Breaking barriers: The reality

We all are liv­ing in a con­nect­ed world. We and every­thing sur­rounds us for that mat­ter is con­nect­ed or at the verge of con­nect­ing to the extent that we are wit­ness­ing a world with­in a world! Let’s dive into my naive way of look­ing at the real­i­ty as a nan­otech entre­pre­neur and how the oppor­tu­ni­ties for nan­otech star­tups are emerging.

The Reality

In less than a gen­er­a­tion, we have been equipped with so many gad­gets that we have start­ed los­ing count on them. Tech­nol­o­gy has mod­ern­ized us to the extent that we have become habit­u­at­ed to many ges­tures so heav­i­ly that we con­sid­er many func­tions as grant­ed as breath­ing air! For exam­ple, we are so used to the smart­phone touch screen that some­times in our uncon­scious mind we try to touch as soon we find a dis­play, swipe our old com­put­er mon­i­tor or even TV for chang­ing a screen­shot or channel! 

The orig­i­nal hard­ware plat­forms man­u­fac­tur­ers like Intel, Sam­sung, Fujit­su, Qual­com, Cis­co, Eric­s­son, Nvidia etc. have enabled the mon­u­men­tal progress of the soft­ware which in turn cre­at­ed a gigan­tic soft­ware indus­try. The youngest play­er of the game ARM have rev­o­lu­tion­ized the mobile devices through their ARM proces­sors specif­i­cal­ly designed for hand­held mobile devices. Com­pa­ny like Microsoft, Ora­cle, Apple, Nokia and Sam­sung undoubt­ed­ly not only exploit­ed the hard­ware to build soft­ware plat­form on top, but also have pushed dif­fer­ent tech­nolo­gies through high­er demand on mem­o­ry, proces­sor clock speed, enable­ment from only talk to touch screen smart­phones and oth­er hand­held devices.

Such pro­gres­sive hard­ware plat­form has also had empow­ered a new breed of entre­pre­neurs cre­at­ing a bunch of high­ly suc­cess­ful soft­ware giants like Google, Face­book, Pay­Pal, Twit­ter, LinkedIn, Youtube, Airbnb, Aliba­ba, Ama­zon, eBay, Skype, Spo­ti­fy, Viber, What­sApp, Uber, Snapchat and a lim­it­less list can con­tin­ue. Pro­lif­er­a­tion of hard­ware has also helped in adop­tion of well­ness track­ing, and vir­tu­al med­ical care solu­tions. Dra­mat­ic progress in sen­sor man­u­fac­tur­ing in turn enabling self-dri­ving vehi­cles and arti­fi­cial intel­li­gence (AI) with view, touch and more senses.

So far the exist­ing hard­ware plat­form was suf­fi­cient to wit­ness the pro­lif­er­a­tion of soft­ware indus­try. How­ev­er, as the soft­ware indus­try is pro­gress­ing well beyond its imag­i­na­tions, and these soft­ware giants are able to con­nect bil­lions of peo­ple and gen­er­at­ing tril­lions of data out of the soft­ware ecosys­tem lead­ing the sup­port­ing hard­ware to bleed and face tsuna­mi of dif­fi­cul­ties to coup up with in terms of speed, band­width, pow­er man­age­ment and more.

Hence hard­ware tech­nol­o­gy is unfor­tu­nate­ly reach­ing to its lim­its or going out of steams or becom­ing more expen­sive instead of being economical.

Even bit­ter news, as it looks this data band­width demand will be fueled fur­ther up by the push of Inter­net of things (IoT), smart inven­to­ry, smartc­i­ty, Robot­ics, vir­tu­al real­i­ty (VR), aug­ment­ed real­i­ty (AU), arti­fi­cial intel­li­gence (AI), brain read, autonomous dri­ving, bio­mimet­ic recog­ni­tion, human-machine (HI-AI) inter­ac­tions and final­ly per­haps the inter­net of things (IoE).

These appli­ca­tions undoubt­ed­ly cre­at­ing tremen­dous stress on the exist­ing hard­ware plat­form not only to be large­ly revamped but also to be com­ple­ment­ed with a lot of /​quoutiny lit­tle sen­sors, wire­less mod­ules and self-sus­tain­ing pow­er management.

There­fore, with no sur­prise, we wit­ness a shift in hard­ware plat­form mov­ing from gen­er­al pur­pose to appli­ca­tion spe­cif­ic hard­ware accel­er­a­tion. A per­fect exam­ple of how the soft­ware indus­try today is shap­ing the future of hard­ware indus­try is the intro­duc­tion and imple­men­ta­tion of ten­sor pro­cess­ing unit (TPU) . Fas­ci­nat­ing enough, the TPU is not devel­oped by the world largest hard­ware foundry, but by the largest soft­ware indus­try Google! The devel­oped TPU legit­i­mate­ly per­forms oper­a­tions > 15× faster than exist­ing cen­tral pro­cess­ing unit (CPU) or graph­ics pro­cess­ing unit (GPU) in the mar­ket.¹

Great soft­ware shines even brighter with great hard­ware under­neath it

Norm Joup­pi, Google

Indeed, Google is tak­ing the lead in hard­ware accel­er­a­tions by devel­op­ing its pro­pri­etary ten­sor pro­cess­ing units (TPU), sys­tem on a chip (SoC) etc.! Apple, repeat­ed­ly dis­ap­point­ing hard­ware chip design com­pa­nies by chang­ing their mod­el from out­sourced towards in-house hard­ware design.² The oth­er soft­ware giants will soon join the club.

Obvi­ous­ly it is no longer enough to col­lect & store gar­gan­tu­an amount of data to the cloud, the sys­tem need to be intel­li­gent enough to be able to ana­lyze and gen­er­ate pat­terns, trends, prob­a­bil­i­ty etc. to be able to nav­i­gate real time activ­i­ties or per­haps to mon­e­tize from the big data. Such scheme how­ev­er requires sophis­ti­cat­ed algo­rithm and a lot of com­put­ing pow­er to be quick enough to help with data crunch­ing, demands quick analy­sis and turn­ing data into some kind of intel­li­gence so that real time deci­sion mak­ing can be made by AI!

Com­pa­nies like Wall­mart would like to equip their shops with a lot of sen­sors all around to enable IoT inspired cus­tomer ser­vices.³ Such ginor­mous imple­men­ta­tion undoubt­ed­ly will boost cus­tomer sat­is­fac­tion as well as will be excel­lent trig­ger for waste management.

How­ev­er, to add some oil to the fire, most of such hard­ware need to become tiny, slick, self-pow­ered/pow­er effi­cient, eco­nom­i­cal, sus­tain­able and envi­ron­men­tal­ly friend­ly etc. etc. etc. There­fore, the need for minia­tur­iza­tion of hard­ware, nan­otech­nol­o­gy and nano­ma­te­ri­als come into play.

What makes the sit­u­a­tion for the nan­otech star­tups so com­pelling for hard­ware plat­form mak­ers is that VR, AI, IoT, smart inven­to­ry, machine learn­ing oppor­tu­ni­ty is unfold­ing at a time when the mar­ket for tra­di­tion­al tech­nol­o­gy is either flat­ten­ing or run­ning out of steams to sup­port the mar­ket demand in data acqui­si­tion and analy­sis. We are undoubt­ed­ly wit­ness­ing an inflec­tion point for the most devel­oped and exploit­ed CMOS tech­nol­o­gy lead­ing to the col­lapse of decades of stan­dard ITRS roadmap (≈ a par­a­digm shift ≈ any tech­nol­o­gy adop­tion can hap­pen sce­nario ≈ hard­ware accel­er­a­tion). There­fore, oppor­tu­ni­ties emerge now for the nano­ma­te­ri­als startups.

Next article

In the next and final arti­cle my dive will be on ‘the future’ where we may find the inva­sion of nan­otech and nano­ma­te­ri­als to impact in revamp­ing future of hard­ware and hence will dra­mat­i­cal­ly help to unleash the pow­er of inter­net of everything!

1. Source: Google says its cus­tom machine learn­ing chips are often 15–30x faster than GPUs and CPUs, TechCrunch, April 5, 2015 ↩︎
2. Apple will design its own pow­er man­age­ment chips, says report, The Verge, Novem­ber 30, 2017 ↩︎
3. Source: Wal­mart Calls for Sub-$1 IoT Sen­sor, EETimes, Decem­ber 4, 2017 ↩︎