Box CEO on the AI Adoption Gap | The a16z Show

Watch on YouTube ↗  |  April 08, 2026 at 14:30  |  58:29  |  a16z
Speakers
Martin Casado — a16z General Partner
Steven Sinofsky — Host
Aaron Levie — CEO of Box

Summary

The panel discusses how enterprise software changes when AI agents become primary users. They debate coding agents, agent security and identity, the diffusion gap between startups and large enterprises, and the economics of AI compute. Market implications center on agent-ready software, AI infrastructure demand, and incumbent systems of record like SAP.

  • Enterprise software must increasingly be designed for agents, not only humans.
  • Coding agents and computer-use agents are advancing, but algorithmic thinking remains a human bottleneck.
  • CFOs and CIOs worry that agents and humans doing integrations could break systems of record.
  • Treating agents like employees breaks down on security, identity, oversight, and privacy.
  • Startups can adopt AI faster than large enterprises due to fewer legacy and compliance constraints.
  • Compute, token budgets, and data center capacity are central economic debates.
  • Wall Street may underestimate AI's market-expansion opportunity by an order of magnitude.
Ideas
Aaron Levie CEO of Box 0:52
Software must be built for agents.
If agents outnumber humans by 100x or 1000x, software must be built for agents. Winning enterprise software companies will need high-quality APIs, monetization, and identity/access controls for agents, and enterprise IT stacks must support agent access.
IGV HIGH
Steven Sinofsky Host 29:47
SAP's domain moat protects it from AI.
SAP has deep domain knowledge embedded in its UI, middle tiers, and workflows, not just a data layer. It is absurd to think AI can vibe-code a replacement, so SAP is not going anywhere and will slow AI diffusion on its data.
SAP HIGH
Aaron Levie CEO of Box 32:50
More agents mean more files for Box.
Agents love working with files, so an agent-first world likely creates more files, not fewer. Box is building a platform and CLI for agents to work with data and expects this to be an optimistic outcome for its business model.
BOX MED
Steven Sinofsky Host 42:53
Wall Street underestimates AI's expansion opportunity.
Wall Street is trying to justify AI capex with linear, fixed-revenue-pie models, but the opportunity is likely at least an order of magnitude larger. New technologies like PCs, cloud, and SaaS expanded markets far beyond initial estimates, and AI will create new business models and consumption.
AI-SECTOR HIGH
Martin Casado a16z General Partner 45:08
Agents drive massive AI compute demand.
As more software is written and agents proliferate, consumption of infrastructure and computer resources is accelerating. Martin says his portfolio of infrastructure companies has gone asymptotic in six months, and he expects much more compute consumption as agents and phones consume AI.
AIQ HIGH
Up Next
2:26:11
Bloomberg Surveillance 4/8/2026
Bloomberg Markets · Apr 08
59:04
Henrik Zeberg: "We Have Not Seen The Top Yet" — Why The Nasdaq Could Rally 30%
Julia LaRoche Show · Apr 08
4:07
Markets will soon go back to being driven by AI investment, says Aperture’s Peter Kraus
CNBC · Apr 08
1:17
Sec. Hegseth: 'Achieved every single objective,' says Iran suffered a 'defeat'
CNBC · Apr 08
1:04:40
$1 Million Is the Worst Amount of Money | Animal Spirits 459
The Compound News · Apr 08

This a16z video, published April 08, 2026, features Aaron Levie, Steven Sinofsky, Martin Casado discussing IGV, SAP, BOX, AI-SECTOR, AIQ. 5 trade ideas extracted by AI with direction and confidence scoring.

Speakers: Aaron Levie, Steven Sinofsky, Martin Casado  · Tickers: IGV, SAP, BOX, AI-SECTOR, AIQ