ICYMI: Day 2 Live Coverage from theCUBE at MWC26 in Barcelona

**Kore**: We’re revisiting Day 2 live coverage from theCUBE at MWC26 in Barcelona — four standouts that together map how telecom, edge, and enterprise software are rearranging around AI.

**Achird**: First up: Databricks on how telcos should build AI with an outcomes-first mindset. The problem they framed was familiar — lots of fragmented operational and analytical data, and too many point solutions that don’t deliver measurable business results. The speaker proposed a three-layer approach: start by defining business outcomes, centralize operational/analytical/transactional data, then build apps and AI on top with strong governance. Let’s listen to that core explanation.

> ## Outcomes-First Telecom AI: Define Goals, Identify Data, Build Apps, Monitor Agents — Governance, Unity Catalog, Lineage, Traceability

> It's  a  three- level  house.  They  need  three  levels. "  So,  I  truly  believe,  and  my  friends  often  ask  me  this,  I  say,  " Start  with  the  outcomes.  At  least  get  started.  Start  with  the  outcomes.  Know  what  you  want  to  achieve.  Then,  understand  the  data  that  you  need  to  achieve  that  outcome,  and  then  create  the  application  and  the  AI  to  get  you  there. "  And  working  with  a  company  like  Databricks,  because  telco  is  so  heavily- regulated,  having  that  governance,  full  lineage,  traceability,  access  control,  because  remember,  you  don't  want  anyone  to  access  any  data  because  you  have  100  million  subscribers,  you've  got  a  lot  of  data  in  there.  So,  those  are  the  underlying  principles,  I  would  say,  that  you  need.  You  need  to  know  the  data  you  have.

> You  need  to  have  trust  in  the  data  that  you  have,  and  you  need  to  also  monitor  your  AI  agents.  These  are  not  things  that  you  just  let  run  off.  So,  you  need  to  evaluate  the  monitor  them  because  they  are  actually  delivering  20 %  growth  in  the  telco,  reducing  churn  by  40 %  today.  Now,  imagine  if  you  had  100  million  subscribers  and  you're  churning  at  9%  a  year  and  you  can  reduce  it  by  40 %.  It  has  a  massive  transformational  impact  on  your  P & L  and  you  can  get  to  those  faults  in  your  network  way  before.  So,  start  with  the  outcomes,  ensure  you  have  the  governance  in  your  data,  unify  your  data,  own  your  data,  and  use  both  your  operational  data,  analytical  and  your  transactional.

> [John Furrier] >> They  have  the  keys  to  the  kingdom.  They  have  the  data

**Achird**: A crisp blueprint — the governance point landed hard: lineage, traceability, access controls, and continuous monitoring of AI agents. They even quoted impact estimates — roughly 20% revenue growth and 40% churn reduction potential — which turns the architecture conversation into one about measurable business levers.

**Kore**: That emphasis on measurable outcomes and ongoing monitoring really reframes AI from an experiment to an operational capability. Now, shifting from the core to the edge: how does AI change traffic flows and where compute needs to live?

**Achird**: Chris Lewis addressed that tension between exploding datacenter traffic and an access network that, so far, shows only a small AI footprint. He described edge node rollouts and the prospect of on-device models shifting patterns toward much heavier uplink usage — a potential industry inflection that could force telcos to rethink service models, pricing, and partnerships. Here’s his take.

> ## Uplink Surge and On‑Device AI: How Small Edge Nodes Challenge Data Center Traffic Dominance

> great.  So,  what  is  fascinating  is  that  with  all  the  moving  parts  of  the,  let's  call  it  the  AI  world,  the  reliability,  almost  the  conservatism  of  the  telecom  industry  in  some  ways  plays  to  its  advantage,  because  connecting  things...  Let's  break  it  down  to  the  way  things  connect.  Between  data  centers,  that  growth  has  been  massive.  And  of  course,  luckily  we've  got  the  optics  to  do  that.  So,  the  capacity  is  there.  We're  pumping  more  capacity  in.  At  the  edge,  and  the  edge  is  finally,  I  mean,  after  what,  10  years  of  talking  about  the  edge.  The  edge  is  beginning.  There  ere  some  announcements  this  week,  Telefónica  here  in  Spain  with  Nokia  starting  to  build  some  little  edge  nodes  around  the  country  to  help  deliver.  But  what's  interesting  is  that  the  real  edge,  out  to  you  and  I  as  users  or  individuals  in  the  business,  so  on,  AI's  not  really  impacted  that  traffic  yet.  It's  about  1%  of  access  traffic.

> So,  we've  not  seen  it  go  right  out  to  the  edge.  And  what  we're  seeing  at  this  show  with  the  announcements  from  people  like  Qualcomm  in  those  edge  devices,  in  devices  which  will  have  their  own  models  running  in  the  near- term,  that  may  well  change  it.  And  if  you  talk  to  Ericsson,  they're  talking  about  the  uplink  becoming  much  more  important.  So,  we're  moving  away  from  a  predominantly  downlink  model  to  an  uplink  model.  So,  I  think  we  have  to  be  very  careful  not  to  confuse  the  explosion  of  traffic  going  on  in  data  centers,  between  data  centers  and  what's  actually  happening  out  on  the  access  network,  because  of  course,  on  the  access  network  is  where  the  telcos  actually  make  most  of  their  money.  So,  they're  selling  to  you  and  I,  they're  selling  to  businesses,  they're  selling  into  the  ecosystem  and  partnerships.  So,  that's  not  changing.  The  demand  will  change  in  traffic  flows  in  the  future.  What  we  connect  at  the  edge  will  change.  And  interestingly,  during  the  judging  of  some  of  the  GLOMOs  this  year,  I  came  across  the  first  robo  guide  dog,  which  obviously  I'm  interested  in  that  for  future  years,  whether  that  will  help  me  get  around  the  show  or

**Achird**: That perspective makes the uplink a plausible new bottleneck — especially for consumer-facing services that start sending model updates and telemetry upstream. The implication is clear: network planning, QoS, and commercial models need to be proactively redesigned, not retrofitted.

**Kore**: Next, a product-level example of how vendors are adjusting: Juho Sarvikas on Inseego Subscribe and why the company is repositioning BSS as SaaS.

**Achird**: Juho unpacked Inseego Subscribe as a lifecycle platform that handles telecom expense, catalogs, rate plans, and even complex FED/SLED purchasing workflows — essentially everything short of the billing ledger. He highlighted carrier configurability and self-service portals as ways to lower acquisition and account management costs, plus a strategic move from partner-delivered projects toward a broader SaaS offering. Let’s hear him explain the product shift.

> ## Inseego Subscribe: BSS lifecycle subscriber management platform, SaaS expansion, FED/SLED focus

> [Juho Sarvikas] >> Excellent.  I  think  we  covered  these  two  products  maybe  already  and  the  mobile  announcement  we  made,  which  really  has  reinvented  the  category.  The  other  big  announcement  for  us  is  a  well- kept  secret.  So,  we  also  have  Inseego  Subscribe.  Now,  forget  what  we  just  discussed.  I'll  bring  it  back  together  shortly.  But  for  now,  just  forget  what  we  discussed  with  the  manageability  and  the  devices.  So,  we  have  Inseego  Subscribe,  which  is  our  BSS.  It's  a  business  support  solutions.  We  have  lifecycle  subscriber  management  platform  that  the  carriers  can  use.  We  effectively  have  everything  in  the  platform,  absent  a  biller.  So,  think  about  not  only  telecom  expense  management,  but  a  much  broader  experience.  We  specialize  in  the  FED / SLED  public  sector,  where  you  have  a  lot  of  purchasing  contractual  items  that  you  need  to  be  able  to  manage,  all  of  that  complexity.

> And  one  of  the  key  benefits  from  Inseego  Subscribe  is  that  we  lower  the  cost  of  acquisition.  But  not  only  that,  also  the  lifecycle  cost  of  managing  the  account  by  providing  tools  where  the  customers  can  self- serve.  So,  let's  say  that  you're  managing  a  large  enterprise  in  the  US  as  a  carrier  sales  rep.  You  can  configure  the  catalog,  configure  the  rate  plans,  you  can  serve  the  customer,  but  they  can  also  enter  the  portal  and  self- serve.

> [Dave Vellante] >> So,  you  joke  that  it's  a  well- kept  secret  because  it's  been  in  market,  right?

> [Juho Sarvikas] >> Yeah,  that's  a  good  point.  We  have  a  partner  who  has  been  out  there  well  over  a  decade.  And  now,  the  new  news  is  that  we're  making  it  more  of  a  SaaS  type  of  a  play  and  expanding  that  to  the  broader  audience.  So,  that's  very  exciting  for  us.

**Achird**: That shift to a SaaS BSS — and the focus on self-service for complex public-sector purchases — reduces friction for large buyers and lets vendors scale beyond bespoke integrations.

**Kore**: Finally, IBM’s macro view on core enterprise applications and why ecosystems matter more than ever.

**Achird**: Jason Kelley framed core stacks — ERP like SAP/Oracle, front-ends like Adobe and Salesforce, plus Workday and ServiceNow — as the backbone that must be stitched together. His argument: acquisitions create heterogeneous landscapes, and delivering AI-driven productivity at scale requires ecosystem partnerships and practical integration approaches. He also referenced “Client Zero” — IBM using its own tooling internally to validate value delivery. Here’s his summary.

> ## Core business applications: ERP, CRM, HR, ServiceNow, Adobe — IBM manages partner ecosystem

> [Jason Kelley] >> Core  Business  Applications,  key  word  there  is  core.  So  think  of  those  apps  that  are  core  to  any  of  your  given  enterprises,  there's  always  an  ERP  in  the  backbone.  So  you're  going  to  have  an  SAP  or  an  Oracle.  There's  going  to  be  a  front- end  to  that,  an  Adobe,  a  Salesforce.  You're  going  to  have  capabilities  of  a  Workday  as  well  as  a  ServiceNow.  And  so  those  are  the  core  business  applications  that  my  teams  work  with  globally  and  continuing  to  see  the  business  grow  exponentially.

> [John Furrier] >> You  guys  have  had  great  ecosystem  relationships.  The  stock  price  is  at  all- time  high.  You  guys  are  performing  really  well.  The  Client  Zero  stuff,  all  that  productivity  of  AI,  we  heard  that  this  morning  in  your  booth.  Talk  about  the  ecosystem  relationship  to  the  company  and  why  that's  so  important  to  the  new  IBM,  I'll  call  it  the  new  IBM,  again,  thundering  away  on  the  business  performance.  So,  yeah-

> [Jason Kelley] >> We'll  take  the  new  IBM  any  day  you  want  to  say  that,  as  long  as  the  new  means  the  continually  growing  and-

> [John Furrier] >> All- time  high....

> [Jason Kelley] >> value  to  our  clients.  And  I  tell  you  that,  that's  where  the  thought  of  an  ecosystem  has  really  evolved.  Our  clients  aren't  ignoring  the  fact  that  they  have  multiple  applications.  Many  of  your  large  enterprise  clients  have  grown  over  the  years  through  acquisition.  So  they  may  have  made  a  landscape  decision  at  one  point,  but  then  as  they  acquire  more,  that  landscape  changes.

**Achird**: That “Client Zero” idea is important — it’s not just about building integrations on paper, it’s about proving they deliver day-one value inside your own operations before you sell them outward.

**Kore**: Quick recap — across these clips we heard consistent themes: start with outcomes; centralize and govern trusted data; expect the edge and on-device models to reshape traffic and commercial models; and embrace SaaS and partner ecosystems to stitch messy enterprise landscapes together.

**Achird**: If there’s one takeaway: telcos and vendors that align strategy, tight data governance, and partner ecosystems will capture AI’s real business value — while the rest risk creating complexity without trust or measurable impact.

**Kore**: Thanks for joining our ICYMI Day 2 roundup from theCUBE at MWC26. Stay curious — the network, the edge, and enterprise apps are all rewriting each other right now.

**Achird**: Until next time — keep listening, keep learning, and we’ll see you at the next show.