NLM Photonics is a University of Washington spin-out commercializing silicon-organic hybrid (SOH) electro-optic modulators that push more bandwidth through less power inside the transceivers that AI data centers depend on. The wedge is a proprietary organic electro-optic material stack, licensed from UW and now backed by NLM-owned patents, that drops into existing silicon photonics foundry processes (AIM Photonics 300mm, Tower Semiconductor PH18M) to deliver higher bandwidth-per-watt than incumbent silicon-only or lithium-niobate modulators. The company sells to transceiver vendors, foundries, and hyperscalers on an NRE-plus-IP-licensing model rather than building its own fab.
StoryHouse wrote a $190K check into a $500K uncapped SAFE bridge (5% discount to the next priced round) that closed alongside Emerald Technology Ventures' signed term sheet for a $12M Series A at a $16.5M pre-money valuation — first tranche of $8.5M with Emerald ($3M), Idemitsu ($2M), Oregon Venture Fund ($2M), TOK ($1M) and the $0.5M SAFE bridge. Validation stack going in: signed AFWERX Phase II grant ($1.8M, per the May-30 note), a $300K Synterra commercial contract, an ISA with Google, a commercial agreement with Photonic Inc. (all per the Sep-4 note), and a favorable independent reference call from the Chemical Angel Network. The May-2026 term sheet for a follow-on Series A-5 led by Pangaea Ventures Impact Fund at a $31M pre-money marks the SH position up ~1.9× from the entry reference.
AI-driven data-center power and bandwidth constraints are the forcing function. Every additional 100G-per-lane demanded from a transceiver worsens the joule-per-bit budget; NLM's SOH modulators attack exactly that bandwidth-per-watt bottleneck at the pluggable, mid-board and co-packaged-optics tiers. The market is fragmenting into competing modulator physics — silicon-only, thin-film lithium-niobate, plasmonics, organic — and buyers (hyperscalers, transceiver vendors) will place multiple bets. Being drop-in with existing silicon photonics foundries is what makes this a real 2026–2028 shot rather than a decade-out science project.
| Player | Positioning | Funding / Stage | Edge vs. them |
|---|---|---|---|
| NLM Photonics | Silicon-organic hybrid (SOH) electro-optic modulators, drop-in to existing Si-photonics foundry flows | Series A closed 2024 (Emerald-led) | — |
| Lightwave Logic | Electro-optic polymer modulators for pluggables and co-packaged optics | Public (LWLG); further along on multi-foundry supply chain | NLM material stack is validated inside standard Si-photonics PDKs (AIM, Tower); competitor is scale/timing risk, not tech overlap |
| Polariton Technologies | Plasmonic modulators, Swiss ETH spin-out | Acquired by Marvell (2025) | Validates the "large semi player buys niche modulator IP" thesis; NLM is one of the remaining independent targets |
| TFLN players (e.g. HyperLight) | Thin-film lithium-niobate modulators | Well-funded private | Higher raw modulation performance but requires new PDK and packaging; NLM's foundry compatibility is the wedge |
| Silorix | Silicon-photonics IP/modulator startup (per DD form) | Early-stage private | Direct-adjacent but earlier and narrower in disclosed customer engagement |
Moat: Material chemistry IP (co-owned with UW plus company-owned filings), demonstrated integration on Tower PH18M and AIM 300mm PDKs, and a hyperscaler / transceiver-vendor design-partner base that would be expensive and slow for any competitor to replicate.
Named comparable exits: Cisco / Acacia at ~$4.5B Web, and — most on-point for NLM's category — Marvell's 2025 acquisition of Swiss plasmonics modulator startup Polariton and its December 2025 deal to acquire optical-interconnect startup Celestial AI for up to $5.5B Web. The natural acquirer set for NLM is the same: Marvell, Broadcom, Cisco, Lumentum, Coherent, plus the transceiver / foundry partners already engaged. The May-2026 Series A-5 term sheet at $31M pre-money already implies ~1.9× on the SH-entry reference terms; a strategic-sale outcome in the plausible sector-comparable range would deliver a multi-bagger on the $190K position.
Call with Lewis Johnson confirming the Emerald-led Series A term sheet is signed at $12M on a $16.5M pre-money, with the $500K uncapped SAFE bridge (this deal) rolling into that priced round.
Next steps: discuss with Miles but Josh and Matthew both on board to do remainder of bridge for $190K.
They have a term sheet from Emerald for $3M presuming they raise an additional $9M, on a $16.5M pre. IdeaShip has been working with them since early this year (Michael).
Round updates: had an acceleration on a few projects that sped up burn (Synterra customer came forward with a project that had accelerated need — building them a PIC; to meet the timeline they didn't have enough resources so worked with Enosemi). Started this small bridge of $500K to lead into $12M.
Doing the $500K bridge while Emerald closes, which will count toward the full $9M they need to raise. Term sheet has now been signed with Emerald. Emerald is doing $3M but bringing in their strategic LPs. TOK pencilled in for $1M–$2.5M. $500K bridge — $300K has been committed; Pack did $160K; Lewis and individuals $150K.
Business traction: $300K contract with Synterra; commercial agreement with Photonic Inc.; ISA in place with Google. Technical updates: filed new patent; closed out NASA project. Minimum check size for the A will likely be seven figures.
Round terms / post-money valuation: $12M round on a $16.5M pre-money valuation. Emerald is doing $3M but bringing in syndicate amounts for their LPs and will likely take up $7–9M.
Call with Lewis Johnson: Emerald has offered to lead the Series A at $10M on a $16.5M pre with a board seat and 1× preferred, targeting term sheet end of July and close late August; customer/partner slate now includes two transceiver vendors, a large US AI company, a Japanese company, one hyperscaler and AMF as a fab partner.
Company & insight: higher bandwidths at less power — applicable across all transceivers in the market. More PICs per wafer than existing approaches, saving customers money. Higher bandwidth and lower power affects both opex and capex. Data centers have long-term limits on power capacity.
Round dynamics: Emerald is offering to be the lead. Three-stage program: general / deep dive (currently in process) with site visit and IC meeting. Terms $10M at $16.5M pre / board seat / 1× preferred. They would put in $3M and likely syndicate the rest and have official term sheet end of July. Pack Ventures / TLK / OUP all have pro rata. Targeting late August. Thinks this can get them to exit or strategic B.
Traction: one customer came back and said they don't want just the modulator, they want the whole PIC — so they are now going to internal dev. Made agreement with another company to work on PIC design. Customers: two transceiver vendors; a large US AI company (with business in Israel), 2–3 year project; one Japanese company; one hyperscaler; one fab to manage — AMF.
Lead investor: Emerald VC $3M+. Round terms / post-money valuation: $10M round on $16.5M pre.
Email from Lewis Johnson: Brad Booth on a Bay Area trip; NLM happy to connect StoryHouse with Damoder Reddy (CAN) to share diligence; AFWERX Phase II is milestone-based with ~$75K expected later that month and ~$200K early September against a 21-month, milestone-adjusted schedule.
Email updates from Lewis Johnson. Brad may be able to provide additional updates after his Bay Area trip. Lewis happy to connect us with Damoder to share diligence info.
Marion is working on determining what AFWERX info can be put in the data room. The contract is milestone-based. Anticipating $75K later this month, and $200K in early September as the first two payments on the project. There may be some adjustments to the timing on other milestones from those on the original contract based on interaction with AIM's MPW tapeout dates, affecting timing for both money in and money out, but keeping the total project cost and timeline (21 months) the same.
Reference call with Damoder Reddy of the Chemical Angel Network (CAN), a group of PhD material-science investors. CAN hired a prospective customer to help diligence NLM. Reddy called Lewis Johnson "top percentile" scientifically and singled out the addition of Brad Booth as decisive — Booth's Microsoft/Meta relationships should help unlock hyperscaler customers. He said NLM's IP is strong and he does not see any competitor as advanced; the open question is commercial conversion.
Flagged risks: (a) lab-to-fab transition where no one on the team has done it before, (b) the IP-licensing business model, and (c) connectivity between chip-to-memory and co-packaged ICs. CAN was investing $50K from the fund plus ~$30K from individuals.
Source: Meeting Notes recXs27WWuIl1eGZy (2024-06-06)