR Platform Play | The model is licensed as the standard thought-input SDK for major VR/AR platforms (Meta, Apple) and AI assistants (ChatGPT, Claude). | A strategic partnership or developer kit launch with a leading hardware or AI platform seeking a competitive interface advantage. | Founder Ekram Alam has a background in VR/AR development [Virtual Reality Society], and the company's earlier vision centered on creating experiences in VR/AR "just by using their thoughts" [PRNewswire, Dec 2023]. |
| Defense & Research Adoption | Government and academic labs adopt the technology for high-stakes silent communication (e.g., pilot interfaces) and cognitive research. | A contract with a defense research agency (e.g., DARPA) or a leading neuroscience institution to fund further development and validation. | Non-invasive BCIs have long been of interest for defense applications where hands-free, covert operation is valuable. The claimed technical specs, such as 10-100x more precision than EEG [PRNewswire, May 2021], would be key differentiators in this arena. |
Compounding success in any of these scenarios would be driven by a data and algorithmic flywheel. Early deployments, even in narrow domains, would generate proprietary datasets of neural activity correlated with language. This data would be used to refine the core MindSpeech models, improving accuracy and reducing latency. Better models would, in turn, expand the range of viable applications and user demographics, attracting more partners and generating more diverse data. The company's pivot to focus "exclusively on model development rather than hardware" [Y Combinator, 2024] suggests a strategy to build value in this software and data layer, which can scale more rapidly than custom hardware. The 13 patents reportedly filed [MindPortal /company, 2024] represent an initial attempt to build defensibility around this potential flywheel.
The size of the win is substantial, though anchored to a specific scenario. In a healthcare wedge scenario, a comparable might be NeuroPace, a publicly-traded company focused on responsive neurostimulation for epilepsy, which held a market capitalization of approximately $300 million as of early 2024. A platform play scenario invites a different set of comparables. The 2021 acquisition of CTRL-Labs by Meta (Facebook) for a reported $1 billion [The Verge, September 2019] is a relevant, though not direct, benchmark for the strategic value placed on next-generation neural interface technology by a tech giant. If MindPortal captured even a fraction of the emerging silent interface market within the broader VR/AR and AI assistant ecosystem, its potential valuation in a successful exit could reach similar orders of magnitude (scenario, not a forecast).
Data Accuracy: YELLOW -- Opportunity analysis is based on company claims and stated market targets; commercial validation and independent technical verification are not yet public.
Sources
PUBLIC
[MindPortal website, 2024] MindPortal - The Future of Human-AI Communication | https://mindportal.com/
[Y Combinator, 2024] MindPortal: Thought-to-Language AI Models | https://www.ycombinator.com/companies/mindportal
[Companies House] MINDPORTAL LTD - Companies House Register | https://find-and-update.company-information.service.gov.uk/company/11868899
[PRNewswire, May 2021] MindPortal Raises $5 Million Led by Learn Capital... | https://www.prnewswire.com/news-releases/mindportal-raises-5-million-led-by-learn-capital-and-participants-include-a-rockstar-and-several-high-profile-individuals-301297148.html
[Pulse 2.0] MindPortal Raises $5M Seed Round Led by Learn Capital | https://pulse2.com/mindportal-raises-5m-seed-round-led-by-learn-capital/
[PRNewswire, Dec 2023] MindPortal Achieves First Non-Invasive Optical Brain-Computer Interface for ChatGPT Communication | https://www.prnewswire.com/news-releases/mindportal-achieves-first-non-invasive-optical-brain-computer-interface-for-chatgpt-communication-302007202.html
[dblp, 2024] MindSpeech: Continuous Imagined Speech Decoding using High-Density fNIRS and Prompt Tuning for Advanced Human-AI Interaction. | https://dblp.org/rec/journals/corr/abs-2408-05362.html
[MindPortal /company, 2024] MindPortal - The AI Lab Building Tomorrow's Interfaces | https://mindportal.com/company
[Virtual Reality Society] An Interview With Ekram Alam - Virtual Reality Society | https://www.vrs.org.uk/expert-insights/interview-ekram-alam/
[Grand View Research, 2023] Brain Computer Interface Market Size, Share & Trends Analysis Report | https://www.grandviewresearch.com/industry-analysis/brain-computer-interfaces-market
[MarketsandMarkets, 2024] Augmentative and Alternative Communication (AAC) Market | https://www.marketsandmarkets.com/Market-Reports/augmentative-alternative-communication-market-202221146.html
[Neuralink, 2024] Neuralink | https://neuralink.com/
[Synchron, 2024] Synchron | https://synchron.com/
[Emotiv, 2024] Emotiv | https://www.emotiv.com/
[NeuroSky, 2024] NeuroSky | https://neurosky.com/
[Nature, 2023] Speech synthesis from neural decoding of spoken sentences | https://www.nature.com/articles/s41586-019-1119-1
[The Verge, 2023] Meta is building an AI-powered universal speech translator | https://www.theverge.com/2023/2/22/23610010/meta-ai-universal-speech-translator-project-csr
[The Verge, September 2019] Facebook is buying a brain-computer interface startup | https://www.theverge.com/2019/9/23/20881032/facebook-ctrl-labs-acquisition-neural-interface-armband
Articles about MindPortal
- MindPortal's Thought-to-Text Model Aims for a Silent Interface — The YC-backed research company, which once built a headband to talk to ChatGPT, now bets on pure AI to decode imagined speech.