Smoltek Semi has achieved 1 µF (one micro­farad) of capac­i­tance on an area mea­sur­ing just one square mil­lime­ter. This rep­re­sents a dra­mat­ic leap from nano­farad to micro­farad ter­ri­to­ry – a thresh­old that seemed almost myth­i­cal just years ago for ultra-thin inte­grat­ed capac­i­tors. While we’re not the first to cross this cov­et­ed line – a few oth­er com­pa­nies have achieved this pres­ti­gious mile­stone in the spe­cial­ized field of ultra-thin capac­i­tor tech­nolo­gies – our accom­plish­ment sets us apart in this exclu­sive club. We reached 1 µF/​mm² while keep­ing our capac­i­tor incred­i­bly thin, unlock­ing pos­si­bil­i­ties that oth­ers sim­ply can­not match.

The proximity imperative

Today’s smart­phones pack more com­put­ing pow­er than room-sized super­com­put­ers from decades past. High-per­for­mance com­put­ing (HPC) cen­ters and AI accel­er­a­tors push pro­cess­ing speeds to extra­or­di­nary lim­its. But this per­for­mance comes with a fun­da­men­tal chal­lenge: as chips become faster, they become more demand­ing about their pow­er supply.

Think of tran­sis­tors as incred­i­bly fast switch­es, flip­ping bil­lions of times per sec­ond between ones and zeros. Each flip demands instant ener­gy, cre­at­ing pow­er fluc­tu­a­tions that would crip­ple per­for­mance with­out inter­ven­tion. Capac­i­tors serve as light­ning-fast ener­gy reser­voirs, absorb­ing excess pow­er and releas­ing stored ener­gy pre­cise­ly when needed.

The clos­er these capac­i­tors sit to the chip, the bet­ter they per­form. Dis­tance means delay, and delay means degrad­ed per­for­mance. The ulti­mate prox­im­i­ty lies direct­ly under the chip, nes­tled between the sol­der balls that con­nect the proces­sor to the cir­cuit board. But this prime real estate comes with a mer­ci­less con­straint: height.

The ultra-thin breakthrough

Here’s where most solu­tions hit their lim­i­ta­tion wall. Achiev­ing 1 µF/​mm² while fit­ting under sol­der balls demands an ultra-thin pro­file. The sol­der balls them­selves define the height lim­it – typ­i­cal­ly allow­ing only min­i­mal clear­ance for inte­grat­ed components.

Smoltek has cracked this code with a capac­i­tor fea­tur­ing an active lay­er of just 6 µm. To put this in per­spec­tive, a human hair mea­sures about 70 µm in diam­e­ter. Smoltek’s active capac­i­tor lay­er could fit more than ten times under the width of a sin­gle hair. This isn’t just thin – it’s what engi­neers call ”tru­ly ultra-thin” for the crit­i­cal per­for­mance area.

This achieve­ment trans­forms Smoltek from anoth­er mem­ber of the 1 µF/​mm² club into the only com­pa­ny capa­ble of deliv­er­ing that per­for­mance where it mat­ters most: in the space-con­strained under-chip envi­ron­ment where every microm­e­ter counts.

Building forests, not digging trenches

Under­stand­ing Smoltek’s break­through requires grasp­ing how CNF-MIM capac­i­tors work. Deep Trench Capac­i­tors (DTC) carve deep trench­es – like minia­ture wells – in a sil­i­con sub­strate. In con­trast, CNF-MIM man­u­fac­tur­ing takes a fun­da­men­tal­ly dif­fer­ent approach by grow­ing forests of car­bon nanofibers on a sub­strate. Pic­ture micro­scop­ic trees, each about 10,000 times thin­ner than human hair, stand­ing ver­ti­cal­ly in dense formation.

The mag­ic hap­pens in the lay­er­ing process. Each car­bon nanofiber gets coat­ed with met­al, applied atom by atom using Atom­ic Lay­er Depo­si­tion (ALD). Over this met­al lay­er, Smoltek applies our pro­pri­etary ZrO₂–Al₂O₃ dielec­tric stack – again, atom by atom with atom­ic pre­ci­sion. Final­ly, anoth­er met­al lay­er com­pletes the struc­ture, cre­at­ing a met­al-insu­la­tor-met­al capac­i­tor at nanoscale.

This atom-by-atom con­struc­tion allows Smoltek to uti­lize 100% of the sur­face area pro­vid­ed by the car­bon nanofibers. No oth­er depo­si­tion method achieves this com­plete cov­er­age, mak­ing ALD essen­tial for max­i­miz­ing capac­i­tance density.

SkyTech partnership enables production scale

The pre­ci­sion ALD process­es that make CNF-MIM pos­si­ble require spe­cial­ized equip­ment at indus­tri­al scale. Smoltek has part­nered with SkyTech, a Tai­wanese leader in ALD equip­ment for the semi­con­duc­tor industry.

The part­ner­ship gives Smoltek access to the same indus­tri­al-scale depo­si­tion tech­nol­o­gy used through­out semi­con­duc­tor man­u­fac­tur­ing. SkyTech’s exper­tise ensures our process­es main­tain their per­for­mance char­ac­ter­is­tics as they tran­si­tion from lab­o­ra­to­ry to mass production.

Most impor­tant­ly, the col­lab­o­ra­tion proves real-world scal­a­bil­i­ty. Once the process­es are ful­ly devel­oped, scal­ing up for mass pro­duc­tion becomes straight­for­ward. This isn’t lab­o­ra­to­ry inno­va­tion that might work some­day – it’s pro­duc­tion-ready tech­nol­o­gy with a clear path to market.

The cost equation advantage

Beyond supe­ri­or per­for­mance, CNF-MIM tech­nol­o­gy deliv­ers com­pelling eco­nom­ics. While com­pet­ing tech­nolo­gies, such as DTC, demand mul­ti­ple expen­sive ALD process­es to achieve sim­i­lar results, Smoltek’s approach requires sig­nif­i­cant­ly few­er lay­ers – direct­ly trans­lat­ing to low­er pro­duc­tion costs.

ALD rep­re­sents one of the most expen­sive process­es in semi­con­duc­tor man­u­fac­tur­ing. Since CNF tech­nol­o­gy can achieve the same capac­i­tance with few­er ALD lay­ers com­pared to deep trench alter­na­tives, the cost advan­tage com­pounds with each per­for­mance incre­ment. This isn’t just about bet­ter tech­nol­o­gy – it’s about bet­ter tech­nol­o­gy that costs less to produce.

The eco­nom­ic impli­ca­tions are pro­found: as per­for­mance require­ments increase, tra­di­tion­al approach­es see expo­nen­tial cost growth, while CNF-MIM main­tains cost effi­cien­cy through its supe­ri­or three-dimen­sion­al architecture.

From laboratory to production reality

These results aren’t lab­o­ra­to­ry curiosi­ties achieved under per­fect con­di­tions. The major­i­ty of test­ed devices showed con­sis­tent capac­i­tance den­si­ty with reli­able repro­ducibil­i­ty – the hall­mark of mature tech­nol­o­gy ready for commercialization.

While cur­rent data comes from sim­pli­fied CNF-MIM devices in lab­o­ra­to­ry set­tings, the same dielec­tric stack pro­duces pro­to­types and prod­ucts achiev­ing iden­ti­cal per­for­mance. The 18-month devel­op­ment process has cre­at­ed pro­duc­tion-ready process­es that trans­late direct­ly to com­mer­cial manufacturing.

Devel­op­ment focus has now shift­ed to improv­ing insu­la­tion resis­tance – essen­tial for min­i­miz­ing leak­age cur­rent and ensur­ing long-term reli­a­bil­i­ty in com­mer­cial applications.

Industry validation

The sig­nif­i­cance of Smoltek’s achieve­ment extends beyond tech­ni­cal met­rics. In a recent inter­view, lead­ing capac­i­tor man­u­fac­tur­er Yageo acknowl­edged the ground­break­ing capa­bil­i­ties of CNF-MIM technology.

This indus­try val­i­da­tion con­firms what the tech­ni­cal data sug­gests: Smoltek has devel­oped some­thing gen­uine­ly unique in the capac­i­tor indus­try. Expert recog­ni­tion from a major indus­try play­er under­scores the com­mer­cial poten­tial of a tech­nol­o­gy that solves prob­lems oth­ers can­not even address.

The investment equation

Every smart­phone, AI accel­er­a­tor, and HPC proces­sor requires capac­i­tors for pow­er sta­bi­liza­tion. Physics dic­tates that clos­er place­ment means bet­ter per­for­mance, and under-chip real estate offers the ulti­mate prox­im­i­ty. Yet until now, no tech­nol­o­gy could deliv­er mean­ing­ful capac­i­tance in this space while meet­ing height constraints.

Smoltek has solved this equa­tion. We’ve joined the exclu­sive 1 µF/​mm² club while stand­ing alone in our abil­i­ty to achieve this per­for­mance with ultra-thin pro­files. As elec­tron­ics con­tin­ue their relent­less march toward high­er per­for­mance in small­er pack­ages, this unique capa­bil­i­ty trans­forms from tech­ni­cal curios­i­ty to com­mer­cial necessity.

The path from lab­o­ra­to­ry break­through to mar­ket dom­i­nance runs through prob­lems that only one com­pa­ny can solve. For investors seek­ing expo­sure to tech­nolo­gies that address seem­ing­ly impos­si­ble con­straints, Smoltek offers capa­bil­i­ties that lit­er­al­ly don’t exist any­where else.