TickerTain
TickerTain
NewsroomShortsPortfolioConvergence
NewsroomShortsPortfolioConvergence
←
▶ 24:52 · AI Hardware & Chip Architecture · Semiconductors · Shift from homogeneous to heterogeneous compute drives chip market diversification
episode briefing
Scaling Europe

An interview with Danyal Akarca, Founder at Callosum & Suraj Bramhavar, Programme Director at Aria

2026-02-26 · 10 company · 6 thematic
sentiment
5 bull0 bear5 neu
speakers
danyal akarca

Founder of Callosum, a stealth startup building software infrastructure to orchestrate heterogeneous AI compute across specialized chips. PhD from Cambridge, previously ARIA seed creator. Raised $10M from Plural to enable multi-agent systems on mixed hardware.

seb

Seb hosts the Scaling Europe podcast presented by Deal, interviewing European tech builders and investors.

suraj bramhavar

Programme Director at ARIA (UK's Advanced Research and Invention Agency), running the Scaling Compute programme and launching the Scaling Inference Lab to build neutral testbeds for heterogeneous AI hardware and software co-design.

now playing · AI Hardware & Chip Architecture · Semiconductors
AI Infrastructuretailwindscore 9/10danyal akarca
Software connective tissue for heterogeneous compute becomes critical value layer
As chip diversity increases, the orchestration layer that matches models to specialized hardware and enables co-design of algorithms and chips becomes the key value capture point. Callosum'…
AI Agentstailwindscore 8/10danyal akarca
Multi-agent systems drive demand for heterogeneous compute orchestration
The paradigm shift from single-model reasoning to multi-agent systems (where models interact with tools, environment, and each other) creates inherently heterogeneous workloads. Different a…
AI Economics & Business Modelstailwindscore 8/10suraj bramhavar
Breaking the hardware lottery: government testbeds de-risk novel chip architectures for private capital
Novel chip architectures face a 'hardware lottery' — high capex and long timelines mean startups need guaranteed customers before tape-out. ARIA's scaling inference lab acts as a neutral te…
AI Hardware & Chip Architecturetailwindscore 9/10suraj bramhavar
Shift from homogeneous to heterogeneous compute drives chip market diversification
AI workloads are splitting into specialized tasks (prefill vs decode, training vs inference vs continual learning vs edge robotics) that require different chip architectures. This inevitabl…
Semiconductorsmixedscore 8/10suraj bramhavar
Nvidia's architecture dominance erodes as inference market diversifies beyond training monopoly
Nvidia retains a natural monopoly in gradient-based training, but the exploding inference market (continual learning, edge, multi-agent robotics) is inherently heterogeneous. Historical pre…
Networking & Optical Infrastructuretailwindscore 8/10danyal akarca
Photonic networking emerges as critical enabler for heterogeneous chip clusters
Light-based chip interconnects (photonics) unlock new system architectures for heterogeneous compute but require co-design with compute chips. This networking layer is a ripe innovation are…